Advanced Transportation Engineering
Advanced Transportation Engineering
Advanced Transportation Engineering
Module
Topic Preface Pavement Basic concept of pavement analysis and design Lecture - 1 Basic concept of pavement analysis and design Analysis and design of concrete pavements Lecture - 1 Pavement Materials Lecture - 2 Analysis of Concrete Pavement Lecture - 3 Various design approaches Lecture - 4 Design of dowel bars Lecture - 5 Design of tie bars Few other topics on concrete pavements Lecture - 1Design of runway, taxiway and apron Lecture - 2 Reinforced Concrete Pavement Lecture - 3 Concrete pavement shoulder Lecture - 4 Composite Pavements Lecture - 5 Comparative studies of concrete versus bituminous pavement Construction and maintenance of concrete pavement Lecture - 1 Construction of Concrete Pavements Lecture - 2 Construction of joints Lecture - 3 Quality control issues Lecture - 4 Maintenance of concrete pavements Traffic Engineering Basic concepts of traffic Engineering Lecture - 1 Basics of Traffic Engineering - A Recapitulation Lecture - 2 Speed and Density Lecture - 3 Flow Problem Session Uninterrupted Flow - I Theory of Uninterrupted Traffic Flow Lecture - 1 Introduction & Fundamental Relation Lecture - 2 An Introduction to Macroscopic Models Lecture - 3 Macroscopic Models- Linear Model/ Generalized Polynomial Model of u - k relation Lecture - 4 Macroscopic Models - Logarithmic Model of u-k relation/Exponential Model of u-k Relation Lecture - 5 Interrupted Flow Lecture - 1 Introduction to Interruptions Lecture - 2Traffic Flow at Signalized Intersections Lecture - 3Traffic Flow at Unsignalized Intersections Capacity and LOS analysis Lecture - 1 Introduction and capacity of facilities with uninterrupted flow Lecture - 2Capacity at signalized and unsignalized intersections and LOS Transport Planning Traffic Planning Process Lecture - 1 Introduction, transport supply and demand Lecture - 2 Overview of the planning process, role of government
II
III
IV
VI
VII
VIII
IX
Basic concept of pavement analysis and design Objectives The objective of this lecture module is to introduce the basic concepts of analysis and design of pavement structure. Also a discussion has been placed as past, present and future perspective of pavement analysis and design Types of pavement structures
Pavements, in general, can be classified in two major categories: concrete pavement and bituminous pavement. Concrete pavements are generally called rigid pavements and bituminous pavements as flexible pavements. There could be some other types of pavements which are neither rigid, nor flexible, for example, block pavement, composite pavement. A Pavement is a multi-layered structure. The layers are placed horizontal one over other. In general, the strengths of the layers decrease from top towards bottom except some special situation like inverted pavement. The terminologies used to identify various layers of bituminous and concrete pavements are identified in Figs. 1and 2
. Fig. 1 cross section of a typical bituminous pavement (Chakroborty and Das 2003)
Fig. 2 cross section of a typical rigid pavement (Chakroborty and Das 2003)
The subgrade is a compacted soil layer. The base and sub-base course could be made up of bound or unbound granular layer. As per Indian specifications (MORT&H 2001), some examples of base or
sub-base layers are: Granular sub-base(GSB), Water Bound Macadam (WBM), Wet Mix Macadam (WMM) etc. The binder course is made up bituminous material. As per Indian specifications (MORT&H 2001), some examples of binder course are: Bituminous Macadam (BM), Dense Bituminous Macadam (DBM) etc. The wearing course is the top bituminous layer which is comes in contact to the vehicle tyre. Wearing course provides impermeability to the pavement surface against water percolation (Chakroborty and Das 2003). The binder course and wearing course together are called bituminous surfacing.
Concrete pavement
Concrete pavement is, in general, consists of three layers, subgrade, base layer and the concrete slab. Generally bound base layers are used for concrete pavement construction. As per Indian specification, some example of such base layers are Dry Lean Concrete (DLC), Roller Compacted Concrete (RCC) (IRC:15-2002) The concrete slab is generally of M40 to M50 grade of concrete as per Indian specifications, and is called as paving quality concrete (PQC) (IRC:15-2002).
Joints are the discontinuities in the concrete pavement slab, and help to release stresses due to temperature variation, subgrade moisture variation, shrinkage of concrete etc. There are various types of joints in concrete pavement, e.g. contraction joint, construction joint, expansion joint and warping joint. Fig. 3 schematically shows position of various joints. The functions of these joints are as follows: Contraction joint: Contraction joints are provided along the transverse direction to take care of the contraction of concrete slab due to its natural shrinkage. Construction joint: Construction joints are provided whenever the construction work stops temporarily. The joint direction could be either along the transverse or longitudinal direction. Expansion joint: Expansion joints are provided along the transverse direction to allow movement (expansion/ contraction) of the concrete slab due to temperature and subgrade moisture variation. Warping joint: Warping joints are provided along the longitudinal direction to prevent warping of the concrete slab due to temperature and subgrade moisture variation.
These discontinuities (joints) could be extended to the full or partial depth of the slab. Sometimes iron bars are provided across the joints, the iron bars along the longitudinal joints are called tie bars andalong the transverse joints are called dowel bars. Pavement analysis and design: historical perspective The past pavement design approaches were mostly empirical in nature and were based on experience.
CBR method of pavement design is one of the earlier pavement design approach developed during 1928-29 (Ullidtz 1986). In this method the thickness design charts are developed (based on observation of number of sections), with reference to the subgrade CBR value for the most critical moisture condition. In 1940 this method was adopted by the U.S. Corps of Engineers for design of airfield pavements (Horonjeff and Mckelvey 1983). It is interesting to note that the design initially did not involve traffic as a parameter, which was introduced later as a correction factor. The method was further improved by considering the CBR values of the individual layers and thereby individual layer thicknesses are obtained. In some other approach, Hveem resistance value of pavement materials is used instead of CBR value. Another pavement design approach considers aspect of bearing capacity of the individual layers, and the design is finalized in such a way that the bearing stress does not exceed the bearing capacity of the individual layers. This method was first proposed by Barbar in 1946, and is still in use (TRH4 1996, deBruin et al. 2002), however this method does not seem to account for traffic repetitions. Another approach recommends limiting recoverable deflection as the criterion for pavement design (Huang 1993). Failure theories suggest that the failure of a structure is due to excess stress or strain, thus, deflection may not be attributed as basic pavement design criteria.
Pavement analysis and design : current perspective Present practice of pavement design involves considerations of three aspects: structural design, functional design and drainage design and they are explained briefly in the following: Structural design In structural design the stresses due traffic loading and temperature are estimated, and the thickness of the pavement is designed in such a way that these developed stresses/ strains are below the allowable values. The current practice of pavement design, more popularly, is known as Mechanistic-Empirical pavement design and is followed by a number of organizations around the world (Asphalt Institute 1999, Shell 1978, Austroads 1992, NCHRP 2005, IRC 2001). It is mechanistic pavement design because it uses stress/ strain of a pavement structure using mechanics based principle, and, as well, it is empirical because the expected life for a given stress/ strain level is estimated from empirical relationships obtained from laboratory or field performance studies. The pavement design approach is not governed by the maximum amount of load that the pavement can sustain, rather, it estimates the number of standard load repetitions that can cause failure. Estimation of pavement stress/strain Stress/ strain due to load
For pavement design purpose, the stress/ strain value of a pavement structure is obtained from structural analysis of the pavement (Ioannides et al. 1998). The stress/ strain values at any point of a pavement structure can be estimated when the elastic moduli, Poisson's ratio and the thicknesses of the individual layers are known. The strain values can also measured using strain gauges. Any analysis procedure involves idealization regarding the structure; similarly, measuring strain involves measurement errors - hence the true value of stress/ strain is never known. A concrete pavement slab, in general, has finite dimensions, and thus the analysis approach of concrete pavement becomes different than the analysis of bituminous pavement. For bituminous pavement, in general, the pavement is assumed as infinite in both the directions, whereas for concrete pavement, in general, it is analysed as discrete slabs connected by joints. The concrete pavement is also assumed to have bending moment carrying capacity, whereas flexible pavement is assumed to have no moment carrying capacity.
The change of temperature causes the pavement to expand or contract. The restriction of free movement causes temperature stresses. There exists temperature variation across the depth of the pavement - this causes warping stresses. The temperature stress varies across the corner, interior and edge of the concrete slab, also at different times of the day. The most critical combination of load and temperature stress is used as design criteria. The temperature stress in bituminous pavement is insignificant. Hence, temperatire stress, is not considered in pavement design. However, temperature affects the elastic modulus of the bituminous layer, which needs to be duly considered in pavement design.
The thickness of the pavement is adjusted in such a way that the stress/ strain developed is less than the allowable values obtained from past performance information. The two major modes of structural failure of pavement are fatigue and rutting. o Fatigue: Traffic applies repetitive load to the pavement surface, and the cracks start from bottom the bound layer/ slab and propagate upwards. When the extent of surface cracks reaches a predefined level, the pavement is said to have failed due to flexural fatigue.
Conventionally, for design of concrete pavement stress is used as parameter, and for design of bituminous pavement strain is used as parameter.
Rutting: Rutting is the accumulation of permanent deformation. This is the manifestation of gradual densification of pavement layers, and shear displacement of the subgrade.
The vertical strains on the pavement layers, mainly the vertical strain on the subgrade is assumed to be governing factor for rutting failure. The rutting issue is not considered for concrete pavement design, because it does not have any permanent deformation.
The fatigue/ rutting equations are developed from field or laboratory studies, where fatigue / rutting lives are obtained with respect to respective stress/ strain for fatigue/ rutting. For a given design life, thus, allowable fatigue and rutting stress/ strains can be estimated using the fatigue/ rutting equations. The various other types of pavement failures could be shrinkage, thermal fatigue, top down cracking (for bituminous pavement) etc. Design of joints The spacing of the contraction joint is estimated from the shrinkage potential of concrete. The spacing of the expansion joint is estimated from the coefficient of thermal expansion of concrete, maximum change of temperature and the acceptable joint gap. Since, the concrete is good in compression, the experience over last few decades indicates that concrete pavement can be constructed without any provision of expansion joint (ACPA 1992). The dowel bars are designed by assuming that they participate in the load transfer, when the vehicle moves from one slab to other. The tie bars are designed in such way that they have enough strength to tie the two adjacent slabs. The design of dowel bar and tie bar is discussed in detail later Functional design The functional pavement design involves considerations of skid resistance, roughness, surface distresses, reflectivity of pavement surface etc. The functional pavement design considers mainly the surface features of a pavement. Drainage design A road needs to be designed in such a way that the rain/ snow precipitation is drained off the pavement and its surroundings. A suitable surface drainage system for the pavement is designed for this purpose. Some water, however, will percolate into the pavement from its top surface and needs to be taken out of the pavement - this is done by providing an internal drainage system to the pavement. Water will also try to enter into the pavement from bottom due to capillary rise or due to rise in water table. A suitably designed sub-surface drainage system tries to avoid such a problem. Pavement analysis and design: future perspective
Mix design, quality control, construction method and pavement design together determines the performance and longevity of a pavement. The future pavement design is expected to take an integrated design approach considering all these issues together. The parameters associated with pavement design are stochastic in nature. Thus, the two pavement designs (designed deterministically) having same design traffic may have different levels of reliabilities of survival. Thus, reliability issues of pavement design are gradually becoming important considerations (NCHRP 2005). A pavement designer essentially looks for the most economical design, yet considering the structural, functional, and drainage design requirement. The future pavement design practice is expected to consider the cost optimality over the entire life cycle of the pavement (Abaza and Abu-Eisheh 2003).
Recapitulation The present module highlights the basics of analysis and design of bituminous and concrete pavements. Pavements are analysed as layered horizontal structure with given elastic moduli and Poisson's ratio. Concrete pavement, in general, is made up of discrete slabs - therefore, it has joints both in longitudinal and transverse direction. A
pavement is designed from structural, functional and drainage considerations. Fatigue and rutting are two major modes of structural failure of pavements. For concrete pavement design, temperature stresses are also considered along with stresses due to load. A pavement designer does not design a pavement for the ultimate load the pavement can carry, rather, the number of repetitions that the pavement can sustain.
Analysis and design of concrete pavements Objectives The lectures in this module propose to introduce the modern materials in pavement construction. It discusses about the scope, application potential, evaluation, and performance expectation of the new highway materials. The second part of the lectures focus on the innovative application concepts of the conventional or the modern materials. Usage of modern materials in highway construction and their innovative application is expected to bring economy in terms of material cost as well as better reliability in performance. Bitumen as a pavement material The characterization of bitumen and bituminous mix has been discussed in detail in the web-course Transportation Engineering - I Bitumen is a complex material, its property ranges from viscous liquid to brittle solid. While bitumen shows linear viscoelastic behaviour at small strains, the nonlinear behaviour becomes more prominent at large strains (Monismith and Secor 1962, Pagen 1968, Cheung and Cebon 1997). The deformation of bitumen is loading rate and temperature dependent (Van der Poel, 1955, Deshpande and Cebon 1997). The bituminous mix is manufactured by mixing bitumen and aggregates of specified size distribution at some specified elevated temperature. Then, the mix is transported to the site, laid and subsequently compacted to pack the aggregate particles together. During the compaction process the air voids are brought down to its desired level. The compacted mix, thus, achieves its strength when it cools down and becomes serviceable as bituminous road . Figure-1 shows a typical cross-section of a bituminous mix sample.
Figure-1 A typical cross-section of a bituminous mix sample The mechanical behaviour of bituminous mix has been studied extensively through various tests, and empirical relationships have been developed for mix design and prediction of the performance of the mix. However, prediction of response of bituminous mix through mechanics based models, is a difficult task. Various attempts have been made by the researchers, for example based on, linear viscoelastic principle (Lee and Kim 1998, Kim and Little 2004), elastic visco plastic principle (Uzan 2005), discrete element analysis (Sadd 2004, Abbas et al. 2005) etc., so as to capture the complex mechanical behaviour of bituminous mix. Cement Concrete as a pavement material Introduction Cement concrete is a mixture of coarse aggregates, fine aggregates, cement and water in suitable proportions. Sometimes admixtures are also added to achieve specific behaviour/ property of the material. The components of cement concrete are briefly introduced in the following. Components of cement concrete Aggregates Aggregates are naturally available pieces of rocks. The aggregates could be igneous, sedimentary and metamorphic type depending on its origin. Figure-1 shows a photograph of aggregates being manufactured from a stone query. The details about the physical properties of aggregates have discussed in the webcourse on Transportation Engineering-1 .
Figure 2: A typical stone quarry Cement Cement is manufactured by heating a mixture of limestone, iron ore, gypsum, clay and other ingredients. Two processes, namely dry process and wet process are followed while manufacturing cement. In the dry process, the raw materials are mixed in dry state, whereas in the wet process raw materials are mixed in presence of water to form slurry . After pre-heating, the raw material is passed through rotating kiln inclined with a small angle with the horizontal line. The kiln is progressively hotter towards its lower end, where the raw material gets molten. From this clinkers are formed when cooled, and after grinding the clinkers, cement is produced. An animated description of the whole process can be obtained elsewhere (cement.org 2006). The Ordinary Portland Cement (OPC) is the most popular, all-purpose cement. There are various other types of cements (for example, natural cement, Portland pozzolanic cement, high alumina cement, expansive cement, quick setting cement, high performance cement, sulphate resistant cement, white cement etc.) and are manufactured to serve specialized purposes. For concrete pavement construction, OPC is most commonly used. Water Water participates in the hydration process; also it provides desirable level of workability. About one third of the water added is utilized in the hydration process, rest forms the pores of concrete, and thereby developing porosity to the concrete. Excess porosity reduces strength of the concrete, and however presence of porosity is good for the situations where there is a freeze-thaw problem. Admixtures Admixtures are generally of two types, chemical admixture, and mineral admixture. Air entrainer, retarder , accelerators are examples chemical admixture, and, fly ash, silica fume are the examples of mineral admixtures. One of the important concrete admixtures used in pavement construction is the air-entraining admixture. Air entraining admixtures are derived from natural wood resins, fats, sulfonated hydrocarbons and oils etc (Wright
and Dixon 2004). Air-entraining admixtures provide durability against freeze-thaw situation. Plasticizers may be used for concrete pavement construction purposes which maintain workability without having increased the water-cement ratio. Calcium chloride is also used sometimes, as accelerating agent, which renders an early strength of concrete Mix design Through mix design, suitable proportions of the ingredients (coarse aggregates, fine aggregates, cement, water and admixture, if any) are estimated, keeping in view the strength, workability, durability and economic considerations. These proportions are achieved through iterative experimental procedure in the laboratory. There are number of methods for mix design of cement concrete, and a detailed discussion can be obtained elsewhere (Neville and Brooks 1999). Water-cement ratio is an important consideration in the mix design process. As water cement ratio is increased in concrete, the durability and strength decreases, however, the workability enhances. Depending on the type of construction, workability requirements are different. For large scale production of cement concrete, the proportioning operation is performed in the batch mixing plant . Figure 3 shows a photograph of a typical concrete batch mixing plant.
Figure 3: A typical cement concrete batch mixing plant Properties of fresh concrete Ideally a fresh concrete should be workable, should not segregate or bleed during construction. Constituent properties, their proportions, aggregate shape and sizes, temperature affect the performance of fresh mix. The tests that are conducted on fresh concrete include workability test and air-content test. Some of tests through which workability of can be estimated are Kelly ball penetration test, slump test, compacting factor test, Vee bee test and flow table test etc. Curing of concrete Presence of adequate amount of moisture, at some requisite temperature and for a suitable period of time, is necessary to complete the hydration process of cement. This process is called curing. The curing conditions significantly affect the final strength achieved by the concrete. For pavement construction, only in-situ curing methods are applicable. Curing compounds are sometimes applied to retain the moisture against evaporation. For final curing of concrete pavements continuous ponding or moistened hessain/ gunny bags are used .
Properties of hardened concrete Tests are conducted on hardened concrete to estimate properties like, compressive strength, tensile strength, modulus of rupture, elastic modulus, Poisson's ratio, creep and shrinkage performance, durability, thermal expansion coefficient etc. These parameters are of functions of aggregate type, shape and size, type and quantity of cement and admixtures incorporated, water cement ratio, curing, age etc. Compressive strength of concrete is the failure compressive stress on cubical or cylindrical samples of concrete. Compressive strength of concrete is related to the combined effect of temperature and time, a parameter called maturity. Maturity of concrete is calculated as the time of curing (in hours), multiplied by the temperature, (in degrees) above some specified reference temperature. Various empirical relationships are suggested to obtain the various strength parameters of concrete (elastic modulus, tensile strength, bending strength etc.) from the compressive strength of concrete. Direct tension test on concrete is performed by applying tension to the cylindrical or dumble shaped samples of concrete. Indirect tension is applied to concrete samples by split cylinder test. Modulus of rupture of concrete is estimated by measuring the maximum bending stress on concrete beam subjected to pure bending in static condition. Fatigue test is generally performed subjecting the concrete beams with repetitive flexural loading. The more is the stress ratio (defined as the ratio between the bending stress applied to the modulus of rupture) the less is the fatigue life. The empirically derived fatigue equation by PCA (1974) is the following: (1) and (2) Where, Nf is the number of load applications to failure, SR is the stress ratio with reference to 90 days modulus of rupture. The equation suggested by AASHTO (1993) is the following: (3)
Transportation of concrete The transportation of concrete is to be done in such a way that segregation and premature setting is avoided. Wheel barrow, truck mixer, dumper truck, belt conveyor, pipe-line etc. are the various ways concrete is transported to the construction site. Figure 4 shows a typical truck concrete mixer.
Figure 4: A typical truck concrete mixer Introduction Road is a costly infrastructure to build and maintain. Thus there is always a need ofr development of (i) new road materials as well as (ii) innovative applications of existing/new materials. These issues are discussed here. EMERGING ROAD MATERIALS Modification of Existing Materials Existing materials may require modifications so as to
improve engineering properties of material satisfy general specification requirement of locally available material which in turn would prove to be cost effective meet the demand of special purpose materials having specific properties. Example: high or low permeability, enhanced shear strength etc
Binder (bitumen) modification Binder (bitumen) modification is done with the help of additives which may or may not react chemically with bitumen. Table 1 presents a partial list of various types of binder modifiers, their purpose and examples. Binder modification results improvement of one or more properties of the binder (and hence the mix) viz. fatigue resistance, stiffness modulus, rutting resistance, stripping potential, temperature susceptibility, oxidation potential etc. Table 1. Some examples of binder (bitumen) modifiers (RILEM 1998; ETM 1999; Asphalt Handbook 2000; Widyatmoko 2002, SEAM 2004 ) Purpose Polymers Fillers to improve bitumen Lime, carbon black, fly ash Examples
durability and check rutting Anti-oxidants Anti-stripping additives to check hardening oxidative Zinc anti-oxidants, lead antioxidant, phenolics, amines (like
to achieve better Organic compounds adhesion of bitumen to arnines, andamides) aggregates to act as bitumen Lignin, sulphur substitute and to improve fatigue resistance -
Extenders
Others
Shale oil, gilsonite, silicon, inorganic fibers, Trinidad lake asphalt (TLA)
Non Polymers Fibers Plastics -Thermoplastics to increase the Polyethylene (PE), viscosity and stiffness Polypropylene (PP), Polyvinyl of bitumen at normal chloride (PVC), Polystyrene service temperatures (PS) Ethylene vinyl acetate without compromising (EVA). with fatigue performance Epoxy resins to obtain insoluble, infusible material that do not flow on heating 3. Elastomers - Natural - Synthetic - Reclaimed rubbers to reduce temperature susceptibility and Rubber temperature distresses, age- Styrene-butadiene copolymer hardening, bleeding (SBR), Styrene-butadieneand binder-aggregate styrene copolymer (SBS), stripping. Ethylene-propylene-diene terpolymer (EPDM), Isobuteneisoprene copolymer (IIR) to reduce viscosity, as Polyester fibers, Polypropylene filler material, fibers
-Thermosets
For conventional binders, it is generally observed that the mixes with high stiffness modulus (E) show low fatigue life, and vice versa. However, for an economical pavement design, both high elastic modulus as well as high fatigue life is desirable. Through binder modification, this particular disadvantage can be avoided. Figure 5 presents this concept schematically.
As can be seen in Figure 5, for mixes with ordinary binder, although elastic modulus E value is higher initially at low temperatures, at high E value the fatigue performance generally becomes poor. On the other hand, at high temperature the E value becomes too low and the mix becomes soft. The bituminous mixes with modified binder does not allow the mix to be too hard (high E value) or too soft (low E value) at low and high temperatures respectively. Thus the stiffness versus temperature curve takes a 'Sshape' as shown in Figure 5. Aggregate modification
The marginal or poor quality aggregates can be improved by using some cementing material such as cement, lime, pozzolanic substance etc. The proportions of the cementing material and other ingredients (like water) can be suitably estimated in the laboratory.
Given the fact that good quality aggregates are depleting and cost of material extraction is increasing, researchers are looking for suitable alternative materials. The tests and specifications, which are applicable for conventional materials, may be inappropriate for evaluation of non-conventional materials ( i.e. alternative materials). This is because the material properties, for example, particle sizes, grading and chemical structure, may differ substantially from those of the conventional materials. Thus, for an appropriate assessment of these materials, new tests are to be devised and new acceptability criteria are to be formed. However, with the advent of performance-based tests, it is expected that the performances of the conventional as well as new materials can be tested on a same set-up and be compared.
Industrial and domestic waste products provide a prospective source of alternative materials. These materials are cheaply available. Also, their use in road construction provides an efficient solution to the associated problems of pollution and disposal of these wastes.
Table 2 presents a partial list of industrial waste materials that can be used in road construction. Table 3 summarizes the advantages and disadvantages of using specific industrial wastes in road construction.
Table 2. Industrial waste product usage in road construction (TFHRC 2004; Hamad et al., 2003; Hawkins et al., 2003; Mroueh et al., 2002; Okagbue et al. , 1999; Sherwood 1995; Javed et al., 1994) Waste product Source Possible usage Fly ash Blast slag Thermal station furnace Steel industry power Bulk fill, filler in bituminous mix, artificial aggregates Base/ Sub-base material, Binder in soil stabilization (ground slag)
Construction and Construction industry Base/ Sub-base material, bulk-fill, recycling demolition waste Colliery spoil Spent oil shale Foundry sands Mill tailings Coal mining Petrochemical industry Foundry industry Mineral industry Bulk-fill Bulk-fill Bulk-fill, filler for concrete, crack-relief layer
processing Granular base/sub-base, aggregates in bituminous mix, bulk fill Stabilization of base, binder in bituminous mix Filler in bituminous mix Glass-fibre reinforcement, bulk fill
Cement kiln dust Cement industry Marble dust Waste tyres Glass waste Marble industry Glass industry
Used engine oil Automobile industry Air entraining of concrete Automobile industry Rubber modfied bitumen, aggregate processing Bulk-fill, aggregates in bituminous mix and tile Bulk-fill, aggregates in bituminous mix
Table 3. Suitability of using industrial waste products in road construction (TFHRC 2004; Hamad et al., 2003; Hawk ins et al., 2003; Nunes et al. 1996; Sherwood 1995; Javed et al., 1994) Material Fly ash Advantages Disadvantages of
Lightweight, used as binder in Lack of homogeneity, presence stabilized base/ sub-base due to sulphates, slow strength development pozzolanic properties Higher skid resistance Light weight ( phosphorus slag)
Unsuitable for concrete and fill work beneath slabs. May show inconsistent properties
slag Construction More strength, can be used as May show inconsistent properties and aggregates granular base demolition waste Blast slag furnace Used in production of cement, Ground water pollution due to leachate granular fill formation, used as unbound aggregates Combustion of unburnt coal, sulphate attack in case of concrete roads Burning of combustible materials
Substitute for fine aggregate in Presence of heavy metals in non ferrous bituminous mixes foundry origin, less affinity to bitumen Some are pozzolanic in nature Presence of poisonous materials (e.g., cyanide from gold extraction)
Cement kiln dust Hardens when exposed to Corrosion of metals (used in concrete moisture, can be used in soil roads) in contact because of significant stabilization alkali percentage Used engine oil Good air entertainer, can be used Requires well organized in collection system concrete works Rubber tires Enhances fatigue life used oil
Requires special techniques for fine grinding and mixing with bitumen, sometimes segregation occurs
The incenerated municipal soild waste (MSW), after further processing, can be used as fines in bituminous mixes. Processing is done to remove ferrous and nonferrous metals and to achieve the required particle size gradation. Due to the presence of larger fraction of fines, MSW ash is primarily used as fine aggregate. It is also used as a fill material in road construction. The ash can also be stabilized with portland cement or lime to produce stabilized base/sub-base material (TFHRC 2004). For conventional road materials, a number of tests are conducted and their acceptability is decided based on the test results and the specifications. This ensures the desirable level of performance of the chosen material, in terms of its permeability, volume stability, strength, hardness, toughness, fatigue, durability, shape,viscosity, specific gravity, purity, safety, temperature susceptibility etc., whichever are applicable.
There are a large number tests suggested by various guidelines/ specifications. Figure6 presents a suggested flow chart to evaluate the suitability of industrial waste for potential usage in highway construction.
Evaluation industrial waste for suitability in highway construction Health and safety considerations should be given due importance handing industrial waste materials ( Mroueh and Wahlstrm 2002, Nunes et al. 1996). Other alternative materials Steel slag aggregate is a good example of synthetic aggregates obtained from byproducts of industrial processes. It has good binding properties with bitumen due to its high calcium oxide content (NatSteel 1993). The angular shape of the aggregates helps to form strong interlocking structure. Road paving with steel slag aggregate show
good skid resistance mechanical strength able to withstand heavy traffic and surface wearing.
Also, many industrial and other waste products like fly-ash, cement kiln dust, incenerated refuse etc. have been successfully used to produce synthetic aggregates. Mixing bitumen with rubber (natural or crumb form) sometimes poses difficulty. As an alternative approach, tiny crumb rubber pieces can be mixed with aggregates - known as dry-process. Research shows improved fatigue performance for this kind of materials (Sibal et al. 2000), also, this process does not require any modification to the existing batch mixing plant. INNOVATIVE APPLICATIONS Innovative applications may be construction method based or design principle based. Some of the relevant issues are discussed in the following. 1.3.1 Construction Method Based A mixture of aggregate and binding material, at varied proportions, constitute various specifications for road construction, for example, bituminous concrete, built-up spray grout, wet mix macadam, lean cement concrete etc. Discussion on all these standard specifications have been skipped here, rather, some specific mixes and their construction methods are discussed.
Emulsified bituminous mix Cold emulsified bituminous mix (EBM) is gaining more and more acceptance for its environmental friendliness and less hazardous construction process. A relative comparison between the EBM and hot bituminous mix (HBM) has been presented in Table 4. It may be noted that though the rate of strength gain in EBM is slower (refer Figure 7), the final strength of EBM is comparable to that of HBM. Table 4. Comparison of hot bituminous mix (HBM) and emulsified bituminous mix (EBM) Property Heating HBM Strong heating required, hardening occurs EBM oxidativeNo heating required, oxidative hardening High with high ambient All weather (wet surfaces, rainy seasons, cold) so no
Relatively difficult construction than Relatively easy construction EBM Relatively higher requirement Relatively lower requirement Inherent anti-stripping agents More costly fire and Free from such hazards
Uniqueness Modifiers needed Economy Safety Less costly Hazards from fuming, environmental pollution
Foamed bituminous mix (FBM) is a foamed mixture of air, water and bitumen. It is produced by injecting very small quantity of water into the hot bitumen, resulting in spontaneous foaming and temporary alteration of the physical properties of the bitumen. Figure-8 represents schematically the manufacture of FBM. Although the foamed bitumen technology was developed more than forty years ago, it is now gaining popularity owing to its good performance, ease of construction and compatibility with a wide range of aggregate types (Transportek 1998). Usage of FBM results in reduction in binder content and transportation costs, as it requires less binder and water than other types of cold mixing methods. FBM can be compacted immediately and can carry traffic almost immediately after compaction is completed ( Jenkins et al., 2003 ). The strength characteristics of FBMs are highly moisture dependent. This is because of the relatively low binder content and high void content of foamed bituminous mixes. FBMs are not as temperature susceptible as HBM. Since larger aggregates are not coated with binder, the friction between the aggregates is maintained at higher temperatures. Foamed bitumen can achieve stiffness comparable to those of cement-treated materials, with the added advantages of flexibility and fatigue resistance (Ramanujam and Kendall, 1999). FBMs usually lack resistance to abrasion and raveling and are not suitable for wearing/friction course applications.
Schematic presentation of FBM manufacture ( Romanoschi 2003 ) Some specific situations where use of foamed bitumen technology can be considered are: A pavement which has been repeatedly patched to the extent that pavement repairs are no longer cost effective. A weak granular base overlies a reasonably strong subgrade. Granular base too thin to consider using cementitious binders. Can be effectively used in desert road stabilization etc. (Jenkins et al., 2003). Relatively high cost, requirement of specific equipment for mix production, sensitivity to aggregate grading and stripping risk are some of the disadvantages with the foamed bituminous mix ( Jenkins et al., 2003 ). Fiber reinforced bituminous mix Addition of various kinds of fibers to the binder and aggregates during mix preparation process results in fiber reinforced bituminous mix (FRBM). Fibers are generally blended with bitumen binder before mixing it with the aggregates to achieve complete coating and even distribution throughout the mix. Research shows that FRBMs develop good resistance to aging, fatigue cracking, moisture damage, bleeding, reflection cracking etc. (Serfass and Samanos, 1996; Maurer et al., 1989). Ultra-thin whitetopping
Overlaying technique of pavement rehabilitation is well known and widely practiced. However, ultra thin whitetopping (UTW) of concrete over existing bituminous pavement is a relatively new concept. UTW can be designed for low-speed, low volume traffic areas such as street intersections, aviation taxiways and runways, bus stops and tollbooths. In this technique, a thin layer of high-strength, fiber-reinforced concrete is placed over a clean, milled surface of distressed bituminous concrete pavement to achieve a full or partial bonding. Bonding makes the two layers behave as a monolithic unit and share the load. Due to bonding, the neutral axis in concrete shifts from the middle of concrete layer towards its bottom. This results in a lowering of stresses at the bottom of concrete layer. Thick composite section behaviour causes the corner stresses to decrease. On the other hand, downward shifting of neutral axis may cause critical load location to shift from edges to corners thus increasing the corner stresses. Short joint spacing is used to decrease the slab area that can curl or warp thus minimizing the corresponding stresses (MTTP 2004). A schematic diagram of UTW have been presented in Figure-9.
Figure -9 Flexible composite pavement using UTW UTW is an excellent resurfacing option for deteriorated bituminous pavements which otherwise require frequent repair or overlays. Following are some of the advantages of a UTW system (CAC 2004, Murison 2002): It is beneficial for repair of roads and intersections having problems of rutting, cracks, and poor drainage. It provides improved skid resistance. Its light colour reflects more light than bituminous pavement. Its heat-reflecting property can help to lower the average city temperature. It is less costly to maintain, than conventional flexible pavements, and does not require frequent resurfacing and repairs. The UTW concrete resists bitumen aging. The UTW concrete prevents degradation of bituminous surface due to fuel spills. It causes minimal traffic disruption due to faster construction and repair procedure. Its small panels are ideal for utility maintenance. Bituminous recycling
In recycling method, bitumen and aggregates are separated out (partly or fully) and used again. The specific benefits of recycling of bituminous pavement can be summarized as:
Conservation of energy and construction material. Prevention of undesirable rise in height of finished surface and preservation of the existing road geometrics. Reuse of deteriorated road materials which in turn solves the disposal problem. Solution to the problem of scarcity of good quality material. Preservation of the environment. Reduction in susceptibility to reflection cracking.
Bitumen ages due to oxidation with atmospheric oxygen as a result of which resins get converted into asphaltenes (Petersen, 1984). By this process bitumen loses its ductility and becomes more brittle. Recycling is based on the fact that bitumen obtained from old deteriorated bituminous pavement, may still has its residual properties and recycling helps in restoring those residual properties of the bitumen. To judge the suitability for use as a recycled material, aggregates are tested for their gradation and bitumen is tested for its engineering properties. The optimum quantity of reclaimed material to be mixed with fresh material is generally determined from mix design process. Fresh thin (soft grade) bitumen having low viscosity can be used to replenish the aged bitumen. Rejuvenators (like road oils and flux oils) are sometimes added for improvement in properties of reclaimed bitumen There are four major technologies exist for bituminous pavement recycling (NCHRP452). They are (i) hot mix recycling Here recycled asphalt pavement ( RAP) is combined with fresh aggregate and bituminous binder or recycling agent in a hot mix plant. Mix is transported to paving site, placed, and compacted. (ii) cold in-place recycling In this the existing pavement is milled up to a depth of 75 to 100mm, RAP, if necessary and recycling agent in emulsion form is introduced, then compacted. (iii) hot in-place recycling In hot in-place recycling method the existing asphalt surface is heated, scarified to a depth from 20 to 40 mm, scarified material combined with aggregate and/or bituminous binder and/or recycling agent and compacted. New overlay may or may not be provided. (iv) full depth reclamation Here all the bituminous layers and predetermined thickness of underlying material is pulverized, stabilized with additives, and compacted. A surface course is applied over it. Semi flexible grouted macadam
Grouted Macadam consists of a single sized porous bituminous layer whose voids can be filled with the selected fluid grout or cementitious slurry.
The porous bituminous skeleton is designed to achieve a high void content while maintaining a thick bitumen coating on the aggregate particles (Zoorob et al., 2002). Grouted macadam gives advantages of both flexible and rigid pavements namely, Flexibility and absence of joints by use of bitumen, High static bearing capacity and wear resistance (as for concrete) by use of cementitious grout.
Micro-surfacing
Micro-surfacing is a fast and economical surface treatment technique used for preventive maintenance of bituminous and cement concrete pavements. Polymer modified emulsion, cold blended with fine graded aggregates, mineral fillers, additives and water, gives the high performing micro-surfacing mixture. Micro-surfacing is generally used to restore the top wearing surface of pavement as a maintenance measure, thereby extending the pavement life. Its thickness may be varied to achieve desired objective(s) such as rut-filling, skid resistance improvement, surface sealing, surface texturing, noise reduction, repairing abraded wheel path channels etc (ODOT 2004; Miller 2004).
Design Principle Based This section discusses about the design principle based innovative applications of road materials. Discussion has been divided into two parts viz.,
Optimum pavement design thickness In Mechanistic-Empirical pavement design, generally sustainability of a pavement structure against fatigue and rutting failures is considered, for which the critical responses are: (a) the tensile strain at the bottom fibre of bituminous layer and (b) the vertical strain at the top of the subgrade. A number of design thickness combinations of bituminous and granular layers are possible which satisfies the above mentioned requirement. Standard design charts developed by various organizations (Shell 1978; Austroads 1992; Asphalt Institute 1981; IRC:37-2001) are available; these design charts generally provide thickness composition of bituminous and granular layers, depending upon other input parameters viz. temperature, traffic, design life, subgrade strength, material type etc. A designer can choose any suitable granular layer thickness, and, corresponding thickness of bituminous layer can be read from these charts.
Figure 10. Typical pavement design chart POINT A - Safe from rutting but oversafe from fatigue considerations. POINT B- Safe from rutting but unsafe from fatigue considerations. POINT C- Safe from fatigue but insafe from rutting considerations. POINT D- Safe from fatigue but oversafe from rutting considerations. POINT E - Unsafe from both rutting and fatigue considerations. Point F- Oversafe from both rutting and fatigue considerations POINT O- Just safe from both rutting and fatigue considerations. Figure 10 illustrates a typical design chart. The design chart consists of two curves: fatigue curve and rutting curve. The fatigue curve shown as COD in Figure 10 represents the points, which are just safe from fatigue consideration. Similarly, the rutting curve shown as AOB in Figure 10 represents those points which are just safe from rutting consideration. Figure 10 shows various points like A, B, C, D, etc. They are safe, oversafe or unsafe from fatigue or rutting considerations. The reader can point the cursor on the respective points to know about their status. In the design chart the fatigue curve and the rutting curve intersects at a point (point O in this case) that may be called as structurally balanced design point (Narasimham, et al., 2001). Thickness of pavement layers chosen according to this point will result in a pavement deign which would fail due to fatigue and rutting simultaneously. Similarly, there could be cost optimal point, where bituminous and granular layer thicknesses are selected such that the total cost of materials used is minimized, without compromising with the structural adequacy of the pavement. The cost optimal point may or may not coincide with the structurally optimal point (Narasimham et al. 2001, Das 2004 ). Perpetual pavement
A perpetual bituminous pavement may be defined as a pavement designed and built to last longer than fifty years without requiring major structural rehabilitation or reconstruction (APA101 2001). This pavement may only require periodic replacement of top wearing surface and recycling of old pavement material (TRL 2001; AA-2 2001). The concept of full depth bituminous pavement is in vogue from 1980s in USA. Nunn and his associates of Transport Research Laboratory, UK found (Nunn et al., 1997) that thick bituminous pavements tend to show long lasting performance and may require only minor surface repairs. California Department of Transportation in collaboration with University of California, Berkeley (Monismith et al., 2001) first implemented concept of perpetual pavement in a rehabilitation planning project. In full depth bituminous pavement, the thickness is so designed that the fatigue and rutting strains developed are below the permissible limit (MS-1 1999 ). If the thickness is chosen to be sufficiently large so that the fatigue strain is close to the endurance limit, then the fatigue life becomes very long, and the pavement may be said to have attended 'perpetual life'. A perpetual pavement, in general, is made up of the following layers:
The top wearing surface is designed in such a way that it is water-tight as well as removable and hence replaceable. Stone Matrix Asphalt (SMA) or Open Graded Friction Course (OGFC) are recommended. They also produce less noise due to tyre-pavement interaction. The intermediate layer is constituted with good quality aggregates and designed to be strongly resistive to rutting. The bottom part is made resistant to fatigue cracking by making it rich in bitumen and choosing a gradation that has less voids.
Figure 11 Layer composition of a perpetual pavement. Figure-11 schematically represents the layer composition of a typical perpetual pavement. A perpetual pavement is a full depth bituminous pavement in most of the cases. The principles based on which it is designed (mix design and structural thickness design) are the following:
The pavement layers are chosen in such a way that they are rut resistive. The pavement is chosen to be adequately thick such that the vertical subgrade strain is low. Since subgrade contributes to the major part of rutting, low vertical subgrade strain would cause low level of rutting. The wearing surface should be adequately water-proof. The surface should be so designed that it can be repaired or recycled and the whole pavement will not require any major reconstruction (AA-2 2001). The thickness of the bituminous layer is chosen in such a way that the horizontal tensile strain (t) developed is less than the endurance limit (refer Figure-12) of the bituminous mix, hence its laboratory fatigue life (N) becomes infinity (AA22001, Nunn et al. 1997). It is justifiable to design the pavement as 'bottom rich' (refer to next section), which shifts the endurance limit to higher level. The temperature gradient tends to be steeper towards the surface of the pavement (TRL 2001, Newcomb 2001) as shown schematically in Figure-12. Therefore the bituminous mixes with temperature susceptible binder should be avoided as surface course. Use of modified binder could be helpful in this regard.
Figure 12
Rich bottom bituminous pavement Increased binder content above the optimum content can appreciably enhance the fatigue life. Higher bitumen content increases the thickness of the binder film between aggregates resulting in lower stress in the binder film, and thus the fatigue life is improved (Sousa et al., 1998; Harvey et al., 1996). However, with increased amount of binder content, the bituminous mix tends to be softer and thereby its stiffness modulus value may fall. A mix designer's objective would be to achieve both high stiffness and high fatigue life. This mutually contradictory situation can be handled by using a bituminous pavement layer where it is made richer in binder content towards the bottom layer(s). Since fatigue cracks start from bottom of bituminous layer, higher bitumen content helps to give greater restraint against fatigue cracking.
This concept has been termed as 'rich bottom pavement' (Monismith et al., 2001; Harvey et al., 1997; Harvey and Tsai 1996).
Figures 13 and 14 provide two such options of achieving this condition. In Figure-13, quantity of bitumen is used more towards the bottom of the layer. In Figure-14, two different bituminous mixes are used in two layers. Out of three possible alternatives, alternative-II turns out to be the best alternative.
Inverted pavement system, or inverted base, is a high depth pavement whose supporting layers are thicker and stiffer than top layers. The system consists of a thin bituminous concrete (BC) layer provided on top of a graded aggregate base (GAB) layer. A portland cement-treated stiff base layer is provided at the bottom. This arrangement causes the critical stress/strain plane to be located at the interface of the BC and GAB layers. Thus only the top portion of the inverted pavement structure absorbs the traffic loads as compared to conventional design where thick sections are required for load distribution. Research by South African Roads Board (SARB 2004) and Georgia Department of Transportation, has shown that an inverted base provides enough structural performance to support traffic loadings up to 100 million Equivalent Single-Axle Loads (ESAL s) with a maximum two inch bitumen riding course (Halsted, 2002). According to SARB, this type of system proves to be more cost effective for construction of long lasting pavements.
The cemented bases are derived from aggregates mixed with some binding material. Since it is bounded layer, it also has some fatigue life. Thus, unlike the unbound granular base, the cemented base layer contributes to some fatigue life, which may give rise to comparative reduction of design thickness of bituminous layer (Das and Pandey 1998). The stiffness modulus of cemented layer is generally found to be much higher than granular sub-base; however, due to shrinkage cracks, the stiffness modulus falls rapidly. This change in stiffness values at different stages of the design life has been schematically shown in Figure-15(a) and Figure-15(b) presents a typical design chart for design of bituminous pavement with cemented base made up of limesoil mixture.
A typical design chart of bituminous pavement with cemented base (LS lime soil) Mix design considerations Non-standard gradation The fatigue life of the mix can be increased by increasing the bitumen content.
But, Voids in Mineral Aggregates (VMA), being fixed for a given gradation and compaction level, increase in bitumen content will result in less Air Voids (VA), which is undesirable for a mix.
However one can deviate from the specified gradation in order to come up with a new gradation, which possibly can give rise to better fatigue performance, yet without compromising with the VMA and Marshall-stability requirements. Stone matrix asphalt Stone matrix asphalt (SMA) is a gap-graded bituminous mix with high percentage of coarse aggregates with high bitumen content. Gap gradation aims at maximizing stoneto-stone contact. This gives a structurally strong mix due to efficient load distribution through the stone-matrix skeleton. The drawback of this method is the absence of medium sized aggregates due to gap gradation. This may arise possibility of drain-down of low-grade penetration bitumen through the stone matrix . To check this possibility, modifiers, such as cellulose fibers, are used to hold the bitumen in place (Better Roads 2003; GDOT 1995; Decoene et al., 1990). Porous pavement
Porous pavement is a special type of pavement which allows surface water to pass through it, thereby keeping the road surface water-free, as well as providing drainage outlet to storm water. Porous pavement may be effectively used in light traffic areas like parking areas, airport taxiway and runway shoulders, footpaths, playgrounds etc. provided that the subsoil drainage, groundwater level and topography of the area is suitable (Michele, 2003; USEPA 1999; DEQ 1992). Pavement structure consists of a top porous bituminous layer placed over a filter layer below which a highly permeable open-graded stone layer (known as reservoir course). A geotextile layer is placed at the bottom to screen off fine soil particles. Porous bituminous layer consists of gap-graded aggregates (lower percentage fines), held together by a fiber-bitumen blend, giving a matrix structure which allows movement of water through its fine voids. Besides load bearing, the reservoir course stores the runoff water (in the void spaces in aggregate layers) until it can infiltrate into the soil beneath. Porous pavement has been found (RPL 2001) to be quite effective in reducing noise levels, splash and spray during rains, and aquaplaning tendency thereby improving the wet skid resistance.
CLOSING REMARKS Certain standard methods are followed for road design and construction. They are modified from time-to-time to match with the technological advancements. Certain modifications in the mix design or structural design can give rise to substantial economy in terms of the longevity of the pavement or the cost of the material concerned. Recapitulations Cement is manufactured by heating a mixture of limestone, iron ore, gypsum, clay and other ingredients. Cement concrete is a mixture of coarse aggregates, fine aggregates, cement and water, in suitable proportions. Through mix design, suitable proportions of the ingredients of concrete are estimated considering strength, workability, durability and economics. Workability test and aircontent test are the tests generally conducted on fresh concrete. Compressive strength, tensile strength, modulus of rupture, elastic modulus, Poisson's ratio, creep and shrinkage, durability, thermal expansion coefficient etc are the tests conducted on hardened concrete. Various modifications and inovatory applications of pavement materials and pavement design brings in better performance and economy.
The objective is this lecture is to introduce the basic principle of concrete pavement design, and discuss briefly the provisions prescribed in various design guidelines.
Introduction The concrete pavement can be of three types, jointed plain concrete pavement (JPCP), jointed reinforced concrete pavement (JRCP) and continuously reinforced concrete pavement (CRCP). There could be another variety - pre-stressed concrete pavement, which has specialized application, such as airport runway pavement. The relative advantages and disadvantages are mentioned in Table-5 (Swanlund and Vanikar 2002).
Table -5 Advantages and disadvantages of various types of concrete pavements (Swanlund and Vanikar 2002) Pavement type Jointed plain pavement (JPCP) Advantages Disadvantages
concrete Reliable design and can be Joint maintenance is costly used at all locations. and may impair future performance after rehabilitation.
Jointed reinforced concrete Fewer joints than JPCP. - dopavement (JRCP) Damage due to pavement cracking is less severe than JPCP, because the reinforcement holds the
cracked portion in position. Continuously reinforced No joints, hence smooth ride High initial cost, complex to concrete pavement (CRCP) and long service life. construct. Maintenance is costly and difficult.
JPCP is the most popularly used pavement. The present lecture discusses various design approaches for design of JPCP pavements. The design approaches for JRCP and CRCP have been discussed elsewhere . Design parameters The design parameters can be primarily divided into three categories, material, traffic and environmental parameters and are discussed in the following. Material parameters The elastic modulus, Poisson's ratio, fatigue life, modulus of rupture etc. are the engineering parameters used for the structural design of the concrete pavement. The input parameters are either found out experimentally or estimated from various recommendations provided in the design guidelines (ACI 2001, ASTM 2003, IS-456). Statistically suitable design value is to be adopted if there are variations in the input parameters (Chakroborty and Das 2003). The modulus of subgrade reaction (k) is generally used for characterization of subgrade strength. k in idealized model represents the spring constant of a dense liquid foundation. The k value is obtained by performing plate load test on the subgrade. The IRC:58 (2002) provides a table with suggested k values when CBR values of subgrade are known. Concrete slab, in general, are constructed on bound, stabilized, or unbound sub-base layer, and not directly on subgrade. The sub-base layer provided below the concrete pavement serves certain purposes (Austroads 2004), such as, 1. It provides uniform support to the concrete slab. 2. It limits pumping at joints and slab edges. 3. It controls shrinkage of concrete slab or swelling of subgrade soil The effective k value that includes the sub-base layer and the subgrade should be used for design of thickness of the concrete slab. This effective k value can be obtained by performing plate load test on the constructed sub-base, or by computational means. Various guidelines (IRC:58 2002, AASHTO 1993, PCA 1984, Austroads 2004) suggest suitable values of effective k when the type and thickness of the sub-base layer is known. Traffic parameters The concept of axle load, wheel configuration, traffic volume, lane traffic, traffic growth rate, and calculation of axle load repetitions have already been introduced in the other web course on Transportation Engineering-I. Interestingly, vehicle damage factor (VDF) (also known as Truck Factor (MS-1 1999)) is not considered in concrete pavement design. This is due to the fact that typically for concrete pavement design considers the individual damages created by individual axle load groups are considered. To take care of the variability of traffic axle load measurement, and unpredicted heavy truck load, a factor called the load safety factor
(LSF) is generally used in concrete pavement design and is multiplied with the total equivalent standard axle load repetitions (Austroads 2004, IRC:58 2002). Environmental parameters The temperature and the subgrade moisture are the environmental parameters that affect the concrete pavement design. The temperature variation and the temperature differential induces temperature stresses to the concrete pavement. Temperature issues in concrete pavement has been discussed in the lecture 'analysis of concrete pavement' . The subgrade moisture level affects the subgrade strength and also it induces curling stresses due to moisture gradient. Design life Design life is the number of years (or number of standard axle repetitions) for which the pavement is being designed. A pavement is expected to serve satisfactorily within the design life. For concrete pavements 20 to 40 years may be assumed as the design life (PCA 1984, IRC:58 2002). Basic design principle Though different approaches for concrete pavement design are suggested in various guidelines, the design principles tend to remain similar across different guidelines, for example, PCA (1994), Austroads (2004), NCHRP (2004), IRC:58 (2002) etc., except the AASHTO (1993) provisions, which is based on empirical approach. The basic steps involved in the design of concrete pavement method can be summarized as follows:
The developed stresses due to load for a trial thickness of the concrete slab are calculated for various loading configuration and the critical one is chosen. The axle loads are generally divided into different axle load groups and the load stresses are calculated individually. The ratio between the load stress and Modulus of Rupture (MOR) is known as stress ratio . The stress ratio determines how many repetitions the pavement can sustain (i.e. allowable traffic repetitions ) for the individual axle load group. If the stress ratio is 0.55 or lower, it can withstand virtually infinite number of traffic repetitions (PCA 1984). The ratio between the allowable repetitions to the expected traffic repetitions is the damage fraction. The calculation process is repeated for various axle loads (sometimes, for various seasons, or various timings of the day), and the sum of individual damage fractions ( cumulative fatigue damage ) should be less than equal to one for pavement design being safe. If found unsafe, the trial thickness is changed and the design process is repeated. The design process may also include considerations for temperature stress, moisture stresses, and erosion distress.
Various design approaches A brief discussion on concrete pavement design approaches suggested by various design practices, viz. PCA method (1984), Austroads method (2004), AASHTO method (1993), NCHRP mechanistic-empirical method (2004) and the Indian Roads Congress (IRC) method (2002) are placed in the following: Portland Cement Association (PCA) Method
The PCA method is based on Westergaard, Picket and Ray's work and further theoretical analysis by Finite Element Method (Huang 1993). The data used to develop the PCA method is generated from various road tests, like, ASSHO road test, Arlington test (conducted by PCA), Bates test road, and Maryland road test (PCA 1984). The PCA design method is based on the following two considerations (PCA 1984):
The fatigue damage on the concrete slab, due to repetitive application of traffic load is estimated. The cumulative fatigue damage principle is used to estimate the design thickness of the slab. Edge stress between the mid-way of the transverse joint is taken as critical configuration. The stresses due to warping and curling (due to temperature and moisture gradient) are not considered in the fatigue analysis as per PCA recommendation, because, most of the time the stresses generated are subtractive to the load stress. The possibility of erosion of pavement materials placed below the concrete slab is evaluated. The rate at which the slab is deflected due to axle load is used as a criterion for erosion. Theoretically, it can be shown that a thin pavement with smaller deflection basin is subjected to faster rate of deflection, compared to thicker slab. Hence thin slab is more susceptible to erosion. In a similar way the cumulative erosion damage is calculated for individual axle load groups. If this value is greater than one, then the design needs to be revised.
Austroads method Austroads method has been adopted from PCA (1984) approach with modifications suited to Australian conditions (Austroads 2004). A bound mix or lean cement concrete is used as sub-base material. For a given CBR value of the subgrade and given thickness of cemented sub-base, the effective subgrade strength can be obtained from the chart provided. Figure-19 schematically shows such a chart. For different levels of traffic, Austroads (2004) suggests minimum values of the base thicknesses to be provided.
Figure-19 Schematic diagram of Austroads (2004) chart for estimation of effective subgrade strength As per Astroads (2004), the concrete pavement slab thickness, for a given expected traffic repetitions, is designed considering the (i) flexural fatigue of the concrete slab and the (ii) subgrade erosion arising out of repeated deflections. For both the considerations, equations are suggested to calculate the allowable traffic repetitions. If the allowable traffic is less than the expected traffic, the design is revised by increasing
the slab thickness. The concrete shoulders adopted are of integral type or structural type . AASHTO method The AASHTO method (1993) for design of concrete pavement has evolved from the AASHO road test (AASHO 1962). The AASHTO pavement design follows an empirical approach. Pavement performance in terms of present serviceability index ( PSI ), loss of serviceability, subgrade and sub-base strength, cumulative traffic, properties of concrete, joint load transfer efficiency, drainage condition, overall standard deviation and reliability are the input parameters considered in the pavement design. The PSI value of the fresh pavement is assumed as 4.5 and the pavement is deemed to have failed when the PSI value reaches 2.5. The resilient moduli of the subgrade and sub-base materials are determined in the laboratory simulating the seasonal moisture content and stress situation. Suggested values are also available for given moisture content, plasticity index etc. The composite modulus of subgrade reaction ( k) is estimated from modulus of subgrade reaction and elastic modulus of sub-base for various seasons and the depth of rigid foundation and the thickness of the sub-base. Design equation as well nomographs are available to estimate the slab thickness ( D) from these input parameters. NCHRP mechanistic-empirical (M-E) method The NCHRP (2004) has recently developed concrete pavement design procedure based on mechanistic-empirical (M-E) approach. This approach attempts to reduce the extent of empiricism prevalent in the existing AASHTO (1993) guidelines. This proposed NCHRP pavement design system is modular in nature, that is, the design approach can be modified by parts (as and when new knowledge is available) without disrupting the overall design procedure. This approach also can take care of various wheel-axle load configuration (Khanum et al. 2004). As per this approach trial thickness of the slab is first assumed, and the stress, strain and displacement values are obtained. From these values, the performance of the pavement in terms of distresses (such as faulting, cracking) and smoothness and predicted. If these predicted performance parameters does not satisfy the required performance for a given reliability, the design revised. The design approach includes a large data-base as input parameters, for example, average daily traffic, traffic growth rate, traffic composition, hourly weather data on air temperature, precipitation, wind speed, percentage sunshine, relative humidity, pavement material engineering parameters, ground water depth, infiltration, drainage, hydraulic conductivity, thermal conductivity, heat capacity etc. The temperature stress is considered in this method, but the temperature profile is linearized to enhance computational efficiency (NCHRP 2004).
Indian Roads Congress (IRC) method The Indian Roads Congress (IRC) guidelines, IRC:58 (2002), has adopted the Westergaard's equation to estimate load stress and Brdabury's equation to estimate temperature stress. The load stress is highest at the corner of the slab, lesser in edge and least in the interior. The order of variation of temperature stress is just the reverse of this. As per IRC:58 (2002), it is recommended that the design needs to be done for edge stress condition and subsequently check needs to be performed for corner stress condition so as to finalize the design. The following are the steps followed as per IRC:58 (2002) guideline for the design of concrete pavement:
The input parameters are obtained to formulate the design problem. The joint spacing and the slab dimensions are decided. If there is a bound sub-base layer over the subgrade, a suitable value of effective k is to be adopted. A trial thickness of the concrete slab is assumed. The edge stress is estimated for various axle loads from the given charts. Figure20 schematically shows such a chart. The cumulative fatigue damage principle for fatigue is applied to check the adequacy of the slab thickness. The sum of edge stress due to load for the highest axle load group and the temperature stress should be less than the MOR of concrete, otherwise the design is revised. The adequacy of corner stress is checked with respect to MOR value and accordingly the design is finalized. Westergaard's corner stress formula is for estimation of corner stress due to load, and the corner stress due to temperature is assumed to be zero.
Figure-20 Schematic diagram of IRC:58 chart for estimation of load stress at the edge (IRC:58 2002) Recapitulation
The fatigue and erosion damages are generally considered as primary structural factors which govern the concrete pavement design by mechanistic-empirical (ME) approach. Cumulative damage principle is employed to evaluate the expected level of damage. Faulting and smoothness are the other considerations in the design. The design guidelines such as PCA (1994), Austroads (2004), NCHRP (2004), IRC:58 (2002) are based on M-E approach, whereas, the AASHTO (1993) approach of pavement design is purely empirical. While considering the cumulative fatigue damage (in M-E approach), the load stresses are primarily considered. The stresses due to temperature and/or moisture gradient are ignored in PCA (1994) and Austroads (2002) guidelines, but are considered in IRC:58 (2002) and NCHRP (2004) guidelines.
Design of dowel bars Objective The objective of this lecture is to introduce the concept of design of dowel bars.
Introduction Dowel bars are placed at the transverse joints of concrete pavement and they take part in partial wheel load transfer from one slab to its adjacent slab. The dowel bars also allow axial thermal expansion and contraction of the concrete slab along the axis of the dowel. The bars are generally made up of mild-steel-round-bars of short length. Around half of the length of this bar is embedded in one of the concrete slabs and the remaining portion is bonded in the other adjacent slab. One end of the bar is kept free for the movement during expansion and contraction of the slab due to change in temperature. The performance of the dowel bar system drastically falls when voids are created between the dowel bars and the concrete slab (Porter 2001). Such a situation develops stress concentration and failure may take place rapidly. The void space may get filled with water and corrosion action to the dowel bars may get aggravated. Subsequently, the load transfer mechanism tends to fail and the differential settlement of the concrete slab may occur (Porter 2001). Differential settlement may further cause breakage of slabs due to impact loading of the vehicles. Thus, the design adequacy of dowel bar system and proper placement to the concrete pavement slab is an important consideration. Analysis of dowel bar Analysis Recall the formulation developed as Equation 13 , in the lecture on 'analysis of concrete pavement' , representing beam on elastic foundation. By putting appropriate boundary conditions to that equation, for a semi-infinite beam with a moment, Mo, and a point load, P, the following solution emerges for Winker's foundation (Porter 2001, Timoshenko and Lessels 1925 ): (25) where, y = the deflection along the x direction, , and is called as the relative stiffness of the beam on the elastic foundation, k = spring constant, E = elastic modulus of the beam and I = moment of inertia of the beam section. From Equation 1, the deflection of the dowel bar at the face of the joint, y o, can be obtained as (Friberg 1940): (26)
where, relative stiffness of the dowel bar resting on concrete (assumed as elastic foundation), kd = modulus of dowel support, b d = dowel bar width (i.e.
diameter), Ed = elastic modulus of the dowel bar, I d = moment of inertia of the dowel bar, Pd = load transferred through the dowel bar and z =joint width. Though the above equation assumes dowel bar to be semi-infinite, Porter (2001) showed that Equation (26) gives a reasonably good estimate of deflection obtained from more rigorous analysis. Thus, the bearing stress developed, bd, can be expressed as the product of modulus of dowel support ( kd) and the deflection at the face of joint ( yo) (Porter 2001), i.e. (27) For a successful dowel bar design, the value of bd needs to be kept lower than the allowable bearing stress of concrete, bd, specified as (ACI 1956, Porter 2001, IRC:58 2002): (28)
where, bd is expressed in MPa, fck = characteristic compressive strength of concrete in MPa and bd is the diameter (i.e. width) of the dowel bar in mm, Joint load transfer efficiency Ideally, the dowel bar system should be able to transfer the whole load applied to it. However, voids developed due to repetitive loading reduce the joint load transfer efficiency ( JLTE ) of the dowel bar. The JLTE (expressed in percentage) of a dowel bar can be defined as (Porter 2001, Ioannides and Korovesis 1992 )
where, Pa is the load applied to the dowel bar. The force transmitted being somewhat difficult to measure, joint load transfer efficiency is sometimes measured in terms ratio of deflections. If impact loading (generally, falling weight deflectometer ( FWD ) is used for this purpose) is applied near the transverse dowelled joint of a slab, the JLTE can be defined as the percentage of the deflection of the unloaded slab with reference to the deflection of the loaded slab. Figure 1 schematically shows two extreme situations as JLTE = 0% and JLTE = 100%. Various researches and documents (Porter 2001, Yoder and Witczak 1975, ACPA 1991 ) suggest that the load transfer efficiency varies between 35 to 50%.
Figue-21 Schematic diagram explaining the deflection based load transfer efficiency (Chakroborty and Das 2003) Distribution of load A part of the load applied is shared by the dowel bar system. Essentially, this load is not shared by only one dowel bar, rather, it is shared by a group of dowel bars. (These dowel bars are placed at some designed interval). Thus, it is important to know (i) how many dowel bars participate in load transfer, and (ii) how is this load shared across the various participating dowel bars. Fridberg (1940) suggested that a length of up to 1.8 radius of relative stiffness (refer to Equation (11) of the lecture 'analysis of concrete pavement' for definition) participate in the load transfer. It is also suggested that the load may be taken as linearly varying with maximum share taken by the dowel bar which is just vertically below the wheel. Tabatabaie et al. (Porter 2001) suggested that instead of taking the factor as 1.8 it should be taken as 1.0 . From design point of view, the wheel can be placed in two ways over the transverse edge of the slab, viz. case (i) the wheel at one edge or, case (ii) the wheel is at the middle. Obviously, the maximum load shared by a dowel bar in case (i) will be more than case (ii) . Hence case (i) would govern the design of dowel bar. Design of dowel bar By the term design of dowel bar , it is meant to estimate the dowel bar diameter and spacing. One of them can be assumed as known or fixed, and the other is estimated. Dowel bars are generally designed from two considerations: (i) Bearing stress approach, where the developed bearing stress is set equal to or less than the allowable bearing stress, and (ii) Relative deflection approach, where the relative deflection of the joints is not allowed to exceed some maximum specified value. Example design An example on design of dowel bar system from bearing stress consideration is presented in the following. Problem A design wheel load of 65kN is applied on to the concrete pavement slab, and 50% of
the load is assumed to be transmitted through the dowel bar system. Assume the characteristic compressive strength (fck) of concrete (used in the concrete pavement) is 40MPa, the transverse joint width ( z ) is 15 mm, radius of relative stiffness ( l) of the concrete slab as 950mm, Elastic modulus of steel (E) as 2.0 x 10 5 N/mm2 and modulus of dowel support (kd) as 415N/mm 3 . Design the dowel bar system. Solution Assume, 32mm diameter dowel bars are used. The allowable bearing stress ( ba ) can be obtained from Equation (28) as: = 29.23 MPa Assume, dowel bar spacing as 300 mm center-to-center. Thus, in 950 mm of effective length (using Tabatabaie et al. criterion) , 4 dowel bars can be placed. Assuming, the load carried by the dowel bar which is just below the wheel, is P d, one can write the load distribution equation of the dowel bar system as:
0.0238 per mm From Equation (27), the developed bearing stress in dowel bar
It is noted that < , thus the design is safe. Thus, dowel bar of 32 mm may be provided with 300mm center-to-center. Closing remarks The side of the dowel bar does not take any shear force, thus it is suggested that a dowel bar with elliptical cross-section will provide more effective area for distribution of stress (Porter 2001). This has been schematically explained in Figure 22.
Figure 22. Elliptical cross-section has more effective area for shear stress distribution (Porter 2001) While designing the thickness of the concrete pavement, generally the stresses at edge, corner and interior are considered. However, the stresses at the edge of the transverse joint also need to checked, considering the situation when dowel bars fails to perform load transfer ( Porter and Guinn 2002 ). Corrosion of the dowel bars, usage of fiber reinforced polymer dowel bars, estimation of modulus of dowel reaction, dowel-slab interaction and their fatigue (Porter and Guinn 2002) are some of the emerging issues which demand further research and understanding. Recapitulation
Dowel bars are placed at the transverse joints of concrete pavement and they take part in load transfer from one slab to the adjacent slab. The dowel bars also allow axial thermal expansion and contraction of the concrete slab along the axis of the dowel. Only a part of the wheel load is transmitted through the 'participating' dowel bars. The dowel bar which is just vertically below the wheel carries the maximum load. Loads carried by the other participating dowel bars decrease proportionately as these move away from the point of load application. The dowel bar system is designed from the bearing stress and deflection criteria.
Design of tie bars Objective Objective of this lecture is to introduce the design concepts of tie bar. Introduction Tie bars are used across the longitudinal joint of concrete pavement. The purpose of tie bar is to hold the concrete slabs together (refer Figure 23). Tie bars supposedly do not transfer any wheel load.
Figure 23: Schematic diagram showing location of tie bar in pavement cross-section Design principle The tie bars are designed to withstand tensile stresses, the maximum tensile force in the tie bar is made equal to the force required to overcome frictional force between the bottom of the adjoining pavement slab and the soil subgrade. Estimation of spacing and length of tie bar is explained in the following. Estimation of spacing of tie bar Since the purpose of the tie bar is to tie concrete slabs together, the area of steel per unit length of longitudinal joint is obtained by equating the total friction to the total tension developed in the tie bar system (as explained in Figure 24).
Figure 24: Estimation of area of steel for tie bar Thus, (29) where = co-efficient of friction between concrete slab and the sub-base, W = weight of the concrete slab per unit length (say per meter), st = allowable working stress in tension for the steel used as tie bar, A st= cross-sectional area of steel per unit length (say per meter). The weight of concrete slab per unit length can be written as, (30) where, w = weight of slab per unit volume (say, cubic meter), B = width of the slab, and h = height of the slab. Thus, from Equation (29) and (30), the area of steel per unit length, (31) Assuming suitable diameter of tie bar, the spacing of tie bar can be found out so as to the requirement of steel per unit length. Estimation of length of tie bar Consider a single tie bar. The tensile force developed in the tie bar should not exceed the bond strength between the tie bar and the concrete, otherwise it can be pulled out of concrete. Thus, considering one end of the tie bar,
(32) where, as = cross-sectional area of one tie-bar, P = perimeter of one tie bar, l = length of tie bar inside the concrete slab, Sb = allowable bond strength between the concrete and the tie bar. Thus, total length of the tie bar, can be written as, (33) where z = allowance due to inaccurate centering of the tie bar. Recapitulation
The spacing of tie bar is estimated from the area of the steel required per unit length of the longitudinal joint. The area of steel required is estimated by equating the maximum tensile force in the tie bar system with the force required to overcome frictional force between the bottom of the adjoining pavement slab and the soil subgrade. The length of the tie bar is estimated by equating the tensile force developed in a tie bar with the bond strength between the tie bar and the concrete slab.
Design of runway, taxiway and apron Objective Objective of this lecture is to introduce the basic concepts of airfield pavement design and to perceive the conceptual similarity (or differences) across various guidelines for pavement design. Introduction The airfield pavement includes runway, taxiway, shoulder and the apron. Some of the portions of these components are identified as critical areas , and some are non-critical areas . As a general guideline, the critical areas are those portions where the aircraft speed is low, or the aircraft is in rest. More thickness is required for the pavement in critical areas. In the non-critical areas, for example, the central portion of the runway, the air craft is partly air-borne; also due to larger lateral wander the stress repetitions at a particular spot is small (FAA 2006, PCA 1995). For high speed (i.e. small time of contact), the asphalt shows higher strength ( creep modulus) than slow speed movement. Thus, design thickness for non-critical areas turns out to be less than the critical areas. Same design chart is, generally, recommended for design of airfield pavements, and some factors are prescribed for adjustment of thicknesses for the individual components (i.e. runway, taxiway, shoulders and apron) of the airfield pavements.
Input parameters The various parameters used in pavement design have already been discussed in the section 'design parameters' of the lecture ' various design approaches' . Similarly, the
parameters involved in the airfield pavement thickness design process are the aircraft gear loading (gear configuration, wheel load, tyre pressure, lateral wander), load repetition, material properties (stiffness of individual layers, fatigue behaviour of bound layer(s)) and environmental factors (temperature, subrgade moisture). Bituminous pavement design methods Some of the methods for design of bituminous airfield pavement structures are (i) Corps of Engineers (CoE) method (ii) Federal Aviation Administration method (FAA) and (iii) The Asphalt Institute method and are discussed briefly in the following: Corps of Engineers (CoE) method This method was developed during the World War-II, when data on runway performance was collected and analysed for various airports (Horonjeff 1975). The total thickness above a particular layer is determined by the CBR value (in percentage) of the layer. Thus, the total thickness above the subgrade, sub-base, base etc. can be found out by knowing the CBR of the respective layers. This method is simple, but purely empirical. The total thickness, T, is estimated as:
here, is the load repetition factor, ESWL is the equivalent single wheel load of the gear load assembly (in pounds), A is the measured tyre contact area (in inch2). The values of is suggested for various gear assembly (single wheel, dual wheel, dual wheel tandem axle etc.) used for ESWL calculation, and number of load repetitions (Horonjeff 1975). From the known CBR values of the subgrade, sub-base, and base layer the individual thickness composition are obtained. Some theoretical basis of the CBR method was proposed later on, where the thickness of the pavement is so adjusted that the shear stress developed at a given level is equal to the allowable shear stress of that layer. U.S. Army Corps of Engineers, Naval Facilities Engineering command and Air Force Civil Engineer Support Agency have recently developed a combined manual for design of airfield pavement as Unified Facilities Criteria (UFC 2001) as a modified version of this design approach. Federal Aviation Administration (FAA) method The FAA uses modified CBR method for design bituminous pavement (FAA 2006). As per the FAA method, the subgrade soil is categorized based on the soil classification group, drainage and frost damage conditions. Design charts are available for design of pavements for gross aircraft weight of 30,000 lbs or more, for single, dual, and dual tandem aircrafts. It is generally assumed that 95% of the gross weight of the aircraft is carried by the main landing gear and 5% is carried by the nose gear (FAA 2006). A typical pavement design is presented in Figure 25. The design life is assumed generally assumed as 20 years.
Figure 25: A schematic bituminous pavement design chart as per FAA (2006) The following is the design procedure followed as per the FAA method (FAA 2006):
The subgrade CBR value, gross take-off weight, main gear weight and annual departure of the individual aircrafts are used as the input to the design. Different types of materials can be used as base/ sub-base, and accordingly some variations are observed in the design thickness. The individual design charts (for various categories of aircrafts) are used to find out the total pavement thickness values for different aircrafts. The aircraft whose corresponding data results in the maximum thickness is designated as the design aircraft . Having obtained the design aircraft, the annual departure data of the individual air-crafts are converted to equivalent annual departure of the design aircraft. This is done by two sequential steps, as follows: o The departures of different aircrafts which have different gear configurations are converted to equivalent departure of the design aircraft, by multiplying with suitable factors. Various factors are suggested by the FAA (2006) using which departure of any gear configuration can be converted to departure of another gear configuration. o The departures of individual aircrafts adjusted for the gear configuration of the design aircraft are further adjusted for the wheel loads. This is done by using the following empirical formula: (35)
where, Rd is the equivalent departure of the design aircraft, R a is the departure of any aircraft adjusted for the gear configuration, W d is the single wheel load of main gear of the design aircraft, and Wa is the single wheel load of main gear of any aircraft.
Having obtained the equivalent departures of the design aircraft adjusted for gear configuration and wheel load, all the departures are summed up. Pavement design is performed for this total equivalent departure with the help of the specific design chart applicable for the design aircraft . From the known CBR values of the subgrade, sub-base, and base layer, the individual thickness composition are obtained. The dashed line indicates the order of progression. The asphalt thicknesses are generally specified for the critical and non-critical areas.
To take care of the frost problem (freeze-thaw), first, the freezing index of the site is obtained. Predictive charts are developed to estimate the depth of frost penetration for various soil types and freezing index. Now, either of the following three approaches can be used for design purpose. o The thickness of the pavement can be made at least equal to the depth of frost penetration. o The frost susceptible subgrade soil may be replaced with non-frost susceptible soil. o Adequate pavement thickness may be provided assuming subgrade will have reduced strength during thawed condition.
A worksheet software for design of bituminous airfield pavement as per FAA can be downloaded from the FAA website The Asphalt Institute Method Asphalt Institute (MS-11 1987) recommends full depth asphalt pavement for air-field pavement (for gross aircraft weight higher than 60,000lbs). Mechanistic-empirical (M-E) method is used for design of pavement. As per the M-E design method, the two modes of structural failure are considered to be governing the pavement design. These are fatigue and rutting failure. The horizontal tensile strain below the bituminous layer and vertical strain on the subgrade are considered as causative factor for fatigue and rutting respectively. The field calibrated fatigue/ rutting equations are obtained by relating the initial critical strains and the number of repetitions it take for the failure due to fatigue or rutting. The allowable fatigue and rutting strains, for a design traffic, are obtained from the field calibrated fatigue and rutting equations. In M-E approach the thicknesses of the pavement layers (asphalt surfacing and base/ sub-base) are adjusted in such a way that the developed fatigue and rutting strains are comparable to the allowable strains. Concrete pavement design methods Some of the methods for design of concrete airfield pavement structures are (i) Corps of Engineers (CoE) method (ii) Federal Aviation Administration method (FAA) and (iii) Portland Cement Association (PCA) method and are discussed briefly in the following: US Army Corps of Engineers (CoE) method The CoE method uses fatigue equation developed from full scale load tests conducted during 60's and 70's (Smith et al. 2002). The current Unified Facilities Criteria (UFC 2002) considers fatigue of pavement slab due to edge stress. The edge stress is calculated as per Westergaard's analysis and assumes that load transfer joints reduces the stress by 25% ( Thuma and Lafrenz 2003 ). The method allows reduction of slab thickness if the subgrade strength is high. Federal Aviation Administration (FAA) method The standard design curve assumes aircraft wheels are either placed parallel to or perpendicular to pavement joint. Westergaard analysis of edge stress calculation is employed. It is assumed that 25% of the edge stress is carried by the joints at the edges (FAA 2006). The following are the considerations for design of concrete pavement as per FAA (2006):
The modulus of subgrade reaction is estimated by performing plate load test on the subgrade. If the subgrade is multi-layered, or sub-base (granular or stabilized) is put over the subgrade, the effective modulus of subgrade reaction needs to be determined. Charts are provided, similar to the Figure 19, presented in the lecture on 'various design approaches'. Same concept, as discussed above for the design of bituminous pavements, is applied to find out the design aircraft . Design charts are available, which involves the input parameters as, concrete flexural strength, effective modulus of subgrade reaction, gross weight of the design aircraft and the annual departure of the design aircraft. The cumulative fatigue damage principle is employed to arrive at the design thickness of the slab. A typical concrete pavement design chart as per FAA (2006) is schematically presented in Figure 26. The dashed line indicates the order of progression. Similar strategies are adopted, as discussed for design of bituminous pavement by FAA method, to take care of the frost penetration problem. Standard schemes are recommended (FAA 2006) for providing contraction, construction and expansion joints to the pavement. Provisions are also provided for jointed reinforced concrete pavement (JRCP) and continuously reinforced concrete pavement (CRCP). A worksheet software for design of concrete airfield pavement as per FAA can be downloaded from the FAA website:
Figure 26: A schematic concrete pavement design chart as per FAA (2006) Portland Cement Association (PCA) method The fatigue equation of PCA method is developed from laboratory testing of fatigue behaviour of concrete beams (Smith et al. 2002, Packard 1995). The PCA method (Packard 1995) requires the (i) concrete properties (ii) effective strength of subgrade (or subgrade-sub-base combination) (iii) type of aircraft, load, and approximate frequency and (iv) type of pavement being designed, such as runway, taxiway, apron hangar etc. Use of 90 days strength test results of concrete is generally recommended for design purpose. The design procedure can be summarized as follows:
The modulus of subgrade reaction is estimated by performing plate load test on the subgrade. If the subgrade is multi-layered, or sub-base (granular or stabilized) is put over the subgrade, the effective modulus of subgrade reaction
needs to be determined. Charts are provided, similar to the Figure 19, presented in the lecture on 'various design approaches'. From careful estimate of the present and future traffic, an appropriate safety factor is estimated for each of the aircraft types. The working stress is obtained by dividing the modulus of rupture of concrete with the safety factor chosen for individual aircrafts. Suggested values of safety factor range from 1.7 to 2.0 for critical areas and 1.4 to 1.7 for non-critical areas. From the pavement design chart, the design thickness slab thickness is estimated for a particular type of aircraft. Figure 27 presents a typical pavement design chart as PCA (1995). The cumulative fatigue damage principle is employed to arrive at the design thickness of the slab. The dashed line indicates the order of progression. This process is repeated for various aircrafts and the highest value of the thickness is chosen as the design thickness. The stresses created due to less critical aircrafts are also checked (by using the design charts in the reverse direction), and if the values are less than some specified value (350 psi), it is assumed that these do not add to the fatigue of the pavement slab.
Figure 27: A schematic concrete pavement design chart as per PCA (1995) Recapitulation
Some of the methods for design of bituminous airfield pavement structures are (i) Corps of Engineers (CoE) method, (ii) Federal Aviation Administration method (FAA), (iii) The Asphalt Institute method etc. The CoE and FAA methods are primarily empirical, and based on CBR value and thickness relationship. The Asphalt Institute method is mechanistic empirical, and considers fatigue and rutting failure of the pavement. Some of the methods for design of concrete airfield pavement structures are (i) Corps of Engineers (CoE) method, (ii) Federal Aviation Administration method (FAA), (iii) Portland Cement Association (PCA) method etc. These approaches estimates the flexural stresses in concrete slab which is modeled to be resting on elastic springs. The cumulative fatigue damage principle is employed to arrive at the design thickness of the slab.
The objective of this lecture is to present a brief introduction to reinforced concrete pavements. Introduction Reinforced concrete pavement could mainly be of three types: jointed reinforced concrete pavement (JRCP), continuously reinforced concrete pavement (CRCP) and pre-stressed concrete pavement. Pre-stressed concrete pavement has a very specialized application, and sometimes used for airport runway pavements. The JRCP and CRCP pavements are schematically shown in Figure-28.
Figure-28: Schematic diagram of JRCP and CRCP pavement (ACPA 2005) Design principle Welded wire fabric or deformed bars or steel fibres are used for reinforcement purpose (Swanlund and Vanikar 2002, IRC:SP:46 1997). Reinforcement in the concrete slabs is mainly provided for counteracting the shrinkage (and associated cracks) due to temperature and moisture changes (PCA 1995). It does not allow crack to become wider (Austroads 2004), and thereby ingress of water and grit is prevented (IRC:101 1991). Since, the reinforcement provided in concrete pavement is not for contributing towards flexural strength, therefore, its position in the concrete slab is not important. Generally reinforcement is placed 50 mm below the surface (IRC:58 1988). The basic slab thickness is designed as per the plain cement concrete pavement design procedure 1 , however, due to provision of steel reinforcement, there is an effective increase of the slab thickness due to provision of steel reinforcement (IRC:101 1991). The requirement of steel reinforcement can be calculated on the basis of maximum force that can overcome the frictional force between the concrete slab and the layer just below it. The shrinkage stress is maximum at the middle, and less at the edges or corners. Though theoretical equations have been developed for design of reinforcement, generally, some empirical relations are used to estimate the amount of reinforcement (Austroads 2004, PCA 1995, IRC:101 1991). For JRCP longitudinal steel reinforcement is about 0.15 to 0.25% of the cross-sectional area of the slab, for CRCP it is about 06 to 1.0% (Swanlund and Vanikar 2002, IRC:101 1991). For steel fibre reinforcement the amount of fibre used is generally 0.75 to 1.5% (IRC:SP:46 1997). Nominal transverse reinforcement is provided to control transverse cracking in both JRCP and CRCP. Recapitulation
There are could be three categories of reinforced concrete pavement, JRCP, CRCP and pre-stressed concrete pavement. The purpose of reinforcement is mainly to prevent shrinkage cracks and to hold the cracked portion together. Various empirical formulae have been suggested to calculate the percentage of reinforcement. The basic thickness design of pavement is designed as plain concrete pavement.
CONCRETE PAVEMENT SHOULDER Objectives Objective of this lecture is to introduce various types of shoulders of concrete pavements, and their role as a part of the pavement structure. Introduction Pavement shoulders, though they are not intended to carry traffic directly, serves a number of purposes to the pavement, such as (Chakroborty and Das 2003): Prevents intrusion of water to the pavement layers. The shoulder is sometimes used as foot-path, or bikeway. The shoulder carries traffic in special situations, and gives psychological comfort to the drivers, and sometimes accommodate vehicle parking. When construction activity is going on the main carriageway, shoulder may be used as temporary driving lanes. A design traffic of 2-2.5% of the main carriageway traffic is generally used for design of concrete pavement shoulders (Ababio-Owusu et al. 2003). In this lecture, different types of shoulders of the concrete pavement are briefly introduced. Types of concrete pavement shoulder The shoulder to the concrete pavement could be unpaved or paved; again, the paved shoulders could be concrete or bituminous - they are discussed in the following. Unpaved shoulder If shoulder is unpaved, it should be well-shaped and attention should be paid to keep it in shape during maintenance operation. This is necessary for proper and quick drainage of the surface run-off. The slope is kept steeper than the main carriageway. Paved shoulder Paved shoulder prevents water percolation into the subgrade of the pavement more effectively than unpaved shoulder. Concrete shoulder The existing concrete slab can be widened (with same or lowered thickness) beyond the carriage-way width to form a paved shoulder, known as integral concrete shoulder (Austroards 2004). This is a preferred design than disjointed concrete shoulder, because this substantially reduces the stress and deflection of the carriage-way slab
(Swanlund and Vanikar 2002, Benekohal et al. 1990) and shoulder drop-off failure is almost eliminated (FHWA 1990). Field investigations show that pavement with widened slab concrete shoulder exhibit little faulting and transverse cracking (Swanlund and Vanikar 2002). Bituminous shoulder The bituminous shoulder is less expensive than concrete shoulder and provides desirable contrast between the carriageway and the shoulder. However, the joint (longitudinal) between the concrete carriageway and the shoulder is difficult to maintain. Infiltration of moisture through the joint causes fast deterioration of the asphalt shoulder (Swanlund and Vanikar 2002) Recapitulation
The shoulder to the concrete pavement could be unpaved or paved. The paved shoulders could be bituminous or concrete type. The bituminous shoulder is cheaper than concrete one, but proper maintenance of the joint between the concrete carriageway and bituminous shoulder is difficult. Integral concrete shoulder substantially reduces the stress and deflection of the carriageway slab. Thus, there is a possibility that the overall pavement design be economical.
COMPOSITE PAVEMENTS Objective The objective of this lecture is introduce pavement analysis and design considerations for composite pavements. ntroduction Composite pavement comprises of two or more pavement layers one of which is bituminous and the other is concrete. The concrete and bituminous pavements have their individual advantages and disadvantages which have been discussed elsewhere. It is believed that composite pavement, if properly designed, can take of the disadvantages associated with the individual types of the pavements. Though there could be composite pavements for fresh pavement constructions, most of the time, the composite pavement evolves due to rehabilitation projects where overlay is applied to the pavement. The composite pavement, thus, can be divided into two categories, (i) concrete overlay over bituminous pavement and (ii) bituminous overlay over concrete pavement. There could another category of pavement, where overlay consists of two layers, and the lower portion of the top overlay is made up of unbound granular material. This is called a sandwich pavement. Figure 29 schematically presents some examples of composite pavement.
Figure 29: Examples of different types of composite pavements General formulation for composite pavement Since the composite pavement is a combination of bituminous, granular, concrete, and subgrade layer, one needs to combine the individual models (Ioannides and Khazanovich 1998) of these layers to develop a formulation for composite pavement. Such an analysis problem will require (i) strain-displacement relationship, (ii) constitutive law, (iii) equilibrium equation and suitable (iv) boundary conditions. Individual layers From mechanics point of view, a composite pavement can be thought to be composed of three basic types of layers (Ioannides and Khazanovich 1998) viz. Burmister layer, subgrade layer and concrete layer . For simplicity in modeling the individual layers can be considered to be made up of homogenous, isotropic and elastic material. The basic characteristics of these layers can be mentioned as follows: Burmister layer The granular layer and the bituminous layer can be modelled as Burmister layer. A Burmister layer transfers the load through grain-to-grain interaction and it does not have flexural rigidity (Verstraeten 1967). It is assumed that this layer is infinite along the x and y direction and single circular loading may be considered as axi-symmetric about z axis. Subgrade layer The subgrade can be assumed as Boussinesq half-space which is infinite along x and y direction and semi-infinite along z (depth) direction (Jumikis 1969). The subgrade may also be modelled as elastic spring, for example, as Winkler spring bed, Pasternak foundation, Kerr foundation etc. and accordingly the different solutions can be obtained. Concrete layer The load in applied to concrete slab is resisted by bending action. The theory of plate bending is generally used in the analysis. If it is assumed that there is no deformation along the thickness of plate, then, z = 0 at any point across the depth. The basic assumption of thin plate (known as Kirchhoff plate ) bending theory is that the thickness of the plate (h) is small as compared to the other dimensions and thus the effect of and on bending is negligible. For detailed discussion on formulation of rigid pavement, the lecture on analysis of concrete pavement may be referred.
Boundary conditions Composite pavement is made up of a combination of these layers. Thus, having developed the governing equations of the individual layers, the analysis can be performed by using suitable boundary conditions. The boundary conditions for various interfaces may be mentioned as follows (Ioannides and Khazanovich 1998): (1) At interface between two Burmister layers and and smooth interface) (for rough or (for rough interface) ; smooth and interface) (for
The superscript t indicates top and b indicates bottom of the interface. The superscript i indicates the ith layer. u indicates the horizontal displacement and indicates the vertical displacement. (2) At interface between Burmister layer and rigid base (for rough or smooth surface) (for rough interface) and (3) For Boussinesq half space at and also, at
(4) At interface between Burmister layer and cement concrete slab and where, is the vertical deflection of the plate.
(5) At interface between Burmister layer and spring bed and , where is the vertical stress of the spring.
(6) At interface between two plates ; provided pressure. (7) At the surface When top surface is Busmister layer, , if ; otherwise circular area Also, When top surface is cement concrete slab, , if , otherwise, layer is not surface layer. (for pure bending), where,
Design approach As mentioned earlier, generally, the composite pavement is designed as overlay. The conventional overlay design applicable to composite pavement by (i) empirical and (ii) mechanistic approach and (iii) a special type of composite pavement as thin concrete pavement over bituminous pavement (known as ultra-thin white topping ) are discussed in the following: Overlay design by empirical approach The pavement thickness, as new pavement design, required to extend pavement life by a given overlay design period is estimated by using deflection, estimated strain, or some serviceability criteria; and the existing thickness is discounted to a some lower value than the original by using suitable factor(s). This may be called as the effective thickness of the existing pavement. The difference (i.e. the thickness deficiency ) between these two determines the overlay thickness. Thus, (36) where, h0 = the overlay thickness (either concrete or bituminous), F = empirical conversion factor for converting bituminous thickness to concrete thickness, or, vice versa, as applicable, hd = estimated overlay thickness (either concrete or bituminous), C = factor to convert the existing pavement thickness to equivalent effective pavement thickness (either concrete or bituminous), n is a factor which takes care of the bonding condition of the overlay. These factors are estimated from traffic data, or, by performing non-destructive testing (NDT), or, are derived from experience. The recommendations in Asphalt Institute (1983), AASHTO (1993), FAA (2006) etc., for estimating the bituminous overlay thickness over existing concrete pavement, are variants of this principle. Overlay design by mechanistic approach As per this approach, the composite pavement with some assumed overlay thickness is analysed mechanistically (as explained in the formulation). The critical stress-strain parameters at the critical locations are obtained, and the overlay thickness is adjusted so that the values are less than or equal to the allowable. For performing this exercise, the engineering parameters of the existing pavement layers (thickness, modulus, Poisson's ratio) needs to be known, and can be obtained through non-destructive testing. The engineering properties of the proposed overlay (modulus, Poisson's ratio) can be obtained from laboratory study. Some guidelines (FAA 2006, NCHRP 2005) recommend similar procedure for estimating the concrete overlay thickness over existing bituminous pavement. As a simplified approach, it is suggested that, the whole bituminous pavement can be idealized as spring foundation, and the effective modulus of subgrade reaction can be obtained from field study. Ultra-thin white topping High strength concrete mixed with fibers are generally used as ultra-thin rigid overlay over existing bituminous pavement and is known as ultra-thin white-topping (UTW). The UTW pavement system is analysed as a three-layer model (UTW layer, bituminous layer and a base layer with equivalent modulus) assuming a degree of bonding between
the pavement layers (Murison and Smith 2002). Bonding in UTW makes the two layers (UTW and bituminous layer) behave as a monolithic unit and share the load. The neutral axis in concrete shifts from the middle of concrete layer towards its bottom (refer Figure-30). Due to this shifting of neutral axis, the following situations develop:
Stress at the bottom of concrete layer reduces. Thus, the concrete layer can be made significantly thinner for the same loading as compared to a conventional white-topping overlay which generally has no or partial bond to the underlying bituminous layer (Goel and Das 2004). This however increases the corner stress at the top of the concrete layer. This may result in a possibility that the corner stress becomes critical than edge stress. In such a situation, the corner stress can be reduced by providing adequately thick bituminous layer as support to the UTW layer (Murison and Smith 2002).
Figure 30: Stress distribution in bonded and unbonded layers (Goel and Das 2004) Closing remarks Composite pavements have recorded excellent to poor performance in different types of projects employing various combinations (NCHRP 2001, 2005). Design and construction of composite pavements need more precision and understanding. Recapitulation Composite pavement comprises of two or more pavement layers one of which is bituminous and the other is concrete. Composite pavement generally designed as overlay on existing pavement. The overlay can be designed, by empirical approach, where the thickness deficiency is estimated by deflection testing or other NDT methods. In mechanistic based approach, the general formulation of the proposed composite pavement structure, is presented to find out the overlay thickness so that critical stress/strain parameters are within allowable limits. COMPARATIVE STUDIES OF CONCRETE VERSUS BITUMINOUS PAVEMENT Objectives The objective of this lecture is to present a comparison between the concrete pavements and the bituminous pavements. The basic structures of the concrete (rigid) pavement and bituminous (flexible) pavement have been presented in the lecture entitled "basic concepts of pavement analysis and design. In USA about 94% of the surfaced roads are made up of bituminous surfacing (NAPA 2005). In Canada, Qubec has the greatest number of kilometers of concrete pavement, which is estimated as 4% of the total road length (CAC 2005a). The famous Autobahn roads in Germany were made up of concrete
pavements (TFHRC 2001, FHWA 2005). In India, existing roads are mostly bituminous, some concrete roads were built during 1920s and 30s. The concrete road in Hyderabad, Chandni Chowk (Delhi) and Marine Drive (Bombay) were built in 1928, 1936 and 1939 respectively (Jain 2003, Jain and Sharma 1995). The pavement construction of 56 km Delhi-Mathura road was a major modern concrete road construction project (Jain and Sharma 1995) completed in 1997 (HRB 2005). The Mumbai-Pune expressway (concrete pavement) was opened to traffic in 2002 (HRB 2005). A number of concrete pavement stretches are being constructed in India in recent times. The concrete and bituminous pavements have their individual advantages and disadvantages. These have been discussed in the following. Is concrete pavement better than bituminous pavement ?
Yes. The heavy truck movement shows substantial fuel saving (ranging from 6 to 20% for different studies) while plying on the concrete road compared to asphalt road (CAC 2005b). It is postulated that since bituminous pavement undergoes deflection due to application of load, some energy is absorbed for deflecting the pavement, which otherwise could have been utilized to propel the vehicle (Zaniewski 1989) - thus there is a saving in concrete pavement. Yes. The concrete pavement virtually does not show any rutting. Therefore, driving on concrete pavement remains comfortable almost throughout the design period(CAC 2005a). Yes. The performance of the concrete pavement remains unaffected during the freeze-thaw condition compared to normal weather conditions. Therefore, no weight carrying restriction needs to be imposed during the freeze-thaw season for those countries/ regions wherever it occurs (CAC 2005a). Yes. Concrete provides better and durable skid resistance. It is generally less slippery in wet weather than bituminous pavement (CPAM 2005). Yes. The longevity of the concrete pavement is high (it is generally designed for 30 to 40 years). Unless there is design failure, the concrete pavement virtually does not require any maintenance. Only re-texturing of the top surface may have to be done routinely. Yes. The concrete pavement renders improved visibility conditions during night time, because of its better light reflecting characteristics than bituminous pavement (CAC 2005a). It as well offers a better visibility on rainy nights (CPAM 2005). Yes. Concrete does not release odourous petroleum products to the air as that of bitumen (CPAM 2005). Yes. It reflects major part of the sunlight, and therefore does not result into 'heat island' effect as observed in cities where bituminous roads are predominant (CPAM 2005). Yes. Considering Indian context, India is self-sufficient in cement manufacturing. Bitumen, on the other hand, is obtained from crude petroleum, bulk amount of which is imported (CMA 2001).
Yes. A well designed and well maintained bituminous pavement lasts for many years. There are instances where bituminous pavements are performing extremely well for 25 to 35 years without any rehabilitation (NAPA 2005, APA 2005b). With the development of the concept of perpetual bituminous pavement , it is possible to design bituminous pavement which will survive for fifty years or more with minor maintenance (Monismith et al. 2001, Newcomb et al. 2001, Von Quintus 2001).
Yes. The bituminous pavement is joint-less, and therefore provides a smooth ride surface (APA 2005b). Yes. Not only the initial investment for bituminous pavement construction is less than concrete pavement, studies show that the overall life cycle cost of bituminous pavement is also low (Eurobitumen 2005, APA 2005a). However, there are some studies which have found it otherwise (Ohio 2005). Yes. The failure of bituminous pavement is gradual, whereas failure in concrete pavement could be sudden. Yes. The bituminous pavement repair is quick, and curing time is less and therefore can be opened to traffic soon (APA 2005a). Yes. It allows stage construction (APAI 2005, APA 2005a). Pavement design with stage construction becomes quite useful when, there is a fund constraint, or, the traffic on a new pavement could not be estimated reliably (Das and Pandey 1999). Yes. Asphalt is 100% recyclable. It is the most recycled material in the construction industry in various countries (APA 2005c, Holtz and Eighmy 2000, Ikeda and Kimura 1997 ). Recyclibility of asphalt, along with various inherent advantages of recycling, also prevents undue increase in the road elevation which otherwise causes drainage and associated problems in urban roads. Yes. The bituminous pavement produces less sound than concrete pavement (CAC 2005a, APAI 2005, APA 2005a). The wear and tear to the vehicle tyres are also less from bituminous pavements compared to concrete pavements (NAPA 2005b). Yes. Because the bituminous pavements have a greater flexibility of customized design. It can be designed according to the requirement of permissible noise, surface permeability, heavy duty traffic, skid, roughness level etc (APA 2005b). Yes. Bituminous pavements are generally not affected by the de-icing materials (APA 2005b).
Closing remarks As seen from the above discussions, it can be said that it is not possible to choose (between bituminous and concrete pavements) which one is the best option. It is also possibly not a good idea to look for which one is superior to the other. The choice should rather be governed by the project specific evaluation among these two alternatives, and other possible combinations. Recapitulation The concrete pavement is advantageous over bituminous pavement because of its fuel saving, low rutting potential, free-thaw resistance, durable skid resistance, longevity, better night-time visibility, heat reflecting characteristics etc. The bituminous pavement is advantageous over concrete pavement because of its joint-less smooth surface, lesser life cycle cost, possibility of quick repair, provision for stage construction, bituminous material recyclibility, lower noise level from vehicle-tyre interaction etc. CONSTRUCTION OF JOINTS Introduction The various types of joints and their functions have been introduced in the lecture entitled 'basic concepts of pavement analysis and design. Schematic diagrams of some of the typical joint configurations have been presented in Figure-39.
From construction point of view, concrete pavement joints could be formed or sawed. Formed joints could be keyed or butt-type and can be of tied or untied variety. Expansion joints are in general formed and tied, contraction joints are generally sawed type (could be tied or untied).
(e) Longitudinal joint with tie bar (f) Tongue and groove longitudinal joint Figure-39 Schematic diagram of some typical joint configurations. Sawing is done as concrete achieves adequate strength, usually after 4 to 12 hours after placement (Swanlund and Vanikar 2002) of concrete. The timing of sawing is important. Premature sawing will cause spalling and raveling of concrete, and if sawing is late, shrinkage cracks will appear (refer Figure-40). Sawing is generally done in two stages - the initial saw cut is further widened in the second stage of sawing. The transverse and longitudinal joints in sub-base and those of in the concrete pavement are generally placed in staggered way.
Figure- 40 Joint sawing time window. Joint sealant Each joint has some gap. This joint gap is filled up by joint fillers . Then, joint sealers are poured over the joint fillers, so as to prevent water to percolate through the joints and also to make a smooth surface for travel. As per the present day practice, a single material is used to both fill and seal, and a term joint sealant has subsequently evolved (ACPA 2005). Three basic categories of joint sealants are (i) hot poured liquid sealants (example, bituminous, polymer, rubber materials ) (ii) cold poured liquid sealants (example, polysulfides, polyurethanes, silicone ) and (iii) preformed elastomeric sealants (example, extruded neoprene rubber ). Elastomeric sealants exert resistive force towards the movement of the joints. Prior to filling, the joint is cleaned by compressed air jets (MORT&H 2001). Beaker rod of circular cross-section, typically made up of foam like material, is inserted to the joints before putting sealants. Joint sealants should be poured into the groove/ gap when the gapis in maximum expanded form. Suitable primer is sometimes applied to ensure good bonding between the cut surface and the sealant (IRC:15-2002). The height of filled joint sealant should be less than the thickness of the concrete slab. Schematic diagram of contraction joint with field poured and preformed type of sealants have been presented in Figure-41.
Figure-41 Schematic diagram of contraction joint with field poured and preformed type of sealants. Joint finishing After putting the joint sealant, finishing the surface is essential so as to prevent rough ride. Squeegee, flat pate or overbanding are the various methods of finishing the surface. The technique adopted for surface finish depends on the type of sealant used. Traffic should be allowed only after the sealing material is adequately set. Figure-42 presents photograph of various stages of concrete pavement joint construction.
Figure- 42 Various stages of joint construction ( photo courtesy: Mr. Jamil Ahmed, Consulting Engineering Service, New Delhi ). Recapitulation
Concrete pavement joints could be formed or sawed. For sawing proper selection of time window is important, failing of which concrete will either spall or crack. Some of the basic varieties of joint sealants are hot poured liquid, cold poured liquid and preformed elastomeric sealants. Elastomeric sealants exert resistive force towards the movement of the joints. Ideally, joint sealants should be poured into the joint gap when it is in maximum expanded form.
Quality control issues Objective The objective of this lecture is to introduce the concept of statistical quality control applied to pavement construction. Introduction Construction of pavement involves processes like, acquiring of material, mixing, transportation, laying, compaction, curing, finishing etc. At every stage there exists some variability with the construction materials as well as with the processes involved. Thus, any parameter which is assumed to be indicative of quality, can be represented with a set of individual values of the prameter recorded during the construction. These values may be represented in the form of a distribution. The poorer is the quality control in a particular project, the more will be the spread of such distribution. Most of the time, quality control is performed through statistical parameters. Quality control through statistical parameter brings in more flexibility in quality acceptance / rejection , and does not involve subjective judgment while taking a decision, and thereby, substantially reduces dispute cases (Livneh 2002). Two terms are often mentioned with reference to the quality control issues, one is quality assurance (QA) and the other is quality control (QC). Quality assurance is the overall plan that guarantees the quality of the construction process. This, in turn, would provide necessary confidence that the constructed facility (i.e. pavement in the present case) will perform satisfactorily during its service period. Quality control is the series of
tests and measurements which are taken during the construction so as to ensure its quality (TRB 2005). Various tests A good construction practice involves uninterrupted control on the material and construction process. This is done by maintaining records of various engineering parameters before, during and after the construction process. Table-6 and 7 summarizes some of the pavement material and pavement construction parameters (respectively), with reference to a typical concrete pavement construction. Table-6 Some quality control tests for pavement material related to concrete pavement Material testing Soil Aggregates Fresh concrete Hardened concrete Quality control tests Subgrade modulus, field moisture content, maximum dry density, optimum moisture content Moisture absorption, toughness, shape Unit weight, strength gradation, strength, durability,
Air content, water content, density, slump, temperature compressive strength, flexural fatigue
Table-7 Some quality control tests during construction of concrete pavement Construction item Quality control tests
Subgrade, subbase and Moisture content, density, thickness, gradation, base slope Concrete paving Joint/ surface Thickness, level of compaction Profile, smoothness, skid resistance
Different pavement construction parameters are recorded depending on the recommendations of the particular specification being followed for the construction project. Various types of specifications are in use in different construction practices. Some of these specifications are, method/recipe based specification, performance related or performance based specification (Anderson et al. 1990), material and workmanship based specification, end result specification, warranty based specification (Hancher 1994) etc. The frequencies of the various tests for achieving quality assurance are generally specified in the respective guidelines. This frequency of testing is decided by considering the cost of the testing and reliability achieved with the acquired data (McCabe et al. 1999, 2002). The frequency value for a specific test is generally recommended in the construction guideline itself. Various tests A good construction practice involves uninterrupted control on the material and construction process. This is done by maintaining records of various engineering parameters before, during and after the construction process. Table-6 and 7
summarizes some of the pavement material and pavement construction parameters (respectively), with reference to a typical concrete pavement construction. Table-6 Some quality control tests for pavement material related to concrete pavement Material testing Soil Aggregates Fresh concrete Hardened concrete Quality control tests Subgrade modulus, field moisture content, maximum dry density, optimum moisture content Moisture absorption, toughness, shape Unit weight, strength gradation, strength, durability,
Air content, water content, density, slump, temperature compressive strength, flexural fatigue
Table-7 Some quality control tests during construction of concrete pavement Construction item Quality control tests
Subgrade, subbase and Moisture content, density, thickness, gradation, base slope Concrete paving Joint/ surface Thickness, level of compaction Profile, smoothness, skid resistance
Different pavement construction parameters are recorded depending on the recommendations of the particular specification being followed for the construction project. Various types of specifications are in use in different construction practices. Some of these specifications are, method/recipe based specification, performance related or performance based specification (Anderson et al. 1990), material and workmanship based specification, end result specification, warranty based specification (Hancher 1994) etc. The frequencies of the various tests for achieving quality assurance are generally specified in the respective guidelines. This frequency of testing is decided by considering the cost of the testing and reliability achieved with the acquired data (McCabe et al. 1999, 2002). The frequency value for a specific test is generally recommended in the construction guideline itself. Statistical analysis The quality control data can be represented in the form a time series (control chart) plot (refer Figure-43(a)) or can be represented in the form of histogram plot (refer Figure43(b)). From the histogram plot, one can perform goodness-of-fit test and obtain the type of known probability distribution it resembles the most.
Figure 43(b) Example of histogram plot In simplistic way, the statistical quality control can be performed in the following manner. The standard deviation of a particular parameter of a given type of project is a characteristic to that type of projects and may be assumed to be known. This can be termed as true standard deviation. This standard deviation value is obtained from prior experience of similar such projects. By knowing the standard deviation of the concerned parameter, one can calculate (provided the parameter is normally distributed) the number of samples to be tested (in other words frequency of testing), so that the population mean can be confidently estimated. If this mean value is outside the specified tolerance limit , the contractor may be penalized . Conversely the average test result is better than the expected, then the contractor may get bonus . Various disincentive and incentive pay factors are suggested in various construction specifications to estimate the bill payable to the contractor (TRB 2005). It may be noted that even if the quality of a parameter is adjudged as acceptable or nonacceptable based on the mean value , there is always a probability that some samples would have behaved otherwise (Livneh 2002). This is explained further in Figure-2. Figure-44(a) shows a situation where the estimated population mean value is lesser than the specification lower limit, hence, the quality is assumed to be unacceptable. However, as shown in Figure-44(a), there is a probability of that it is wrongly rejected. Similarly, Figure-44(b) shows a situation where the estimated population mean is higher than the specification lower limit, hence it is accepted. However, as shown in Figure44(b) there is a probability of that it is wrongly accepted. is called the contractor's risk and is called agency's risk. The objective of developing a quality assurance scheme is to recommend suitable sample size such that the agency's risk, as well as the contractor's risk, are minimized (Livneh 2002). At the same time, large sample size involves higher cost towards quality control process (e.g. costs towards acquiring data, data processing, testing), therefore this aspect also needs to be looked into.
Recapitulation Some parameters are identified during the construction process as the quality control parameters. Values of these parameters are noted at a specified frequency. The population mean of each parameter is estimated from the collected sample data. The quality of construction for each parameter is assessed considering the specified limits(s) and the estimated mean value. Accordingly, the pay factors for the construction items are decided. MAINTENANCE OF CONCRETE PAVEMENTS Objectives This objective of this lecture is to discuss the maintenance issues specific to concrete pavement. Introduction Any in-service pavement shows signs of distresses, and subsequently requires maintenance to retain its serviceability. Various distresses that may occur to the concrete pavement can be, in general, classified as follows (TFHRC 2006): 1. Cracking Corner breaks Longitudinal cracking Transverse cracking 2. Joint deficiencies Joint seal damage Spalling of joints
3. Surface defects Map cracking and scaling Polished aggregate Popouts Loss of skid resistance 4. Miscellaneous distresses Blowups Lane-to-shoulder drop-off and separation
Pictorial description and discussion on various types of pavement distresses can be obtained from reference TFHRC 2006. The following section discusses various maintenance actions undertaken on concrete pavement towards rehabilitation and preservation of concrete pavements. Rehabilitation and preservation of concrete pavements Repair of early distresses Some of the early distresses in concrete pavement and their maintenance actions are discussed in the following: Plastic shrinkage cracking: Plastic shrinkage cracks appear in the form of parallel surface cracks on the surface due to rapid evaporation of water from the surface. The depth of such cracks can be estimated through coring technique . This can be repaired by sealing the surface cracks by injecting low viscosity epoxy, high molecular weight methacrylate (IPRF 2003) etc. Edge slump: When the slip-form paver moves forward, the edge of concrete slab being un-supported, sometimes slumps at isolated locations. This edge slump seriously affects the performance of the constructed pavement. The repair can be done by separately placing fresh concrete (and not mortar) to the affected portions. Such repair should be conducted before application of any curing compound (IPRF 2003). Joint spalling: Joint spalling may occur due to (i) early joint sawing, (ii) use of wrong blade type, or (iii) poor operation of the sawing equipment. Reforming manually the affected portions with fresh concrete is recommended for isolated joint spalling problems. Full depth cracking: Full depth cracking of newly constructed pavement can occur due to one or combination of the following reasons (IPRF 2003): Provide a link to the 'Slip form paving system' discussed in lecture 'construction of concrete pavement' Late transverse joint sawing or insufficient depth of sawing. Misaligned dowel bars. Excessive curling and/or warping. Rapid surface cooling. Early age loading by construction equipment. Excessive drying shrinkage. Excessive base frictional restraint.
Complete replacement of the affected slab is generally recommended for such a situation. The removal of the affected slab should be done panel by panel using some saw cutting tool (IPRF 2003), and not by any impact-type tool. Repair of other common distresses Joint seal failure: If the pavement is not badly deteriorated, but the joint seals are showing signs of failure (i.e. missing or debonded sealant) - joint resealing is an appropriate action (ACPA 1993). Hot or cold poured liquid sealants can be used for resealing purpose. The old sealant needs to be removed, and the side walls of the joints need to be cleaned and made free from dust and moisture before applying fresh sealants. Loss of support: The locations of loss of support can be detected by falling weight deflectometer, or other non-destructive techniques, like infrared thermography, ground penetrating radar etc (Gordon et al. 1998, Moropoulou et al. 2001). One of the popular ways of treating loss of support is by slab stabilization, where slab is not lifted, but the voids are filled by under-sealing, sub-sealing or pressure grouting (ACPA 1994). Examples of these grouting materials are flyash grouts, asphalt, polyurethane etc. Partial depth cracking: Partial depth repairing is done when the distress is confined to upper surface of the slab, spalling or fatigue cracking is less severe, and the joints are still functional etc. Conventional portland cement, polymer based concrete or other cementations materials are used for repair purpose. The type of the material chosen is based on (i) curing time, (ii) ambient temperature, (iii) depth of cracks, (iv) expected performance etc (ACPA 1998). The repair area can be decided by sounding technique with the help of a simple hammer. Concrete is removed by sawing and light hammering from the chosen rectangular area, cleaned and bonding and patching materials are applied (ACPA 1998). For cracks of low severity, crack-stitching can be performed, which involves applying steel reinforced cement, which helps to hold the cracks tightly (ACPA 1997). Full depth cracking: Full depth repair is recommended for cracks extended to the full depth of the pavements. Full depth repair takes care of blowups, corner breaks, transverse cracking, longitudinal cracking, severe joint spalling etc (ACPA 1995). Conventional portland cement concrete, with suitable additives, if required, is generally used for full depth repair. Same procedure, as that of partial depth repair, is adopted for the full depth. Epoxy coated dowel bar/ tie bars are suitably replaced in their designed positions. Compacting, finishing, curing, sawing, joint sealing, texturing etc are done as required (ACPA 1995). Surface irregularities and loss of skid resistance: The improvement of skid resistance and removal of surface irregularities can be done by grinding and grooving . Equipment fitted with closely spaced diamond blades/ discs are used for this purpose (ACPA 2000, Villemagne et al. 1996 ). Sometimes, riding quality related defects are treated with suitable bituminous surfacing (Villemagne et al. 1996) or thin concrete overlay . Weak load transfer at joints : Weak load transfer at the joints happen due to failure of dowel bars. New set of dowel bars (or wedging system ) can be placed by machining by impact or by sawing (Villemagne et al. 1996). However, replacement of dowel bars is a difficult, costly and time-taking process. Blocking of the whole joint with pressure grouting could be another alternative. Closing remarks The proper timing at which maintenance needs to be initiated, is an important consideration. Sometimes certain limiting values are suggested on the functional or structural serviceability (say, in terms of present serviceability rating, international roughness index etc.) conditions of the pavement. If the serviceability rating falls below (or exceeds) certain specified limits, maintenance of pavement is initiated
Recapitulation
Various distresses of concrete pavement can be grossly categorized as, cracking, joint deficiencies, surface defects and other types. Maintenance strategies adopted for treatment of early stage distress of concrete pavement are generally different than the treatment of distresses which show up gradually with the movement of traffic. Most of the structural repair techniques of concrete pavement generally involve application of fresh concrete or other types binding agents. Slab replacement is generally suggested for major distresses. Joint filling, re-grooving, re-texturing can be considered as some of the functional maintenance strategies. Selection of proper timing for maintenance activity is an important consideration. Sometimes, certain limiting values of structural or functional serviceability conditions are suggested; if the serviceability rating falls below/ exceeds certain specified limits, maintenance of pavement is initiated.
BASIC CONCEPT OF TRAFFIC ENGINEERING Basics of Traffic Engineering - A Recapitulation In this section we shall use a narrow definition of Traffic Engineering. Here we shall look at the vehicular traffic on roadways. We shall study in traffic engineering how to analyze vehicular traffic on roadways and how to design various facilities which aid in the smooth flow of traffic. First, however, we shall recapitulate some things we learnt in Transportation Engineering I
Figure-1: A two lane one-way road where L1 and L2 are two lanes. In the above example there are two streams moving in the same direction. This type of road is called a one-way road.
This is a two way road; the streams move in opposing directions. Strict adherence to lanes is referred to as maintaining proper lane discipline. In India often drivers do not maintain proper lane discipline. Let us now look at Vehicles A and B in Figure3 (press the arrow button). Let us also assume that distance of a vehicle is measured from the line marked L-L. Let us also assume that time = 0, when vehicle A is at L-L. Hence as vehicles move up they trace a path on a corresponding Time-Distance plot.
Similarly, if we trace the path of many vehicles, we get a distance-time plot, as shown in Figure 5. The figure also explains various other traffic engineering terms.
The terms speed, flow and density are important as these parameters describe the state of a traffic stream. These parameters are explained in detail in the following lectures. Basic Concepts of Speed Typically speed of individual vehicles are measured at an instant or over a very short period of time. Such a speed is called Spot speed. If u i represents spot speed of vehicle i (say at time 't') then it is equal to the slope of the distance time line for the i th vehicle at time 't'. For example see the definition of u 6 in Figure 5 (reproduced from Lecture 1).
In order to determine a representative speed for the entire stream typically the average speed of vehicles over a certain period of time (usually not more than 15 minutes) is determined. Two types of averages are typically determined; Time Mean Speed
is a more appropriate measure of the stream conditions as it gives the average stream speed as the average of average speed of individual vehicles over a certain distance. Advanced Concepts As can be expected the speeds are not all the same; individual u i 's are actually distributed randomly around a certain average value. Various speed studies have been done and it has been seen that speed is distributed Normally (in order to learn more about Normal distribution click here) around the observed mean. Figure 6(a) shows a typical frequency distribution of speed on not-so-free-flowing expressway. The figure states that at this site 8 times out of 195 observations the speed was between 49.5 kmph and 50.5 kmph (this is plotted as a vertical line of height 8 on a x-axis value of 50 kmph- the mid point of the class 49.5-50.5)
Figure 6(b) shows the same information as a relative frequency (drawn as vertical lines) versus speed plot. Relative frequency of a particular data is the frequency of that divided by the total number of observations. Also shown in the figure is a superimposed plot of a normal distribution fitted to the data. In this case it is meaningful to superimpose the density function on the relative frequency graphs i.e. Figure 6(b), as the class intervals are of unit length, like 48.5 to 49.5 kmph, 49.5 to 50.5 kmph, etc (Click here if you still don't understand). In order to fit the normal distribution the mean and standard deviation of the observed data are set equal to the mean and standard deviation of the theoretical distribution. Studying speed distributions is important from various standpoints, like selection of design speed vis-a-vis speed limit, simulation of traffic streams , etc. Basic Concepts of Density
Density is defined as the number of vehicles per unit length.It is measured at a given time over a specified distance. For example,as per the data given in Figure 5 the density at time T measured over a distance of D is 5/D vehicles per km. Note that 5 is the number of times the distance-time lines cross the vertical line at time = T. This indicates that at T there were 5 vehicles between 0 and D. Density is also equal to the reciprocal of average distance headway.The distance heading is shown in Figure 5 (reproduced from lecture1) and indicates the distance between (similar points of) two consecutive vehicles.
Advanced Concepts Density by definition is an average measure. It, as stated earlier, is related to the average distance headways. Individual distance headways however are all different and vary with time. A good and detailed description of distance headways will be provided in the discussion on Traffic Flow Theory. Basic Concepts of Flow Flow, unlike density, is a point measure. It is defined as the number of vehicles that cross a section of the roadway per unit time.The units of flow is vehicles per hour(vph) or vehicles per hour per lane(vphpl). As per the data given in Figure 5 (reproduced from lecture 1), the flow across a section at a distance D, over a period of T is 8/T vph. Note 8 is the number of times the time-distance plots cross the line drawn at D between 0 and T. Physically, this means that someone standing at D from time = 0 to time = T will count 8 vehicles crossing him/her. Flow, in a way , measures how many vehicles are being serviced by the road. It is of good interest to the service providers and not much to the individual drivers. Notice, that parameter speed and density (which implied proximity of other vehicles) are of interest to and do affect the behaviour of drivers. Sometimes microscopic measure of flow (i.e. Time Headway) is of interest to individual drivers (we shall see this later). It can be seen from Figure 5 (reproduced from lecture 1), that flow is equal to the reciprocal of the average time headway.
Advanced Concepts Microscopically flow is related to time headways; time headways are simply referred to as headways. As is expected not all time-headways are the same. Studying the distribution of time headways therefore becomes imperative. Studying the probability distribution of headways helps in studying time losses at signalized intersection, delays at unsignalized intersection delays at merging operations and various other things. The study of headways distribution is also extremely important in order to able to simulate traffic streams realistically. A lot work has been done on headway distributions. Generally, the Pearson Type III family of distribution is used to represent headway distributions. Here, a brief description is provided through the following table. In the table, 'h' represents the variable headway, represents the mean of the observed headway data and h s represents the standard deviation of the observed headway data.
An example of a typical frequency and relative frequency headway distribution for headway is shown in Figure 7(a) and 7(c) respectively.
The mean headway for the data is 2.46 second and the standard deviation is 1.196 seconds. Hence the parameters of the Erlang distribution is k = 4, = 1.63. This fitted Erlang distribution is shown in Figure 7(b). The comparison of the predicted relative frequency and the observed relative frequency (try to remember the definition of relative frequency) is shown in Figure 7(c).
UNINTERRUPTED FLOW Introduction & Fundamental Relation: Theory of Uninterrupted Traffic Flow Uninterrupted traffic flow refers to flow of those streams where vehicular motion is not interrupted by stoppages. Movement of vehicles on expressways and certain arterial sections can be classified as uninterrupted traffic flow. In fact, to some extent, all flows except at or near intersections can be said to be within the scope of uninterrupted traffic flow. Our discussion on uninterrupted traffic flow is divided into four sections. The first section presents the fundamental relation of traffic flow. The second section talks about certain macroscopic relations between speed and density. The third section deals with viewing the traffic stream as an outcome of interactions between drivers; this section is titled "Microscopic Models". The fourth section analyzes situations when two different traffic streams meet. Fundamental Relation The fundamental relation of traffic flow is valid for interrupted as well as uninterrupted traffic flow. It is discussed here as this is the first instance we discuss models of traffic flow. Consider the following traffic stream;
Assume that vehicle V1 is touching section A-A and vehicle V n is touching B-B at time, t = 0. Also assume that the average speed of the stream is u kmph. For simplicity, assume that each vehicle in the stream is moving with u kmph. Further assume that the density of vehicles between A-A and B-B is k vpkm (vehicles per km). Now, if an observer standing at A-A starts counting vehicles at time, t = 0 the first vehicle he/she would count is V 1. Since, Vn is u km away from A-A and since every vehicle is moving with u kmph V n will reach A-A exactly at t = 1 hr (press replay button to see). Hence in one hour the observer will count all vehicles between A-A and B-B. That is, it will count all vehicles over the distance u km (see Figure). Now, as per our initial assumption, there are k vpkm (i.e. density = k vpkm). That is, the number of vehicles between V1 and Vn (including both), N is given by N = u (the distance between A-A and B-B) x k(density) Hence in one hour the observer will count N or u x k vehicles. By definition, number of vehicles counted in an hour is flow, q in vph (vehicles per hour). Hence, in this case, q = N or q=uxk The last equation, which says flow of a traffic stream is equal to average speed of the stream times the average density of the stream is called the fundamental relation of traffic flow. An Introduction to Macroscopic Models In example 3 of Problem Session 1 we saw that the speed and distance headway are related. As distance headway reduced, speed also reduced. Now, density is inversely related to distance headway. Hence it is expected that density and speed will be inversely related. Figure 2 shows some typical speed-density plots from three different Indian sites exhibiting uninterrupted flow. These sites were on G.T. Road and Kalpi Road in and around Kanpur, in Northern India.
As you can see, as expected, density and speed are inversely related. Macroscopic models of traffic flow primarily deals with developing models of relation between speed and density. Over the last seven decades various types of models have been developed. Here we discuss three types of models:
1. Linear/Generalized Polynomial 2. Logarithmic 3. Exponential Macroscopic Models- Linear Model/ Generalized Polynomial Model of u - k relation Linear Model Greenshields [1], based on limited data, proposed that u = a + bk with the boundary conditions that when k = 0, u = u f and when k = kj , u = 0. Note uf is defined as the free flow speed; it is the average speed at which vehicles move when there are very few vehicles on the road. Also note that k j is defined as jam density and represents that density at which jam conditions exist and no vehicles can move. Greenshields model with the boundary conditions become Clarification An example of how the linear relation may fit a u-k data set is given in Figure 3. Figure 3 shows a linear fit to the data shown in Figure 2(a).
Visually, it looks like an acceptable fit; the estimated values of u f and kj.(from the figure) are kmph and passenger cars per km. A related issue here is how to find the best fit line, given the data. Note, Figure4 shows various lines going through the same data as shown in Fig 2(a). All of them "looks" like they are acceptable fits but each one gives very different estimate of u f and kj. Obviously, a better method needs to be evolved to determine the best fit line.
Generalized Polynomial Model The generalized polynomial evolved from the work done by Gazis et al [2,3] and May & Keller [4]. Mathematically, the generalized polynomial model looks like
The model has four parameters (namely a,b,c and d) it reduces to the linear model when a = 0 and d =2. Further, the model is valid only for and d > 1. The reasons for this will be explained under microscopic theory of traffic flow. Using the same boundary conditions as before it can be shown that b = u f and c = kj. Clarification Typically, in this case a and d are chosen from expectations on optimal density and optimal flow. For a preview into definitions of optimal speed and optimal density click here. Easa and May [5] developed a nomograph which uses estimates of u f, kj, optimal speed, u0 and optimal density, k0 to obtain a (referred to as m in the nomograph) and b (referred to as l in the nomograph). This procedure is quite unique to this model (as it is not regression based); similar procedures can also be used for the linear model and the other models described later [Think how this can be done]. The nomograph is reproduced here as Figure 5. Also note 1 kmph = 0.62 mph. As can be seen from the nomograph, one starts from two directions. Given an estimate of kj one locates the appropriate point on the lower right of the nomograph. From this point one moves to the left, parallel to the x-axis till one reaches ones estimate of k 0. At this point one moves vertically up (parallel to this y-axis) At the same time, given ones estimate of u f one locates the point in the top left area of the nomograph and moves vertically down till one reaches ones estimate of u 0 ; at this point one moves horizontally (parallel to x-axis) to the right till the line intersects the vertical line of the previous paragraph. This point gives the values of a (i.e. m in nomograph) and b (referred to as l in nomograph). In order to understand the process better press the animate button on the nomograph. This then will show you how to get m(i.e.. a) and l(i.e.. b) when u f = 58 mph (93 kmp), u0 = 30 mph (48 kph) kj= 182 vpm (113 vpkm) and k0= 50 vph (31 vpkm). Note that for this case a = m = 0.79 and b = l = 2.55.
Figure : 5 Nomograph developed by Easa and May to determine a and b n order to see a realistic example of how the above method works click here (being developed) . However it is strongly suggested that the reader visit this link after going through this section completely.
Macroscopic Models - Logarithmic Model of u-k relation/Exponential Model of u-k Relation Logarithmic Model of u-k relation The logarithmic model is of the form
0 It was first proposed by Greenberg [6] in 1959. Using the boundary condition that u = 0 when k = k j one can deduce that b = k j Clarification The other boundary condition that u = u f when k = 0 cannot be used here. As k tends to zero, as per this model, u tends to infinity. This implies that in the logarithmic model free flow speed is infinite. This, of course is a major drawback of the logarithmic model. In order to estimate the parameters, a and b one can use linear regression to obtain the best fit line. Although, this will give the best fit line the parameter (i.e. a and b) estimates will not have the properties ordinarily expected from linear regression estimate techniques. [For further classification to the previous statement see the link on best fit or refer to any good book on econometrics]. The following shows how a logarithmic model can be fitted to a u-k data.
or or or
u = a ln b - a ln k u = c - ak, u = c + AK,
Note once C and A are obtained (see best fit) one can get a = -A and
The following figures (Figure 6) shows the best fit logarithmic model to the data given in Figure 2(c). Figure 6(a) shows a plot of u versus ln k(or k); from the figure it is clear that c = 59.31 units and A = 735 units (can you tell what the units are?) Figure 6(b) shows the same data and the same logarithmic model on a plot of speed versus density. [although this line is a best fit notice that kj estimate is 3979 pc per km; this is an unacceptable estimate. This implies two things: (i) the logarithmic model may not be a good model to fit to the given data, and (ii) the best fit line is not always the line which gives the best traffic parameter values].
Before leaving the section it must be pointed out that the parameter "a" has a physical meaning. Towards the end of this section we revisit this model to see what "a" means. Exponential Model of u-k Relation These models are of the form
Undowood [7] initially proposed an exponential model with b = 1. Later the Northwestern model emerged with b = 2. Typically, the parameter b is assumed and parameters a and c are estimated from a given set of data on u and k.
Using the boundary condition that when k=0, u = u f , we get a = uf. Like in the previous case one can use linear regression to estimate a and c; again, as in the previous case, the observations on the properties of the estimates apply. One can rewrite the above model as
or
or Note that b is a pre-defined value (like 1 for Undewoods model and 2 for Northwestern model). Once A and B are estimated, a and c can be obtained as a = e A and . Figure 7 shows the best fit Northwestern model (i.e. b = 2) for the data shown in Figure 2(b). In figure 7(a) the plot is between ln u versus k 2/2 From the plot the values of A and B are 3.5787 units and -0.0002units.(can you tell what the units are?)
) on a graph of u
Before leaving this section it must be pointed out that c has a physical meaning and will be discussed later. Also try to figure out what happens to density when u tends to zero. Is this acceptable? Next we look at macroscopic models of u - q (speed - flow) and q - k (flow - density). These macroscopic models are directly obtained from the macroscopic u - k relations. Uninterrupted Flow Introduction As stated earlier the speed-flow (u-q) relations and flow-density (q-k) relations can be obtained from using the u-k relation and the relation. However, before looking at the connection between the u-k relation and the u-q and q-k relations, first, their general nature and idealized views are presented. It must be understood here, that these relations are still the topic of research and new observations and understandings of the process are still emerging. The discussion here has been kept traditional because the purpose is to introduce the reader to these relations and not necessarily to make the readers experts in the area of flow theory. Figures 1 and 2 show some typical u-q and q-k plots that are generally observed on roads.
Figure 2 Flow - density relation generally observed from the field Few points emerge from these plots: 1. In either of the curves two portions are clearly visible. In Figure 1, these are the top and bottom portions, and in Figure 2 these are the left and right portions. 2. In the top portion of Figure 1, the points are closely packed where as in the bottom portion the spread is much more. Similar observation can be made from Figure 2 when one compares the left and right portions. 3. The figures show that a maximum value exists for flow, q. This is known as capacity, q max . 4. The clarity in the figures is least around the capacity value. 5. There always seem to be fewer points just after capacity than anywhere else on the curve
The observed relations (as depicted in Figures 1 and 2) have traditionally been idealized as those shown in Figures 3 and 4. Figures 5 and 6 show a more recent idealization of the relations depicted in Figures 1 and 2. As this is an introductory lecture on this topic the idealizations shown in Figure 3 and 4 are used for the rest of the discussion.
Figure
more
modern
idealization
of
the
speed
flow
relation
Figure 6
The reader must also note the following: 1. q = 0 when either u = uf or k = kj (the reader should satisfy himself/herself as to why it is so). 2. uo, ko are referred to as optimal speed and optimal density and are by definition the speed or density at which the flow is maximum. 3. As mentioned before, the value of maximum q is often written as q max and referred to as capacity.
Let us now look at how one can obtain a u-q or q-k relation given the u-k relation. In order to show this, two examples are taken.
Example 1 Determine the implied u-q and q-k relations for a linear (or Greenshields') u-k relation.
Hence,
Or, This then is the implied q-k relation. In order to obtain the u-q relation one can proceed as follows:
Hence,
Or, The shapes of the q-k and u-q relations obtained from the linear model are shown in Figures 7 and 8.
Figure 8: The u-q relation implied by Greenshields' linear u-k relation Few more observations can be made from the above relations. First, let us see at what speed and density do the flow become maximum. The process of determining this is illustrated using only the q-k relation.
As shown before,
Hence,
max
, or
Note that the above expressions for optimum speed and density are only true for linear u-k relation. Example 2 Determine the implied u-q and q-k relation for a logarithmic (or Greenberg's) u-k relation.
Hence, This then is the implied q-k relation. In order to obtain the u-q relation one can proceed as follows:
Hence, The shapes of the q-k and u-q relations obtained from the logarithmic model are shown in Figures 9 and 10.
Figure 10: The u-q relation implied by Greenberg's logarithmic u-k relation
Few more observations can be made from the above relations. First, let us see at what speed and density do the flow become maximum. The process of determining this is illustrated using only the u-q relation.
As shown before,
Hence,
At q = q max , ; therefore at u = uo (note at this point q = qmax ) the derivative must be zero. Therefore,
, or
Interrupted Traffic Flow Introduction to Interruptions In this lecture the flow of traffic under traffic interruptions is studied. Traffic interruptions are collectively a general situation used to denote conditions where a stream of traffic flowing under certain conditions of speed and density meets another stream flowing under some other condition of speed and density. For example, if traffic flowing at high speed meets another stream flowing at a greater density and lower speed, the former traffic stream will face a traffic interruption. There are various ramifications of such interruptions on traffic flow and these are discussed in the next section on shock waves. Within this general description of traffic interruptions, two kinds of traffic interruptions are of special interest. These are the interruptions that take place at signalized and unsignalized intersections. In this section, there are three subsections; the first studies shock waves. The second and third study the traffic flow at signalized and unsignalized intersections, respectively. Shock Waves Whenever a stream of traffic flowing under certain stream conditions (say, speed = u A , density = kA, and flow = qA) meets another stream flowing under different conditions (say, speed =uB , density = kB, and flow = qB) a shock wave is started. The shock wave is basically the movement of the point that demarcates the two stream conditions. This demarcation point may move forward or backward or stay at the same place with respect to the road. The rate at which this demarcation point moves (the direction of motion of the vehicles is taken as the positive direction) is referred to as the speed of the shock wave. In order to see the generation and movement of shock waves consider the distancetime graph shown in Figure 1.
Figure 1 This figure is drawn for the following situation. A slow moving vehicle with speed u B (whose distance-time plot is shown as a dotted line) enters a traffic stream originally moving at uA , kA, qA. This slow moving vehicle slows the traffic and creates another flow condition denoted by uB , kB, qB. From the graph it can be seen that there exists a line (a bold line marked Shock wave 1) that is the locus of the point that demarcates the two flow conditions at any given time. After a while at point Q the slow moving vehicle leaves the traffic stream and the congested condition created by the slow moving vehicle is released, say, at some other stream condition (denoted by u C, kC, qC). Obviously, the point that demarcates the flow conditions u B , kB, qB from the flow conditions uC, kC, qC also causes a shock wave. The locus of this point is marked as Shock wave 2 in the figure. At some time these two shock waves meet signalling the end of the flow conditions u B , kB, qB. But this time the flow conditions uA , kA, qA meet the flow conditions uC, kC, qC starting Shock wave 3 (see figure 1 ). Interestingly, the vacant area in the figure bounded by the line denoting the motion of last vehicle to go without being caught behind the slow moving vehicle, the dotted line denoting the motion of the slow moving vehicle, and the line denoting the motion of the first vehicle to get released after the departure of the slow moving vehicle, also represent a flow condition; namely the free flow conditions. In the figure this zone is indicated as a stream with conditions u = uf (the free flow speed), k = 0, and q = 0. Hence when the stream condition u B , kB, qB meets the stream condition u = uf , k = 0, and q = 0 a shock wave should and does emanate. The only difference here is that the point which demarcates the two conditions is the same as the point representing the slow moving vehicle; hence, the shock wave that emanates here moves with the slow moving vehicle and the time-distance diagram of this shock wave is the same as the time-distance diagram of the slow moving vehicle. Similarly, the time-distance diagram of the first vehicle that is released is the timedistance diagram of the shock wave that emanates when flow condition u C, kC, qC meets u = uf, k = 0, and q = 0 From the above discussion it is clear that the key parameters of a shock wave are (i) the point at which it starts, (ii) the point at which it ends, and
(iii) the speed of the shock wave. As will be apparent later, of these, the only parameter which needs to be studied in detail is the speed of the shock wave which forms the topic of the next subsection. Speed of shock waves Let a stream flowing under condition A (with u A , kA, and qA) meet another stream flowing under condition B (with u B , kB, and qB). Assume that the speed of the resultant shock wave is at a speed of . Then relative to the shock wave vehicles in condition A are moving and those in condition B are moving at a speed of .
Now, recall that the shock wave is a demarcation between the two conditions. Hence, it can be said that in a time duration of the number of vehicles crossing over the shock wave from condition A is . Similarly, it can be said that in a time duration of the number of vehicles crossing over the shock wave from condition B is . Note that, physically, some vehicles are crossing over the shock wave from one condition to the other. Since vehicles are neither created nor destroyed in the process of crossing over, the number of vehicles crossing over the shock wave from the perspectives of conditions A and B must be equal. Therefore,
by
It should be noted that the above equation for speed of shock wave can also be obtained using principles of geometry from a distance-time plot of the type shown in Figure 1. Further, the above equation also has a simple graphical interpretation; it states that the speed of the shock wave is given by the slope of the line joining the points representing the two conditions (on a graph) whose confluence gives rise to the shock wave. Figure 2 shows this graphically.
Fig. 2: Illustration of the speed of shock waves on a There can be three types of shock waves:
plot.
(i) the forward moving shock wave, i.e., speed of shock wave is positive (see Figure 3 (a)), (ii) the stationary shock, i.e., speed of shock wave is zero (see Figure 3 (b)), and (iii) the backward moving shock wave, i.e., speed of shock wave is negative (see Figure 3 (c)).
plot.
As can be seen from the figure3, the first type of shock wave will occur when a stream with lower flow and lower density meets a stream with higher flow and higher density or when a stream with higher flow and higher density meets a stream with lower flow and lower density. Stationary shock waves will occur when the streams meeting have the same flow value but different densities. The third kind of shock waves will occur when a stream with higher flow and lower density meets a stream with lower flow and higher density or when a stream with lower flow and higher density meets a stream with higher flow and lower density.
In the following web page an example is worked out in order to show how knowledge of shock waves can be used to obtain different traffic flow parameters of interest and also to illustrate how we can obtain information about where a shock wave starts and where it ends (which are the other two parameters related to the description of a shock wave). Shock Waves Example Traffic is moving on a one way road at q A = 1000 vph , and kA = 16 vpkm. A truck enters the stream at a point P (which is at a distance of 1 km from an upstream benchmark point BM) at a speed of uB = 16 kmph. Due to the decreased speed the density behind the truck increases to 75 vpkm. After 10 minutes, the truck leaves the stream. The platoon behind the truck then releases itself at capacity conditions, q C = 1400 vph and kC = 44 vpkm. Determine (i) the speed of all shock waves generated (ii) the starting point of the platoon (behind the truck) forming shock wave (iii) the starting point of the platoon dissipating shock wave (iv) the ending points of the platoon forming and platoon dissipating shock waves (v) the maximum length of the platoon, and (vi) the time it takes for the platoon to dissipate, and also plot the (vii) location of the front of the platoon and the rear of the platoon versus time, and (viii) length of the platoon versus time. Solution Consider the distance-time diagram shown in Figure 4 plotted for the scenario described in the problem. This diagram is shown here to help the reader understand the problem better; strictly speaking the complete diagram is not necessary for solving the problem. However, an understanding of the physical scenario is of definite help.
Fig 4: Distance time diagram illustrating the example problem on shock waves. (i) Speeds of the various shock waves generated (shock wave can be obtained directly by using Equation 1 as follows: = = = 3.39 kmph is denoted as SW )
= -6.45 kmph
= 14.29 kmph
= 16 kmph
= 31.8 kmph
ii) Shock wave 1 is the platoon forming shock wave. It starts at Point P and at the time when the truck enters the stream. (iii) Shock wave 2 is the platoon dissipating shock wave. It starts at Point Q (i.e., where the truck leaves the stream) and 10 minutes (note the truck remains in the stream for 10 minutes) after the truck entered the traffic stream. Point Q is 16 x (10/60) = 2.67) km downstream of Point P.
(iv) Both shock waves 1 and 2 will end if the platoon condition (i.e., condition B) ends. This condition will end whenever Shock waves 1 and 2 meet. Say they meet at time t hours after the start of Shock wave 1, where their positions must be the same. Their positions at time can be determined from their starting positions and distances by which these travel during time t. Thus, knowing that P is 1 km from BM and Q is 3.67 km from BM, one can write the following
or,
Hence, the two shock waves end 22.84 minutes after the start of Shock wave 1 and at a distance of 1+ 3.39 x 0.381 = 2.29 km downstream of BM. (v) The maximum length of the platoon will be at the instant where Shock wave 2 is just about to start. The platoon at any given time is defined by the length between the front of the platoon (Shock wave 4) and the rear of the platoon (Shock wave 1). Hence, the length of the platoon grows at a speed of 16 - 3.39 = 12.61kmph. The length is maximum at 10 minutes after the platoon starts forming. Hence the maximum length is equal to 12.61 x (10/60) = 2.1 km. In terms of number of vehicles the maximum length of the platoon is kB x 2.1= 75 x 2.1=157.5 158 vehicles. (vi) In part (iv) it was determined that the platoon ceases to exist 22.84 minutes after the start of platoon formation. Out of this for the first 10 minutes the platoon only grows (and there is no dissipation). Hence it takes 12.84 minutes for the platoon to dissipate. (vii) Figure 5 (a) shows the required plot. In the plot, time is assumed to be zero when Shock wave 1 starts; the distances are as measured from BM. In the figure, Lines 1 and 2 represent the front of the platoon and Line 3 represent the rear of the platoon. Further, the slope of Line 1 is equal to U SW4 , the slope of Line 2 is equal to U SW2 , and the slope of Line 3 is equal to USW1. (viii) Figure 5 (b) shows the required plot. In the plot, time is assumed to be zero when Shock wave 1 starts. Slope of Line 1 is equal to the rate of growth of the platoon. This value, as determined in part (v), is 12.61 kmph. The slope of Line 2 is basically the rate of dissipation of the platoon; however, this need not be calculated since one knows the maximum length of the platoon and when the platoon completely dissipates.
Fig. 5: (a) Plot of platoon front and rear locations versus time for the example, and (b) Plot of length of platoon versus time for the example.
Traffic Flow at Signalized Intersections An intersection is a location where two or more roads carrying traffic streams in different directions cross. The space which is common to all these roads is referred to as the intersection. At such a location, obviously, different traffic streams compete with one another for the use of the common space or the intersection. If left on its own, the flow at an intersection will always be chaotic; the safety and efficiency at such locations will be low. Hence, various strategies are used to control the flow of traffic at an intersection in order to improve the safety and efficiency of traffic flow. Among the strategies that are used, signalization is the most common. At a signalized intersection, the common space is periodically given to certain flows while the other conflicting streams are barred from entry at that time. In a manner of speaking, the common space is time-shared among the various flows. Although, there are various kinds of time sharing strategies like pre-timed, partially actuated, and fully actuated signalizations, in this section the pre-timed signalization and its effect on flow is studied. The other strategies, where the time sharing mechanism changes more frequently than in the pre-timed strategy, are not studied here as the basic traffic flow analysis process is the same as that of the pre-timed strategy. In the pre-timed signalization, the time sharing between the different conflicting flows occurs according to a pre-defined strategy which repeats at a fixed interval. This fixed interval is referred to as the cycle length. During the cycle length, the time for which a particular stream can utilize the intersection is referred to as the green time for that stream or movement, the time during which a particular movement cannot utilize the intersection is referred to as the red time for that movement. Invariably during the change over from green to red an amber signal is shown to warn the driver that a red signal is impending. During the amber time for a movement, the vehicles of that movement can use the intersection. Of course, the sum of green, amber and red times for a particular movement is equal to the cycle time. A detailed discussion and design of
pre-timed signalization is provided in the next lecture. The above brief introduction is given here in order to initiate the reader to the various terminology used in signalized intersection analysis and design. Obviously, at a signalized intersection, there is interruption to flow of traffic. The type of interruption and its effect on the flow is described in the next several web pages. Flow characteristics at signalized intersection The interruption to traffic flow at a signalized intersection is orderly and deterministic. Consider the following scenario. The signal has just turned red for a particular stream or movement. All vehicles on this stream come to a stop and remain stopped (thereby forming a queue) till the light turns green. Once the light turns green the first vehicle in the queue departs followed by the other vehicles. Movement of vehicles continue unabated till the light turns amber at which point vehicles close to the intersection generally go through while the ones farther away from the intersection initiate maneuvers to come to a stop. The same pattern follows for every cycle. Given this interruption pattern, the following processes become important for analysis: (i) the arrival process of vehicles (ii) the departure process of vehicles (iii) the queue of vehicles, and (iv) the delay to vehicles. In this subsection, the first two processes are described. The next subsection analyzes the other two processes. The last section in this topic analyzes the related matter of capacity and level of service at signalized intersections. Arrival process at signalized intersection Arrival processes at intersections could be of three kinds (i) random arrivals, (ii) grouped arrivals, or (iii) mixed arrivals. In random arrivals vehicles seem to arrive at the intersection randomly. Such arrival pattern is seen at isolated intersections, i.e., intersections at locations where there are no other upstream intersections in the vicinity (say within 3 to 4 km). In these cases the inter-arrival times (or headways) are often distributed more or less according to the negative exponential distribution. That is, if h is headway between vehicles, then the probability that a particular headway is between H 1 and H2 is given by (2) where, is the reciprocal of the mean or average headway. The above expression is derived by integrating the negative exponential probability density function (for headways in this case) between H1 and H2. The negative exponential probability density function, f(h) (for all h > 0) is given as (3) An assumption (or observation) of negative exponential distribution for headways also implies that the vehicle arrival process is a Poisson process. That is, the probability that the number of vehicles Nt, that arrive in a time interval t, is equal to k is given by:
(4) where, is the reciprocal of the mean or average headway or, alternatively, it is the mean or average arrival rate. Grouped arrivals are seen at intersections which are located close to (say within 2 km) another upstream intersection. In such cases, arrival process seems to be uniform and vehicles can be assumed to arrive at reasonably constant headways. This phenomenon occurs because vehicles arriving at the intersection are the ones which have been released by an earlier intersection and therefore are in a platoon. Mixed arrivals are seen at intersections which are located at intermediate distances (say between 2 to 4 km) from another upstream intersection. Here, the arrival cannot be characterized either as purely random or as purely grouped. This is because the distance between the upstream intersection and the intersection being studied is large enough for many of the released vehicles to disperse from the discharged platoon and arrive independently; yet the distance is not large enough for the entire platoon to disperse. Hence, some vehicles still arrive in a grouped manner. Departure process at signalized intersection When a signal turns green from red the first of the stopped vehicles initiate manoeuvres to move and cross the intersection. The next to cross the intersection is the second vehicle in the queue and so on. If one measures the headways, i.e., the time gaps between successive vehicles when they cross a pre-specified point on the intersection (generally the stop line on the road), then an interesting and expected pattern emerges. Figure 6 shows a typical plot of headways versus position in queue. The ordinate value for an abscissa value of gives the headway between the and vehicles in the queue when they cross the pre-specified point in the intersection. Further, the ordinate value for an abscissa value of 1 indicates the time-gap between the light turning green and the first vehicle crossing the pre-specified point in the intersection.
Fig. 6: A typical plot of departure headways versus position in queue at a signalized intersection. From the figure6, two features emerge: (i) the headway stabilizes to a value h s, referred to as the saturation headway; this value basically states the maximum number of vehicles that can ever be released during a specified green time and (ii) the initial headways are larger than hs. The second feature highlights that although vehicles can move at a headway of h s, the initial vehicles take a longer time due to perception /
reaction time (to the light turning green) and the extra time taken to accelerate to a reasonable speed (note that the latter vehicles more or less achieve this speed when they cross the specified point as they start moving from a distance further upstream from the specified point. In a sense then, some time is lost due to the fact that the initial vehicles take longer than h s. The sum of these excess times is referred to as the startup lost time, ls. That is, (5) Near the end of the departure process some time is also lost. This happens because invariably some part of the amber time remains unutilized because vehicles come to a stop even when some part of the amber time is still remaining. This loss of time, referred to as movement lost time (or sometimes as clearance lost time), l m , is primarily due to the fact that drivers are never aware of the remaining amber time. Delay and queue analysis at signalized intersections In this section the delay faced by vehicles at signalized intersections and the queues developed at signalized intersections are studied. Consider the plot shown in Figure 7. In this figure, the abscissa is time and the ordinate is the cumulative number of arrivals as well as the cumulative number of departures for a given stream (or approach) at an intersection. There are two lines in the figure. One shows a typical graph for cumulative number of arrivals on the given approach at a signalized intersection, the other shows the cumulative number of departures from the given approach at the signalized intersection. On the abscissa, the time is divided into slots named Cycle I, Cycle II, and so on. These slots represent the cycles at the intersection. Each cycle is further subdivided into R and G. The R represents the duration of effective red (i.e., the time during which no vehicle on this particular approach crosses the intersection); similarly the G represents the duration of effective green (i.e., the time during which vehicles on this particular approach crosses the intersection).
Fig. 7: A typical plot of cumulative number of arrivals and departures on an approach to a signalized intersection.
Figure 7 gives a reasonably complete picture of the arrival and departure processes at the intersection. From such a plot, we can obtain information about both delay and queues. The horizontal distance at a value of n on the ordinate, for example, will give the delay faced by the n th vehicle to arrive at the intersection. Hence summing all such horizontal distances (or equivalently the area between the two lines) will give the total delay faced by all the vehicles arriving at the intersection. The total delay divided by the total number of arrivals will provide the average delay. Similarly, the queues on the approach can be easily determined from this figure. For example, the vertical distance between the two lines at time t will give the queue length at the intersection approach at time t. This implies that the queue lengths at any given time can be obtained easily from the figure and hence the parameters such as average queue length, variance of queue lengths, and the like can also be obtained. However, obtaining such graphs for each and every intersection at all times is not feasible. Hence it is imperative that we analyze the delay to vehicles and queues with an aim to derive equations which can give these quantities once data on arrival rates, cycle lengths, green times, red times, etc. are known. In the following, such a description of the analyses procedures is provided Delay analysis at signalized intersection To begin with assume that the arrival process is deterministic and vehicle arrive at a uniform rate. Further, assume that the system in unsaturated, that is, the total number of vehicles that arrive in a period is less than the total number of vehicles that can be served by the system. These two assumptions mean that the arrival rate is such that all the vehicles that come in a cycle are cleared within the same cycle (like the situation in Cycle I of Figure 7). The average delay to vehicles for this case can then be easily determined from the figure shown in Figure 8. The figure shows a typical cumulative arrival/departure graph against time for an unsaturated, uniform arrival rate approach to an intersection. The slope of the cumulative arrival line is , where is the uniform arrival rate in vehicles per unit time. The slope of the cumulative departure line is sometimes zero (when the light is red) and sometimes (when the light is green); where is the saturation flow rate obtained as the reciprocal of the saturation headway explained earlier; is expressed as vehicles per hour of green per lane or vphgpl.
Fig. 8: A typical plot of cumulative number of arrivals and departures on an unsaturated, uniform arrival rate approach to a signalized intersection.
From the figure 8 (where C is the duration of the cycle length and g the duration of the effective green period) it can be seen that the total delay (under the assumptions stated above), , is given by
(6) Assuming that n is large enough so that the discrete sum of d(i) is equal to the area of the triangle in the figure, the following can be written: (7) From the figure, where, can be easily determined by noting that
can be written as
Hence, (8) From the above and noting that n, the total number of vehicles that arrived, is average delay under the above assumptions, , can be obtained as: (9) However, an approach to an intersection may not always stay unsaturated. There may be periods of over-saturation during which the arrivals from one cycle spill over to the next and so on. A similar situation can be seen in Cycle II of Figure 7. Under this assumption, and the assumption that arrival rate is still deterministic and uniform, the plot of cumulative arrivals/departures versus time can be drawn (see Figure 8). In the figure it is assumed that the arrival rate from 0 to time T is , and that is large enough to cause oversaturation (i.e., C greater than sg, the maximum number of vehicles that can be served during a cycle). C the
Fig. 9: A typical plot of cumulative number of arrivals and departures on an oversaturated, uniform arrival rate approach to a signalized intersection. From the figure 9 it can be seen that the total delay under these assumptions, i.e. , is given by the sum of the area marked with horizontal stripes (Area I) and the area marked with inclined stripes (Area II). This implies that the average delay, i.e. , under these assumptions is the sum of the average delay due to Area I and the average delay due to Area II. The average delay due to Area II can be easily determined by assuming the dashed line as the cumulative arrival line and using Equation 9. However, first the slope of the dashed line needs to be determined and later substituted for in Equation 9. If the slope of the dashed line is taken as (and noting that the slope of the inclined part of the cumulative departure line is , the saturation flow rate) then
or
for
(10)
The average delay due to Area I can be determined by looking at the average time between the cumulative arrival line and the dashed line. If the horizontal distance (i.e., time in this case) between Point P and the dashed line is taken as Z, then it can be said that the time between the cumulative arrival line and the dashed line increases linearly from 0 to Z over a time period of 0 to T. From this it can be said that the average delay for vehicles arriving between time 0 and T is Z/2 (Note that the area of the triangle formed by the cumulative arrival line, a horizontal line from P and the dashed line is given by Zn/2, where is the number of vehicles to arrive till time T). Similarly, the time between the cumulative arrival line and the dashed line decreases linearly from Z to 0 over a time period of T to M. Following the same logic as above, the average delay due to Area I to vehicles arriving between times T and M is Z/2. Hence, it can be said that the average delay due to Area I is Z/2 irrespective of when a vehicle arrives. If one assumes the vertical distance of Point P from the dashed line as y, then
where, , represents the maximum number of vehicles that can cross the intersection from a given approach per unit time. Thus the average delay, is given as (12) In reality, however, more often than not the arrival is not deterministic, it is stochastic. Once it is assumed that the arrival is stochastic the above relations cannot be used and can at best function as approximate estimates. Assumption of stochastic arrivals lead us to analysis which is beyond the scope of this section. Hence, in this text, only the relations generally used to determine delay under the above assumptions are described. One of the relations often used in determining delay is due to Webster [259]. Webster assumed that the arrivals are according to a Poisson distribution, departures occur uniformly and at a maximum rate of , the average arrival rate is such that , and that the queueing process runs under similar arrival and departure conditions long enough for the system to stabilize to a steady state (where, for example, cycle to cycle variations in average delay, average queue length, etc. are minimal). Based on these assumptions and some simulation runs (where the arrival and departure processes at a
signalized intersection are simulated for various arrival pattern and signal settings) Webster proposed the following expression for average delay, : (13) The first term in the equation is the same as that given in Equation 9, the second term is the additional term which results from analyzing the process by assuming stochastic arrivals, the third term is a correction factor obtained from simulation studies. It is seen that this term is often between 5 and15 per cent of the sum of the first two terms. Hence the following simplified form of the above equation is sometimes used.
In practice, it is found that the estimates of delay are not good for the entire range of /C values. In general, when /C values are close to one (that is, C is close to , this implies an increase in the chances of oversaturation when the stochasticity
in the arrival rate causes the number of arrivals to be greater than ), overestimates the average delay. One possible reason for this is that the derivation of the quantity assumes steady-state behaviour, which is never achieved at real intersections as oversaturation by design occurs only in short spells. Other researchers around the world have developed other equations which are supposed to predict delays more realistically. All of them predict values which are close to when /C is not high (say less than 0.8) while their estimates are much lower when /C is high (say greater than 0.95). The 1985 Highway capacity manual of USA [104] proposes the use of one such expression for delay, , reproduced here as Equation 14. The relation was developed based on a large database on intersection delay. (14)
The 1985 HCM [104] cautions users against using this relation for . In the 1998 HCM [103] the calculation of delay has been further modified. That expression is not provided here as it uses many site specific empirical constants which are not valid for Indian conditions Queue analysis at signalized intersections An involved discussion on analysis of queues at signalized intersections is not possible in this text as it requires substantial knowledge of stochastic queuing processes. However, certain characteristics of the arrival and departure processes from the point of view of queueing analysis are presented here so that the interested reader may pursue this topic further.
The primary purpose of doing a queueing analysis at a signalized intersection is to be able to determine the probability distribution of queues that form. That is, the aim should be to answer questions like, what is the probability that there will be vehicles in a particular queue at any time. This is important since not only values such as average queue length are important, what is more important is an idea of the queue length distribution so that one can determine lengths of auxiliary lanes (see next lecture for details) for pre-designated values of probability of overflow (where the queue of vehicles is larger than the space provided by the auxiliary lane) and probability of blockage (where the queue of vehicles on the lane adjacent to the auxiliary lane is so large that it blocks the entrance to the auxiliary lane). However, the theory of queues available so far is not able to model the queueing process at a signalized intersection in a simple manner. The primary reason for this is that the departure process is not a Poisson process. If one assumes (and correctly so) the time a vehicle waits at the top of the queue to be its service time, then it can be easily seen that the service time duration (to which the departure process is integrally connected) is not negative exponential. In fact the service time is either zero (if one reaches the top of the queue when the light is green) or equal to the duration of the red period (if one is the next to the last vehicle to have crossed the intersection); the probability of either of these service times being the service time of a particular vehicle is not easy to calculate. Matters can be further complicated if there is a permitted phase in the signal (as is sometimes the case for turning movements) where a vehicle can cross the intersection (during the non-green time) if a sufficient gap in the opposing stream exists. Some researchers (see for example Kikuchi et al. ), however, have attempted to determine the probability distribution of queues through a Markov chain analysis of the queueing process. The interested reader may refer to Kikuchi et al.'s work cited above for a good understanding of the analysis procedure. Data collection at signalized intersection At signalized intersections, other than collecting data on arrival rate and pattern (which can be done in the usual manner of counting volume and recording time headways), data on delay and saturation flow rates may need to be collected. In this section, procedures for collecting data on delay and saturation flow rates are described. Collecting data on average delay In this section a procedure which can be easily used to collect data on stopped delay at a signalized intersection is described. The procedure relies on the principle that the area between the cumulative arrival and departure plots (see figure 7) gives the total time that all the vehicles spend stopped at the intersection; its unit is vehicle seconds. This area divided by the total number of arrivals obviously gives the average delay. The area can be obtained by summing either all the delays or all the queue lengths, that is, Total area between cumulative arrivals and departure plots where all the variables are as explained in figure 7 Hence the average delay can be obtained as
The data collection procedure uses the above relation to evaluate the average delay. The method relies on determining the area by observing the queue lengths at short
intervals of time over the entire experiment time period, and counting the total number of vehicles during the entire test period. The procedure is explained step-by-step as follows: Step 1. Decide the time period P for which the data will be collected. Decide the interval of time I at which, the queue length at the intersection will be counted. The cycle length C should not be an integral multiple of I. Let the number of intervals in P be m. Step 2. Count the queue length qi at the end of each interval. Continue counting till P is over. Also over the time P, count the total number of vehicles that arrive at the intersection, i.e. Step 3. Estimate the area as . See Figure13, which is a figure similar to
Figure 7. The area obtained using is generally seen to be more than the actual area, and hence the estimate obtained here is reduced by 10% in order to get a closer estimate of the true area.
Collecting data on saturation flow rate The saturation flow rate is the reciprocal of the saturation headway. The saturation headway, as suggested in the subsection on DEPARTURE PROCESS, is the headway at which latter vehicles discharging from a queue cross the stop line. In general, the HCM [103] suggests that the average value of the saturation headway s i for sample i can be obtained using
where Tj,i is the time at which the jth vehicle of the queue crosses the stop line for the i th sample, and L stands for the last vehicle in the queue. The assumption here is that from the fourth vehicle onwards all vehicles more or less maintain the saturation headway. According to the HCM [103], the average value of the saturation headway should be estimated as the mean of all s after repeated sampling
Example for signalized intersection On an approach to a signalized intersection, the effective green time and effective red time are 30 s each. The arrival rate of vehicles on this approach is 360 vph between 0120 s; 1800 vph between 120-240 s, and 0 vph between 240-420 s. The saturation flow rate for this approach is 1440 vphgpl. The approach under consideration has one lane. Assume that at time = 0 s the light for the approach has just turned red. 1. Plot the arrival rate of vehicles versus time. 2. Assuming arrival and departure processes to be continuous, plot the cumulative number of arrivals and departures versus time. 3. Determine the average delay to vehicles arriving between 0-120 s. 4. Determine the average delay to vehicles arriving between 120-240 s. 5. Determine the average delay to vehicles arriving between 0-240 s. 6. Determine the delays to the fourth and the sixtieth vehicles that arrive at the intersection. 7. Determine the maximum delay faced by a vehicle on this approach. 8. Determine the maximum queue length on this approach. At what time does the queue length first become equal to the maximum? 9. Determine the percentage of time for which there exists a queue on this approach. 10. Determine the average queue length between 120 and 420 s. Solution The following parameter values are provided:g = 30 s, C = 30 + 30 = 60 s , s = 1440 vphgpl, from 0-120 s = 360 vph, from 120-240 s = 1800 vph, and from 240-420 s = 0 vph, and T, the time for which there exists a flow higher than , is 120 s. Note, c = (g/C)s= 720 vph. (i) Figure 14 gives the plot of arrival rate of vehicles versus time. (ii) Figure15 gives the plot of cumulative arrivals and departures versus time. (iii) Between 0-120 s the intersection is operating under unsaturated conditions. Further, the arrival is deterministic and uniform. Hence we can use Eq 9 to determine the average delay. Therefore,
Fig. 14: Plot of arrival rate versus time. We can also determine the average delay directly from the graph, by noting that the area of either Triangle I or Triangle II in Figure 16 (which is the same as Figure 15 but with few extra annotations) divided by the total number of arrivals during a cycle will give the average delay. Therefore, Average delay = Area of Triangle I or II Number of arrivals in a cycle
(iv) Between 120-240 s the intersection is operating under oversaturated conditions. The arrival is deterministic and uniform. Hence we can use Equation 12 to determine the average delay. Thus,
We can also determine the average delay directly from the graph (see Figure16), by noting that,
Fig. 15: Plot of cumulative number of arrivals and departures of vehicles versus time.
(v) The average delay to all the vehicles between 0-240 s can be obtained by dividing the total delay (faced by all vehicles) by the total number of vehicles. Hence,
where,
is the average delay to a vehicle coming during 0-120 s is the number of vehicles that arrive during 120-240 s is the average delay to a vehicle arriving during 120-240 s. Hence,
Of course, we can also find the average delay here from the graph (the reader should do this).
(vi) The arrival rate of vehicles from 0 to 120 seconds is 360 vph or 0.1 vps. Assuming that the fourth vehicle arrives before 120 seconds, the time of arrival of the fourth vehicle is 4/0.1 = 40 seconds. (Hence the assumption is not violated). The departure rate of vehicles is 1440/3600 = 0.4 vps. The time of departure of the fourth vehicle, assuming that the fourth vehicle gets discharged during the first green, is 30 + 4/0.4 = 40 s. (Since the departure time is less than the start of the next red, the assumption is valid.) The delay to the fourth vehicle therefore is departure time - arrival time = 40 -40 = 0 s The same observation can be made from Figure 16.
Fig 16. Example: Plot of cumulative number of arrivals and departures of vehicles versus time. The delay to the sixtieth vehicle can also be read from Figure 16 as 144 seconds. (vii) As can be seen from the Figure 16 the maximum delay is 180 seconds. (viii) As can be seen from Figure 16 the maximum queue length is 36 vehicles. At time = 240 s, the queue length first becomes equal to 36 vehicles. (ix) As can be seen from Figure 16 there are no queues from 40-60 s and from 100-120 s. For the rest of the time, there is a queue at the intersection. Hence, the percentage of time for which there is no queue at the intersection is (40/420)100 = 9.52. Hence the percentage of time when there exists a queue is . 100 - 9.52 = 90.48% (x) We can determine the average queue length directly from Figure 16, by noting that,
Capacity and LOS analysis Introduction and capacity of facilities with uninterrupted flow Introduction Capacity, in its broadest sense refers to the maximum vehicle carrying ability of a transportation facility. Two terms in the above definition needs further clarification; they are "vehicle carrying ability" and "transportation facility." Vehicle carrying ability typically implies the number of vehicles that can be "processed" or "transported." In other words it is the flow , q . Hence, capacity of a facility is the maximum flow at the facility. A transportation facility is any engineering structure created for the purpose of transporting people or goods from one place to another. Examples, of such facilities include: (i) multi-lane expressways, (ii) two-lane roads, (iii) interchanges, (iv) signalized intersections, (v) unsignalized intersections, and others. Typically, one transportation facility differs from another in the way vehicular streams behave at these facilities. From an engineering standpoint it is imperative that one knows what the capacity of a particular facility is; only then will one be able to design a facility which is adequate for a given demand (which typically would be in units of flow ). In the rest of this module, the concepts of capacity and level-of-service are presented under three sub-headings: (i) capacity for facilities with uninterrupted traffic flow, (ii) capacity at intersections, and (iii) level of service. Further, the intention of these lectures is not to provide details of how capacity is computed but to bring forward the factors that impact capacity and the principles on which capacity computations should be based. The reason for not going into specific details of how capacity is computed is that Indian provisions for capacity computations are still in its infancy and codes in this regard are still in their formative stages Finally, it must be said that a detailed description of capacity is beyond the scope of this course; only a course specifically dedicated to traffic flow theory can do justice to this vast area of study. Capacity at facilities with uninterrupted traffic flow Capacity is qmax and occurs at u = u 0 and k = k0. Capacity analysis is that field of traffic engineering which tries to relate the design features of a transportation facility to the capacity it offers. In the case of facilities which offer uninterrupted traffic flow, like expressways, (or arterial sections with large inter-signal spacing), the design features are (i) the number of lanes, (ii) the width of lanes, (iii) the width of shoulders, and (iv) layout (like the frequency of curves and their curvatures, the general gradients, etc.). The purpose of capacity analysis is to see whether any of them affect capacity and if so what type of relation exists between them. The question as to whether a design feature affects capacity is ultimately related to whether the design feature has any impact on the behaviour of drivers; i.e., whether the feature affects the speed-density relation (recall that speed-density relation is a product of driver behaviour). It has been observed, that all of the features mentioned above does have an impact on the speed-density relation and therefore on the capacity. Most of the time the reason for the impact is that drivers are affected psychologically by the presence or absence of a certain feature. For example, although, physically there is no reason why a narrower lane should impact driver behaviour yet the closeness of the edges affect a driver in such a way that a driver drives slower than usual.
For example, if the u-k relations on two roads with varying lane (road) widths are plotted then conceptually they may look something like what is shown in Figure 1. (The reason for using straight lines for the u-k relation is purely to make the figure easier to comprehend). As can be seen, chances are that both free flow speed and jam density will be lesser in the case of the narrow lane width road than in the case for a wide lane width road. Further, and more importantly, the speed at any given density is expected to be lower on the narrow road. This is because individual drivers will drive at a lesser speed at a given distance headway than they would have done on the wider road possibly due to a higher perception of threat to ones safety on the narrow road.
Figure 1: Conceptual u-k relations on lanes (roads) with different widths Lower speeds at every density then would imply that flow at any density (or speed) will be lower for the narrower road and hence capacity will also be lower. The analysis of capacity concentrates on determining how the capacity value changes when any of the design features (like lane width) changes. In this sense, the scope of capacity analysis is narrow. However, recent trends show that traffic engineers are trying to determine the entire speed-density (or speed-flow) relationship with changing design features. In the following, some observations on capacity of roads with different design features are listed: 1. Wider lanes increase capacity (per lane); however, beyond a certain value, the width of the lane stops playing any role in increasing capacity. 2. Wider shoulders increase capacity; like in the previous point, beyond a certain value the width of the shoulder does not affect capacity. 3. Straight, flat roads increase capacity; grades (or slopes in the vertical direction) affect capacity because it sometimes affect the performance of vehicles; horizontal curves always affect the way a driver drives because of the impact of the centrifugal acceleration and because of restricted sight distances. In order to give the reader some idea of how these observations have been taken into account in working relations we discuss a method used by one of the most detailed capacity manuals, namely the Highway Capacity Manual (HCM) published by the Transportation Research Board of USA. In the HCM method it is assumed that free speed ( uf ) of a stream on a given road indicates how drivers will behave over the entire range of density on that road. Hence being able to predict the free speed on a road with given specifications will allow one to determine the capacity of the road. This, in a nut-shell, is the basis on which capacity is determined by this method. Various tables and charts are presented which allow the engineer to determine what the free speed will be on a road with certain characteristics. Different speed-flow diagrams are provided and one needs to choose that which closely matches the free speed expected on the given road. This speed-flow diagram is then
assumed to be the one for the given road. The capacity is simply read as the maximum flow as per the selected speed-flow diagram. Other methods for determining capacity are also viable. For example, there are methods which define an ideal condition (in terms of road width, shoulder width, etc.) and a capacity value for this condition (often referred to as the ideal capacity ). Next these methods, define factors which reflect the proportion by which capacity has to be reduced for a given departure from the ideal condition. For example, if lane width is far removed from the ideal condition then one could define a proportion by which the capacity has to be reduced for such a departure. Typically, the product of all such reduction factors (from different parameters which are non-ideal) are determined and used as the factor by which the ideal capacity is reduced to reflect the capacity for nonideal conditions. Before leaving this discussion, it must be pointed out that the maximum number of vehicles that can cross a section (i.e. flow) obviously depends on the vehicle mix of the traffic stream. If the percentage of heavy vehicles (like trucks and buses) is more then it is expected that the total number of vehicles that cross a section will be lesser than in the case where all vehicles are, say, automobiles. Typically, the analysis of capacity is done assuming that all vehicles are automobiles. Hence when a stream contains different types of vehicles then the vehicles in the stream are converted to automobiles using an equivalency factor. For example, if a truck is considered to be equal to 2.5 automobiles then if a stream has 100 trucks per hour it is assumed that instead of trucks the stream has 250 automobiles per hour. The equivalency factors are referred to as passenger car (i.e., automobile) equivalency factors. Generally the passenger car equivalency factors for all types of vehicles are determined. These factors are used to convert a stream with a given mix of vehicles to an equivalent stream with only automobiles. The determination of equivalency factors is an issue which needs attention from researchers in the area and is beyond the scope of this lecture. Capacity at signalized and unsignalized intersections and LOS
Capacity at signalized intersections The idea of signalized intersections was introduced in Module VII of this lecture series. It may be recalled that at an intersection more than one road (referred to as approach) meet. Hence, in general, it is not meaningful to talk about the capacity of the intersection; what is more appropriate is to talk about the capacity of individual approaches. In fact, if on an approach, movements (like through, right-turn, etc.) are separated spatially (like right-turn vehicles on a separate right-turn lane) and temporally (like separate green times for different movements) then it is appropriate to talk about capacity of each movement. Figure 2 shows a typical signalized intersection. In this intersection the road carrying traffic from east to west have right-turn movements separated (both spatially and temporally) from other movements. Here it is best to talk about capacity of the right-turn movement and the capacity of the through + left-turn movements separately. Yet on the road carrying traffic from south to north all the movements share the same road space and are not separated. In this case it is best to talk about the capacity of the approach (which is the road from south to north).
Figure 2 n the rest of this lecture, we shall generally take the approach as the basic entity (as that is more appropriate for Indian conditions) and talk about the capacity of an approach. The capacity of an approach depends on the saturation flow and effective green time to cycle time ratio . The idea of saturation flow was introduced in Module VII. Saturation flow, si , gives the maximum number of vehicles that are expected cross a point on the approach i in an hour if the approach has green time for the entire hour; typically the unit of si is vphg (vehicles per hour of green). If the effective green time (i.e., the green time minus the lost times - see Module VII for a description of lost times) on an approach i is gi and the cycle time (i.e., the time between successive greens on the same approach; also see Module VII) is C then the capacity on the given approach is
The unit of capacity is vph. The above relation does not need further explanation as the ratio basically denotes the fraction of time (say in an hour) the vehicles on approach i are allowed to move. It may be beneficial to the reader if the contents under departure process at signalized intersections in Module VII are read with this section. The determination of saturation flow is akin to the determination of capacity at facilities with uninterrupted flow. Saturation flow primarily depends on the number of lanes (or width of the road if lanes are not demarcated), type of movement and the types of movements which share an approach. It can also depend on the gradient of the approach and the available sight distances at the approach. As before issues of vehicle mix exist at intersections and can be taken care of in similar fashion. However, the passenger car equivalencies of different vehicles generally are different from the corresponding values in the case of uninterrupted flow
Capacity at signalized and unsignalized intersections and LOS Capacity at unsignalized intersections It is suggested that the reader recapitulate the discussions in Module VII under unsignalized intersections.
As in the case of signalized intersections, it is meaningful only to talk about capacities of different movements or approaches to the unsignalized intersection, and not about the intersection as a whole. Given the earlier description of the flow characteristics at such intersections (see Module VII), it is evident that it is best to proceed from an understanding of the factors on which the capacity of a movement at an unsignalized intersection depends to an understanding of the capacity of an approach. The factors on which capacity of a movement depends are: Gap availability: Any traffic stream (or more specifically, movement) at an unsignalized intersection moves by accepting gaps in the conflicting streams. Obviously then, the more the number of conflicting streams the lower is the gap availability; similarly, the higher the volume in the conflicting streams the lower is the availability of acceptable gaps. Further, greater the size of the critical gap, the smaller is the number of acceptable gaps. Therefore, the total number of gaps available for use by a movement depends on the number and volume of conflicting streams and the size of the critical gap. Hierarchical position: Not all gaps available for use can, however be used by the vehicles of a movement. As stated earlier, there exists a hierarchy of movements at any unsignalized intersection. The movements which are higher in this hierarchy get a preference over the movements lower in the hierarchy when it comes to actually using an available gap. In effect, the lower the priority of a movement the more it is impeded from using a gap because the chances of higher priority movements wanting to use the same gap are more. Hence, even if a particular movement has a lot of available gaps it may get to use only a few of them because of its position in the hierarchy. Thus the capacity of a movement depends on: (i) number of conflicting movement and conflicting volume, (ii) critical gap size, and (iii) hierarchical position of the movement amongst all movements vying for a gap at the intersection. Often, at unsignalized intersections, a lane (or an approach) may be shared by more than one movement. In such cases, it may happen that a particular gap of interest to a given movement never gets used because the vehicle of that movement is caught behind vehicles of other movements (to whom the gap may not be of interest) sharing the same approach. That is, even if a gap is available (an unclaimed by others) it may not be used because the vehicle which could have used it had other vehicles (from other movements) ahead of it in the queue. For example, a vehicle wanting to go straight (which generally has available gaps all the time) may not be able to do so because a right turning vehicle is ahead of it. To the right turning vehicle these gaps (which are of use to the through vehicles) are of no interest. These situations can occur more frequently if the number of movements sharing the approach is more. These factors are sometimes incorporated in capacity calculations in a manner outlined below. The discussion is limited to principles only. Step 1: The movement whose capacity is to be determined is identified (this is referred to as the test movement, TM). Given the TM and the intersection geometry, all conflicting streams are identified. Step 2: Based on the conflicting volumes and the critical gap size an estimate of the total available gaps for TM is determined. The determination is typically done through empirical or quasi-theoretical curves. Step 3: Given the hierarchical position of TM, all streams which are higher in the hierarchy (and compete with TM for gaps) are identified. For each such higher priority
movement, an impedance factor is determined. The impedance factor from a particular movement, in a way, states the number of times that movement uses a gap which TM could also have used. Again, these impedance factors are determined from empirical observations. Typically at the end of these steps one can approximate the capacity of the TM. Often different movements share an approach. In such cases, sometimes gaps are not utilized as the vehicle which could have utilized the gap is stuck behind others. The chances of this happening depend on how many different types of movement share an approach and what are the relative volumes of these movements. These values are used in conjunction with the capacities of all the movements (obtained at the end of Step 3) which share an approach to determine the capacity (often referred to as shared capacity) of the approach. Obviously, if an approach has only one movement then the capacity of the approach is same as the capacity of the movement.
Level of service (LOS) - The concept Level of service of a facility at a given time is defined as the conditions (as perceived by the driver) under which a driver is operating at that facility. Typically the LOS is divided into six categories --- LOS(A) through LOS(F). Each category defines a set of conditions which describes the ease or difficulty with which drivers can operate. For example, LOS(A) stands for a situation where the driver can drive without being impacted by other vehicles, can freely choose his speed, is not inconvenienced, and so on. LOS (F) on the other hand, is a situation where congestion is unbearable, drivers are subjected to stop-and-go traffic, they have to be extremely alert at all times, has no freedom to maneuver, etc. The crux of the matter however is to define a measurable traffic parameter which can tell a traffic engineer (who is not in the stream) how a driver in the stream at a given time might be feeling. Different parameters have been suggested for different facilities as the indicative parameter. For example, density has often been suggested as the parameter which can act as an indicator for LOS on facilities with uninterrupted traffic. Delay has been suggested as the parameter at signalized intersections, etc. Once a parameter has been identified, typically cut-off values are decided for each of the LOS categories. For example, it may be decided that if density is less than kA then LOS is A, if it is between kA and kB then it is LOS (B) and so on. It must be said that the limits are often ad hoc and does not have very strong basis. Even the selection of a single parameter as an indicator may be questioned, nonetheless they remain. Interested readers may look at various codes that exist around the world for an idea of the prevailing practices. Traffic Planning Process Introduction, transport supply and demand Introduction Transportation planning is an area which deals with the process of planning new facilities and the process of determining how best to utilize the available facilities. These two processes are really not independent of one another but are two sides of the same activity of efficiently utilizing the available resources. Alternatively, one can think of transportation planning as a macroscopic study of the inter-relationship between demand and supply of transport facilities.
Traditionally, engineering has been involved with only the supply side analysis. For example, engineering (in any discipline) deals with devising techniques to better handle the needs (or demand). For example, given the load (which is the demand) civil engineers have the task of designing a structure (something which meets the requirements); or given the requirements of water in a particular area a civil engineer had to design the size of the channel or pipe. That is, in engineering the demand is always assumed to be known and fixed. Similar is the case with transportation engineering, which primarily deals with providing services like roads, intersections, parking lots, which cater to a given demand. Transportation planning, on the other hand, looks at both transportation demand and transportation supply to evolve better plans to provide mobility to goods and people. The study becomes all the more important because transportation is one of the few engineering areas where demand depends on supply and vice versa. For example, how many travels on a road depend on how good the road is which is again driven by the demand there is for the road. In the next few lectures the following topics will be covered: 1. The interdependence of transport supply and demand 2. Overview of the planning process 3. Role of government in transportation planning (Goal setting) 4. Study of needs or requirement (Demand analysis) 5. Development of alternatives (Supply analysis) 6. Evaluation of alternatives (Impact analysis) Interdependence of transport supply and demand The demand for transportation is a derived demand. The demand does not exist because people gain directly from moving around (except in those rare cases when people go on a joy ride) but because people have to move in order to satisfy some other requirement, like education (going to school), livelihood (going to office), etc. Hence, transportation demand and its pattern have a lot to do with where people's origins and destinations are. For example, if peoples' homes are located in the suburbs and the offices are located in the city center then one can expect a lot of demand on the arterials which link the city center to the suburbs. In other words, transportation demand and the land-use pattern (land-use is a word used to describe what use a particular parcel of land is being put to; for example, land-use could be commercial, residential, agricultural, industrial, etc.) of an area are integrally connected. What is interesting, is that land-use of a particular parcel of land is impacted by the transportation facilities in and around that parcel. The facilities in turn are impacted by the requirements (or transportation demand). This interplay (also shown in Figure 1) leads to a unique (at least in civil engineering) situation where land-use impacts demand which in turn impacts land-use; and the cycle continues. Any transportation planner must understand this and land-use planning and transportation facilities planning must go hand in hand. Unfortunately, this course does not offer us the scope to discuss this issue further.
Figure 1: The inter-relationship between land-use, transportation demand, and transportation infrastructure. Overview of the planning process, role of government
The following diagram gives a good overview of the transportation planning process:
Figure 2 : Overview of the transportation planning process As can be see from the figure, the planning process takes in two inputs, namely, the goals and policies set by the elected officials or government bodies, and the requirements from the transportation infrastructure over the planning horizon. The latter is an outcome of studies which predict transportation demand. Demand analysis is the topic of the next two modules. Given the two inputs, knowledge of traffic engineering is used to develop various alternative plans which achieve the goals (these plans could be like signalize intersections, build grade-separated intersections, introduce rapid transit,
etc.). This may be referred to as supply side analysis and is largely within the purview of traffic engineering. Once the alternatives are in place, they need to be evaluated and this is done through various impact analyses. These will be briefly described in Lecture 3. Finally, if an acceptable plan is found then the process ends with the initiation of an implementation guideline; else the planning process continues till an acceptable plan is formulated or the goals modified to realistic levels. Role of Government The government plays a role in transportation planning in three different ways: (i) The government sets the goals and objectives which drive transportation infrastructure planning and development. For example, in the recent years, the government had come up with a policy of connecting all reasonable sized villages with the nearest road through an all-weather thoroughfare. Another such goal was to develop high speed road connections between all metropolitan cities. Often these goals are stated through enactments or declaration of schemes. For example, laws allowing private parties to run airlines in India could be thought of as an offshoot of the goal to improve air-connectivity in India. (ii) The government develops laws which have an indirect impact on transportation planning process. For example, the laws relating to clean air may give rise to various impacts on the planning process; like removal of older vehicles, encouragement of public transport, introduction of modes which use clean fuels or renewable energy sources, development of means to reduce congestion, etc. Laws relating to land-use also fall under this category. (iii) The government leads the planners in certain directions by allocating funds preferentially. For example, the government may come up with a scheme where local bodies can be allocated more funds to widen existing roads (which, say, can alleviate congestion and hence reduce pollution) if at the same time these bodies promise to phase out older autos from the roads (see this also helps the governments agenda to improve air-quality) Demand analysis, supply analysis and evaluation Demand analysis Demand analysis deals with predicting the flow on road sections in an area given the land-use data, demographic data, and various other socio-economic characteristics of that area. This is the subject of the next two modules and hence is not discussed here. Development of alternatives (Supply analysis) As stated before, development of alternatives include designing, building and operating traffic and mass transit facilities. These are the subject matter of traffic engineering and are therefore not discussed here. However, in recent times, effective demand management has also become a strategy which is used as an alternative solution for certain problems or for achieving certain goals. Demand management is any action intended to influence the intensity, timing and spatial distribution of transportation demand (which is, the number of people wanting to go from one place to another). The student should note that for various problems effective demand management is as good a solution as improvements in the facilities. For example, congestion on certain roads can be reduced if schools and offices do not start at the same time which is the same goal one would achieve by widening the road. Yet, the first method does not cost much and is much easier to implement. In the following various demand management strategies are described. The list is not
complete and innovative techniques are developed every now and then to control demand effectively. Demand management strategies: (i) Alternative work schedules: This strategy reduces the peaks in demand by spreading the demand over a wider time window. For example, if demand on a particular road around 9:00 am is 1000 vph (out of which 300 vph are due to school going kids and the rest due to office goers) then by staggering the school and office start times the peak demand may come down to only 700 vph. (ii) Pricing: For example, tolls could be introduced for single-occupancy vehicles trying to access the business district during the morning peak hours. This kind of congestion pricing dissuades people from coming in their own vehicles to office and encourages car-pooling, or shifting to public transport. There can be many other innovative ways in which tolls can be used to control demand. (iii)Employer incentives and dis-incentives: Employers can play a crucial role in managing transportation demand (this is in addition to Point (i) mentioned above). Most employers provide free car-parks for their employees; discontinuation of this perquisite has definite impacts on how many people come to office in their own vehicles. Employers can provide incentives to employees who use the public transport by subsidizing the monthly passes; this again may encourage employees to shift to public transport. There are various other ways in which employers can impact the way their employees commute to office. It must be realized that demand management becomes important during the peak travel hours and that during these hours the largest percentage of people on the road are office goers; hence the role of employers as demand managers. (iv) Land-use reorientation: As has been discussed, demand on transportation networks are impacted by the land-use pattern in the adjoining areas. Hence, by reorienting landuse the government can impact the demand on the network. Typically, such reorientation can be achieved through what are known as zoning laws (regulations specifying the legitimate uses of different parcels of land). For example, the government can decide that only land lying on the outskirts of a city (and preferably near less traveled roads) can be used for office complexes. It is not very difficult to see that this strategy works on the demand in the long run and cannot effect any change to the demand pattern in the short run. Evaluation of alternatives Alternative solutions are evaluated using criteria which can be divided into two distinct groups. The first group consists of those that study the feasibility of implementing the alternative; and second group consists of those that study the various impacts that will follow if the alternative is implemented. Feasibility studies The two major criteria that have to be used in order to establish the feasibility of implementing a project are: (i) technical feasibility and (ii) financial feasibility. Technical feasibility studies relate to studying whether the alternative being evaluated is technically feasible with the given level of expertise available. For example, trains with very high speeds are not possible; hence an alternative which envisages connecting two cities 200 kms apart in half-an-hour through high speed commuter trains is not technically feasible. Financial feasibility studies relate to whether the capital required for implementing and operating the alternative is prohibitively high. If so, even if a project is extremely
beneficial it cannot be considered because the money for implementing it is not available. Impact analysis Broadly speaking, any transportation project should be evaluated for its (i) land-use and demographic impact, (ii) environmental impact, and (iii) economic impact. These are briefly discussed in the following. Before, describing these, one point may be mentioned. In all the impact analyses described here, an important underlying principle is whether the negative impacts and benefits of a particular project are incident upon the same group of people. It should be realized that often the benefits are enjoyed by one group of people and the negative impacts are borne by another group of people. So, even if the benefits outweigh the costs one should always ask the question as to who are benefited and who are inconvenienced. Land-use and demographic impact: The impact of transportation projects on land-use have been described earlier and is not repeated here. The other impacts of transportation projects are related to displacement of people and other social issues. People are often displaced because of the project itself (like building an airport, which requires huge amount of land, may displace people) and sometimes they are displaced because they can no longer afford to live in the same locality any more. For example, development of a good road to an area may increase the land values of that area and also bring in well-to-do families as settlers. This may increase the cost of living in the area and poorer original inhabitants may have to move away. Sometimes, the impact is also positive. Environmental impact: The impacts on the following areas should be studied --- (i) airquality, (ii) noise, (iii) energy consumption (especially if the source is non-renewable), (iv) water quality, and (v) habitat. The transportation system is a major source of air-pollution. Pollutants are typically of the following kinds: (a) carbon monoxide (b) hydrocarbons (c) oxides of nitrogen (d) oxides of sulphur (e) particulate matter (f) ground-level ozone and (g) some air toxins. Often the contribution increases if congestion exists. Transportation facilities often produce high levels of noise and vibration. In India, the problem is further compounded by indiscriminate use of legal and illegal horns. It has been proved that sustained exposure to high levels of noise is detrimental to the physical and mental health of humans. The transportation sector is one of the largest consumers of petroleum products. The negative impacts of large consumption of non-renewable energy forms have been well documented and are not described here. Most transportation facilities have impervious surfaces, like roads, tarmac, parking lots, etc. These then stand as impediments to groundwater recharge as most of the water falling on its surface is converted to runoff. The problem becomes particularly severe in urban areas where the proportion of such surfaces is high. Inappropriate disposal of motor oil is also taking a toll on water quality . Often construction of transportation facilities and the associated pollution results in change in the ecosystem primarily due to the loss or degradation of habitat for both the flora and fauna of an area. The effects are complex and long term. Yet, their evaluation is necessary.
Economic impact: That transportation effects economy is a well-known fact. A cursory look at the flourishing economies of the world and their transportation facilities establishes the positive correlation between the two beyond any doubt. In the following some points are given to provide some idea to the reader as to why transportation has a strong effect on economy. 1. Often a large proportion of a household's budget is spent on heads which are affected by transportation either directly or indirectly. For example, expenses on trips to school, work, market, etc. are directly related to transportation. On the other hand, expenses on grocery, vegetables, etc. are indirectly affected by the efficiency of the transportation system. In the short term, prices of certain commodities go up if certain transportation links are disrupted. In the long run, prices of all commodities go up if the transportation becomes costlier. Hence, cheap and reliable transportation frees up substantial portion of a household's income which can now be spent elsewhere. 2. Transportation impacts the economic health of a region by improving its accessibility to other regions and therefore to newer and possibly better paying jobs. 3. Inefficient transportation costs the society dearly in terms of lost time (in delays and congestions) and therefore lost productivity. 4. Pollution related to transportation also has negative impact on the economy of the society as medical expenses and time lost due to sickness increase. The next two modules deal with transportation demand analyses Demand Analysis I Introduction This lecture introduces the topic of transportation demand and its determination. Transportation demand, simply stated, is the demand for trips that exist in any area. All of this demand, however, may or may not materialize into physical trips (vehicular or pedestrian) -- and some of it generally remains latent and is referred to as hidden demand. The importance of analyzing the transportation demand in order to be able to predict the expected number of trips in a given network cannot be overstated. The demand for transportation forms the primary input in any decision related to creation and management of transportation and traffic facilities, such as roads, intersections, parking lots, transit system, and so on. The lecture is divided into three sections. The first section transportation demand and how it can be analyzed. Then it demand analysis technique -- the most frequently used transportation demand. Finally it briefly describes some mechanisms employed in travel demand analysis. describes the nature of presents the sequential method of determining of the data collection
Transportation demand, unlike demand for other commodities, such as wheat, coffee, housing, clothing, etc., is a derived demand. That is, one demands to be transported not because he/she just wants to move (except for those rare cases when the person goes out for a joy ride!) but because he/she wants to achieve some other purpose such as reaching school, or office or a movie theater. In other words, the need for achieving some goal (like reaching office or a shop) creates the need to travel. Hence, travel demand is primarily generated by the population's need to work, entertain (themselves), socialize, study, etc. Therefore, it is not surprising that two of the major aspects in travel demand analysis are land use and trip purpose. Land use refers to the pattern of land usage in an area. Land use affects transportation demand through generation and distribution of trips. The effect of land use on transportation demand is not necessarily a one-way effect but rather a part of cycle in which land use changes transportation needs which in turn change land use. Figure 1 shows a simple schematic of how land use and transportation demand are related.
Fig.1: Relationship between land use and transportation demand. Trip purpose refers to the purpose for which the trip is being undertaken. Travel demand behaviour changes with the trip purpose. For example, a person hardly exercises any choice for work trips; i.e. does not necessarily decide every time whether to go to work or not. A person obviously does not decide where to go to work (generally it is fixed over a period of time for a large section of the population), even the choice of route and mode are not daily decisions. On the other hand, for recreational trips, an individual makes a large number of decisions, such as whether to go or not, where to go, how to go. Consequently, the travel demand behaviour for work trips varies considerably from that of recreational trips. This example, can obviously be extended to other types of trips such as shopping trips, etc. Given the effect of trip purpose on travel demand behaviour, the analysis of travel demand is done separately for different trip-purposes. Although, the above discussion throws light on some of the factors which affect travel demand, some more understanding of travel demand is necessary before one can analyze the demand and can, with some degree of confidence, predict the volume on various links of a network. Generally, a trip (which is the basic quantity in travel
demand) materializes after the trip-maker makes certain decisions. These decisions can be broadly classified as follows:
The decision to travel. The trip-maker, given his/her requirements, makes a conscious decision to travel so that the requirements can be met. The decision on the choice of destination. The trip maker also makes a decision as to where he/she wants to go; for certain kinds of trip purpose, such as work trips, this decision may not exist; yet for other kinds of trips, such as shopping trips, there may be certain alternative locations to choose from. The decision on the choice of mode. The trip-maker also takes a decision as to what mode of transport to use for a given trip. This decision, however, is only available to those who have access to different modes and are not captive users of any particular mode. The decision on the choice of route. The trip-maker on any given trip takes a definite decision on which route to take so as to reach the destination. Again, this decision is available to only those trip -makers who have access to modes which can use different routes as per the wishes of the trip maker. Such modes would generally include personal automobiles or two wheelers.
Although, there is unanimity on the fact that the above decisions can aptly capture the entire trip-making behaviour of an individual and hence can be used to analyze travel demand pattern of an area, it is difficult to ascertain whether there exists any definite sequence in which these decisions are made. Generally it is assumed, primarily for the ease of analysis rather than anything else, that the decisions are made in a strict sequence as shown in Figure 2. Analysis techniques which assume that such a sequence exists are referred to as sequential demand analysis techniques.
Fig. 2 : Schematic representation of the assumption of sequential decision making. Although, even today transportation demand is analyzed sequentially, the assumption that the four major decisions (see Figure 2) of a trip-maker follow a strict sequence (i.e., are in a series) is possibly not the most appealing. Quite often, the decision to travel is changed because an appropriate destination does not exist; or an initial choice of destination is changed because it cannot be reached by the desirable mode of transport. It is possibly a truer picture of reality if the decision making framework is assumed to have feedback loops. One such possible structure is shown in Figure 3. In this structure, unlike in Figure 2, there are feedback loops indicating that decisions taken earlier can be changed based on a latter decision. For example, the decision to
travel may be aborted because at the mode choice stage if it is realized that none of the available modes suit the requirements.
s stated earlier, even though the assumption of the existence of a strict sequence in the decision-making process of a potential trip-maker may be debatable, generally sequential demand analysis is used to determine the travel demand. As will be seen throughout this lecture, even with this simplifying assumption of sequential decision making, the analysis of transportation demand remains sufficiently complex. Figure 4 shows a schematic of the sequential demand analysis procedure. The figure attempts to not only illustrate the logic of the analysis procedure but also gives the names of the different classes of models used to mathematically describe each decision-making phase of the analysis procedure. Each of these classes of models is described in detail in the next webpages. Before presenting the models, a general overview of the entire process is provided. In this analysis procedure, first the entire study area is divided into various zones. These zones are generally obtained from the land-use pattern of the area. Next, for each zone the total number of trips generated in that zone are estimated using the trip-generation models. The outputs of the trip-generation models are then used to determine the number of trips between all zone pairs using the trip-distribution models. Given the tripdistribution pattern (which is often referred to as the origin-destination (O-D) matrix), the relative shares of these trips for the different modes are estimated using the modechoice models. Finally, the traffic-assignment models estimate the volume on each link of the network by determining the routes that will be used by the trips.
Trip-generation Models The trip-generation models strive to predict the number of trips generated by a zone. These models try to mathematically describe the decision-to-travel phase of the sequential demand analysis procedure. It may be mentioned here that typically the term trip-generation is used to mean trip production -- generally the trips made from households --and trip attraction -- the trips made to a particular urban location or activity. However, it is felt that analysis of trip attractions should not be within the purview of trip-generation models which attempt to quantify a populations urge or propensity to travel. Rather, trip attractions are an outcome of the destination-choice phase of travel behaviour. Similar concerns about trip attractions being part of the trip generation phase of urban demand analysis have been also voiced in Kanafani [132]. In keeping with this, the present section discusses trip-generation primarily in the context of trip productions. Trip attractions are assumed to be an outcome of the destinationchoice phase and are discussed in the section on trip- distribution models. There are basically two different model structures for trip generation models -- the cross-classification models and the regression models. However, both these model
structures incorporate the same basic factors which affect the trip generation of a zone; the models only differ in their characterization of these factors. The factors (for any given trip purpose) which affect the trip generation of a zone are:
The number of potential trip-makers in the zone; this data could be captured by variables like residential density, average household occupancy, age distribution of occupants, and so forth. The propensity of a potential trip-maker to make a trip; this is related to automobile ownership, accessibility to public transportation, and the like. For example, persons who own automobiles make more non-work trips than persons who do not own automobiles. Accessibility of the zone to potential destinations for a given trip-purpose satisfaction; variables like distance to potential destinations can capture this factor. For example, persons who live close to various recreational facilities may make more number of recreational trips than persons who live in areas which do not have nearby recreational facilities.
The Cross-classification Model The cross-classification model, sometimes referred to as category-analysis model, is based on the assumption that the number of trips generated by similar households or households belonging to the same category is the same. According to this model if in Zone there are households in category and if is the average rate of trip generation per household in category k then the relation of trips generated (or produced) by Zone i, Ti is given by (1) The model predicts the trips produced by a zone by simply aggregating the total trips produced by all the households in that zone. However, two basic questions need to be answered here: (i) how do we define similar households, or alternatively how do we define categories of households, and (ii) how do we determine the rate of tripgeneration for a given category of households. The answer to both these questions is: through empirical observations and analysis . What is done is that, first, data on demographic characteristics and trip -making behaviour of a large number of households are collected. This data is then analyzed to see what characteristics of the households are important in defining a homogeneous group -- the households which produce approximately the same number of trips. Based on the above analysis, tables are made which define each category of households by listing its properties in terms of different demographic variables. For example, a particular category of households may be defined as households with 3 to 4 members in the age group 6 to 60, with income in the range of Rs. 30,000 to Rs. 40,000 per month, and one automobile. Finally, for each category of household the average number of trips generated are listed. The listing of the definition of categories and the associated trip -generation rates are generally referred to as trip tables. This method of analysis although simple in its structure has few difficulties. The foremost is the problem with defining categories correctly -- at best it is very difficult. There are other problems like handling additional data on trip-generation behaviour -the trip tables are not amenable to simple updating but generally have to be completely revamped every time new data is available.
Regression model In this model an additive functional form is assumed to exist between the factors which affect trip-generation and the number of trips generated. Generally, a linear function of the following form is used: (2) where, are parameters of the regression function and is the value of the variable (such as income, automobile ownership, number of members in a household, and the like) for the i th zone. As can be seen using this model to determine the number of trips generated by a zone is a simple matter when all the parameters of the regression function are known. Obviously, the parameters are determined by using some parameter estimation technique like Ordinary Least Squares or Maximum Likelihood Technique on empirically obtained data on variables and . For a good description of regression analysis and the parameter estimation techniques mentioned here, one may refer to any book on introductory statistical methods or basic econometrics; for example, Gujarati [95] and Wonnacot and Wonnacot [263]. Discussion: Generally the models of trip-generation include variables which reflect the number of potential trip makers and the propensity of potential trip-makers to make a trip. However, none of the present models incorporate variables which reflect the accessibility factor. This is possibly the single largest factor as to why trip-generation models cannot very well predict the number of trips generated The trip-distribution models strive to predict the number of trips that will be made between a pair of zones for a particular trip-purpose. These models try to mathematically describe the destination-choice phase of the sequential demand analysis procedure. There are various models of trip distribution. However, most of them incorporate the same basic factors which affect the number of trips between an origin zone and a destination zone. The models differ in their characterization of these factors and in the way these factors are assumed to affect the trip distribution. The factors (for any given trip-purpose) which affect the number of trips between two zones are:
The number of trips produced by the origin zone. The degree to which the in-situ attributes of the destination zone attract trip makers. The attributes which gain importance vary with the trip purpose. For example, if one is modeling the number of shopping trips attracted to a zone then the type of attributes of the zone which assume importance will be the total shopping floor area, number of retail outlets, and the like. On the other hand, if one is modeling the number of work trips attracted to a zone then the type of attributes of the zone which assume importance will be the number of offices, the type of offices, and so forth. The factors that inhibit travel between a pair of zones. These factors could be, travel time, travel distance, travel cost, and so on.
Four different models are presented here. These are (i) Gravity model, (ii) Intervening opportunities model, (iii) Choice model, and (iv) Entropy Model.
Gravity model The Gravity model uses the following basic form to determine the trips between an origin zone i and a destination zone j. That is, (3)
tij = where, Ti is the total number of trips being produced by origin zone i aj is a set of in-situ attributes of the destination zone j which attract trip makers dij is a set of factors which inhibit travel between two zones is a calibration constant and Note that are positive monotonically increasing functions. is basically the trip attractions of Zone j.
The expression may be thought of as a factor which distributes the total trips produced by a zone among all the possible destination zones. In this sense, the sum of the expression over all destinations should be equal to unity. Thus, (4) The above equation implies that, (5) Substituting, the expression for in Equation 3, the following relation, often referred to as the origin constrained gravity model, is obtained: (6) Sometimes, in the gravity model, the number of trips attracted to a zone (when such data is independently available), , is used as a surrogate for . When such a substitution is done the gravity model is typically written as (7) Very often, in this case the following two constraints are imposed on the gravity model to obtain the two calibration constants.
and
are
(8)
(9)
Obviously, using this model requires an iterative solution technique as and vice versa. Example
depends on
Consider the following six-zone model of a town. Zones 1, 2, and 3 are fully residential areas and Zones 4, 5 and 6 are purely shopping areas. The shopping areas, shopping trips attracted (per day), the shopping trips produced (per day) and the travel distances are as shown in Table 1. The cells which have a ``-'' imply that those data are irrelevant to the problem. Determine the trip distribution between the zones for the following different scenarios: (a) Use the origin-constrained gravity model, assuming to be a linear function of the shopping area (in square meters) with a slope of 0.01 and constant term of 10. Also assume to be dij2 where dij is the distance in km. (b) Use the origin-destination constrained gravity model with same relevant assumptions same as those in (a). Table 1 : Data for the example on Gravity Model Zone 1 2 3 4 5 6 Shop Area ( 1000 2000 3000 ) Trips Produced 1000 1000 2000 Trips Attracted 800 2000 1200 Distance (km) to 1 4 2 7 2 3 1 6 3 5 2 6 4 4 3 5 5 2 1 2 6 7 6 6 -
Solution (a) The trips of interest here are , where and . Also note, as per the . The trips
problem description f(aj) = 0.01 x (shopping area in m2) + 10, and are given by Equation 6.
(b) In this case, the only difference from (a) is that the trips are given by Equation 7 and the constants of Equation 7 are given by Equations 8 and 9. The initial value of Kig is assumed to be the square root of the corresponding values obtained in (a). These values of Kig are used in Equation 9 to obtain a set of Kig values which are in turn used in Equation 8 to obtain a new set of Kig values. The process continues till all the and values converge. The final values of Final values of , and Final values of , Using these values of , and , and , and and and and and and and in Equation 7, the following tij values are obtained: and and obtained are as follows:
Discussion: It may be pointed out here that the expression in Equation 6 may be viewed as , the probability that destination j is chosen from origin i. Once such a view is taken, then the Maximum likelihood technique can be used to
estimate the parameters of the gravity model. If observations on the trip distribution between different origin-destination pairs are made and denoted as , then the likelihood function can be written as (10) Other constraints like could be accounted for by adding the constraint to the likelihood function and constructing a Lagrangian The postulate of travel behaviour on which this model is based, from Stouffer [221], is that the probability of choice of a particular destination (from a given origin for a particular trip purpose) is proportional to the opportunities for trip-purpose satisfaction at the destination and inversely proportional to all such opportunities that are closer to the origin. The inverse proportionality to closer opportunities can be interpreted as proportionality to the probability that none of the closer destinations (opportunities) are chosen. Thus, in this model, the in-situ attractive properties of the destination are modeled as opportunities and the impedances are measured in terms of the number of opportunities which are closer. In order to formulate the postulate as a mathematical model, the following notation are used: L: Constant of proportionality (or calibration constant). J : Total number of destinations. j: Index indicating the position of a destination (from the given origin), j =1 for the nearest destination and j = J for the farthest. :Cumulative function of opportunities up to and including the origin i. destination from
: Probability that a destination is chosen from the ith origin by the time the jth destination is reached. ij: Probability that the jth destination is chosen from the ith origin. Based on the postulate, the following expression can be written: (11) This relation states that a small addition in opportunities after the jth destination causes a correspondingly small increase in the probability . This increase is proportional to the probability that no destination is chosen by the jth destination and the amount of additional opportunities. Using we get (since it is assumed that ) and solving the differential equation (12)
and since, by definition, , we have (13) Note, however, nothing guarantees that . In fact, could be interpreted as the probability that no destination is chosen. This, however, poses a problem in the sequential demand analysis framework because if no destination is chosen then it leads to a situation where a trip is produced but does not go anywhere. In order to correct this anomaly, the ij obtained in Equation 12 is modified by adding a constraint which states that a destination must be chosen from the set of available destinations thereby precluding the possibility of not choosing any destination. Mathematically this can be done in one of two ways, both leading to the same result. We could use the Bayes theorem and determine the probability that destination j is chosen given that a destination is chosen or we could simply normalize the ij's obtained in Equation 13 by dividing the individual ij's with modified ij, written as , is given by, . In either case the
(14) Note that the denominator can be looked upon as either (i) the probability a destination is chosen, P(J)or (ii) Once the . (The reader should verify these claims).
are obtained the trips between i and j can be obtained as: (15)
Example For the 1200 shopping trips from Zone A, three destinations exist. The destinations are Zones X, Y, and Z. The shopping areas available in each of the zones and their distances from Zone A are given in Table 2. Assuming a proportionality constant of 0.35 and assuming a thousand square meter of shopping area as one opportunity determine the trip distribution from Zone A. Table 2 : Data for the example 2 on intervening opportunities model Zone Shopping area (in '000 sq.m.) X Y Z 2.0 4.0 2.0 Distance from Zone A (in km) 7.0 12.0 4.0
Solution First, the destinations should be arranged from the closest to the farthest; in this case, j = 1 for Zone Z, j = 2 for Zone X, and j = 3 for Zone Y. Since there are only 3 destinations, j = 3.
, and
and
Discussion: Since the model gives the probability of choosing a destination from a given origin we may employ the maximum likelihood technique to calibrate the constant L. The likelihood function is similar to that given in Equation 10. However, it is seen that assumption of a constant L is not very good for most problems. A varying L adds complexity to the model and its calibration. For a more detailed discussion on intervening opportunity models one may see Kanafani [132] or Stouffer [221] or Ruiter [197]
Destination Choice Models The postulate of travel behaviour on which this model is based is that destination j will be chosen from origin i for a particular trip-purpose if the perceived utility derived from choosing j is greater than the perceived utility derived from choosing any other destination. It is further assumed that, , the utility derived from destination j by an individual (or a group of similar individuals) in origin i has two components: (i) the deterministic component, and (ii) the stochastic component, . The deterministic component is the approximate utility that can be obtained from the destination, given the destination's in-situ attributes and the impedances. On the other hand, the stochastic component can be thought of as an approximation of the random variability assumed to be present in the utility because of the fact that this is a quantity which is perceived by humans. Mathematically, (16) Since is a stochastic quantity, the answer to the question as to which destination provides a greater utility will be probabilistic. Keeping this in mind the basic postulate of travel behaviour is modified thus: the probability that destination j will be chosen from origin i for a particular trip-purpose is equal to the probability that the perceived utility derived from j is greater than the perceived utility derived from each of the other destinations. Mathematically, (17) That is,
Once ij are obtained, the trip distribution between any i and j can be obtained by multiplying Ti by ij. It is, however, clear from the above equations that the exact nature of ij will depend on the assumptions about the nature of the 's. If it is assumed that (i) 's are distributed identically for each k, (ii) that they are independent, and (iii) that they are distributed according to the Gumbel distribution, i.e. (18) then the resulting expression for ij can be obtained as follows:
then Now realizing that when k = j, then we can rewrite Equation 18 as (19) Now substituting or (21) The form for ij given in Equation 21 is known as the multinomial logit model. Example For the data given in the Example 2 and assuming the trip distribution using the multinomial logit model. Solution For the given data, , , and and . Hence, the , determine and , we obtains (20)
Discussion: Instead of assuming a Gumbel distribution for the random terms, if it is assumed that they are distributed normally then the resulting form for ij is called the Probit model. This model is, however, a lot more cumbersome than the Logit model.
For a detailed and extensive discussion on the Probit model one may refer to Kanafani [132]. Another point which may be mentioned here is about the calibration of destination choice models. The parameters which need to be calibrated are the parameters of the function . Since the choice models give the probability of choosing a destination from a given origin, we may use the maximum likelihood estimation technique to estimate the parameters. Note that the likelihood function will the same as that given in Equation 10. The Entropy model of trip distribution, unlike the other models discussed here, is not a behavioral model. That is, the entropy model does not strive to predict the trip distribution by modeling the human behavioral aspects related to choosing a destination. This model, on the other hand, attempts to determine a distribution of trips which is most likely to occur assuming that each trip occurs independently of another. For a total number of Q trips in a network, it is assumed that there are a total of Q independent decisions. Obviously, a given trip-distribution matrix can be obtained through different combination of the decisions. Specifically, the number of ways a tripdistribution matrix [tij], can be obtained is: (22) In the entropy models it is assumed that the likelihood of a particular trip-distribution matrix occurring is proportional to the number of ways the matrix can be obtained. That is, the likelihood of a matrix occurring is proportional to the expression given in Equation 22. The entropy models determine the estimates of t ij by determining that set of tij's which maximizes the likelihood of a trip-distribution matrix occurring. Mathematically, the problem, therefore, is to solve the following (23)
Since Q is a constant, this is equivalent to solving (24) or (25) Using the approximation , the above is equivalent to solving (26) Since , this is equivalent to solving (27) The unconstrained optimization problem in Equation 27 is the standard maximum entropy model of trip distribution. However, we can easily incorporate various constraints in this model. Some of the constraints which are often included in the maximum entropy model are , , observed average trip length, and so forth.
Discussion: The constraints can be easily be incorporated in the function to be maximized by constructing a Lagrangian. The Lagrangian can then be differentiated with respect to the tijs and and the Lagrange multipliers to obtain a set of expressions which can be equated to zero. These equations when solved will give the estimate of the trip distribution. Also note that this method does not use any calibration constants and hence the issue of calibrating the model does not arise. Modal Split Model Modal split models aim to determine the number of trips on different modes given the travel demand between different pairs of nodes (zones). These models try to mathematically describe the mode choice phase of the sequential demand analysis procedure. Generally, choice models are used for modal split analysis. That is, it is assumed that the probability of choosing a particular mode is the probability that the perceived utility from that mode is greater than the perceived utility from each of the other available modes. Since, choice models were discussed while presenting destination choice models in the section on trip distribution they are not repeated here. This lecture only discusses the factors which are generally assumed to affect the perceived utility of modes. An example problem is also solved. The factors which affect the choice of a mode (and hence the perceived utility from a mode) are:
Socio-economic factors like income, automobile ownership, age, and so on. Service-related factors like in-vehicle travel time, access to public transport (or transit systems), frequency of transit system operation, out-of-pocket cost, and the like.
Example For a particular zone pair, three modes of travel between the zones exist -- private transport like automobiles (PT), bus (B), and urban rapid transit system like local trains (RT). It is given that all trip-makers have access to private transport and that the perceived utility of a mode m, i.e. , is given by
where, is the in-vehicle travel time in minutes for mode m is the out-of-pocket cost in rupees for mode m is the waiting time in minutes for mode m , and is a dummy variable which is 1 when the mode is private transport, 0 otherwise. Assuming that the variable values are as shown in Table 3 and that 1000 trips are made from the origin zone to the destination zone determine the number of trips made by the different modes. Use Logit model. Table 3: Variable values used in Example 4 on modal split models.
Mode
PT B RT
65 75 25
Hence, 302 trips will be made using private transport, 324 will use buses, and 374 will use the rapid transit system Demand Analysis II Traffic Assignment Models Traffic assignment models aim to determine the number of trips on different links (road sections) of the network given the travel demand between different pairs of nodes (zones). These models try to mathematically describe the route choice phase of the sequential demand analysis procedure. There are various models of traffic assignment. All of these models assume that travel time on the link is the only factor which trip makers consider while choosing a route. These models, however, differ in their assumptions regarding the variation in link travel times with the link volume (or link flow). In this section, three models are discussed, namely, (i) all-or-nothing assignment model, (ii) incremental assignment model, and (iii) user-equilibrium model. The notation used (other than those used in the earlier sections) in describing these models is as follows: : Flow (or volume) on link . : Flow (or volume) on link as estimated in the kth iteration.
:Travel time on link a when flow on link a is xa tij : The total demand (or number of trips) between origin i and destination j. : Indicator variable which is 1 if link a is apart of the kth route between origin i and destination j; otherwise it is zero. : Flow between i and j which uses the kth route n this model it is assumed that (i) the travel time on links do not vary with link flows, i.e. and (ii) all trip-makers (users) have precise knowledge of the travel time on the links. Based on these assumptions about travel times and the postulate that a trip maker will choose that path (or route) which minimizes his / her travel time this assignment model assigns all the trips between a particular origin and destination pair to that route (or path) which offers the minimum travel time. The exact nature of the assignment model is presented through the following algorithm. Step 1: For every pair (i.e., origin-destination pair) with , determine the minimum travel time path (or route) using as the link travel times. The minimum path determination can be done using any of the various existing algorithms like Flyod's algorithm or Djkastra's algorithm. Detailed description of these algorithms can be found in Teodorovic [233] or any other book on theory of networks. Also initialize all . Step 2: Set iteration counter . Select a particular pair. Step 3: Assign the entire tij to the minimum path between the pair. If link a is a part of the minimum path set, Step 4: If (where N is the total number of pairs with tij > 0 ) then report else set
In this model it is assumed that (i) the travel time on links do not vary with link flows, i.e. and (ii) all trip-makers (users) have precise knowledge of the travel time on the links. Based on these assumptions about travel times and the postulate that a trip maker will choose that path (or route) which minimizes his / her travel time this assignment model assigns all the trips between a particular origin and destination pair to that route (or path) which offers the minimum travel time. The exact nature of the assignment model is presented through the following algorithm. Step 1: For every pair (i.e., origin-destination pair) with , determine the minimum travel time path (or route) using as the link travel times. The minimum path determination can be done using any of the various existing algorithms like Flyod's algorithm or Djkastra's algorithm. Detailed description of
these algorithms can be found in Teodorovic [233] or any other book on theory of networks. Also initialize all . Step 2: Set iteration counter . Select a particular pair. Step 3: Assign the entire tij to the minimum path between the pair. If link a is a part of the minimum path set, Step 4: If (where N is the total number of pairs with tij > 0 ) then report . Else, select another i - j pair; set k = k + 1 and go back to Step 3. Example For the network shown in Figure 5 and the trip distribution matrix given in Table 4 determine the link flows using the all-or-nothing assignment technique. Note that the numbers on the links of the network denote the travel times and the numbers in the circles denote the zone numbers. as xa. else set
Figure 5: Network for example problem on all-or-nothing assignment technique. Table 4: Trip distribution matrix (O-D matrix) for the example problem on all-or-nothing assignment model. Origin zone 1 2 3 4 5 Solution Note there are 25 possible zone pairs out of which 9 have t ij = 0. Hence N = 16. Step 1: Destination zone 1 0 0 200 100 150 2 0 0 300 300 50 3 200 300 0 100 100 4 100 300 100 0 0 5 150 50 100 0 0
The minimum path for the 16 zone pairs (obtained using Djkastra'a algorithm) are as follows: i - j pair 1-3 3-1 1-4 4-1 2-4 4-2 3-4 4-3 Min. path 1 > 3 3 > 1 1 > 3 > 4 4 > 3 > 1 2 > 3 > 4 4 > 3 > 2 3 > 4 4 > 3 i -j pair 1-5 5-1 2-3 3-2 2-5 5-2 3-5 5-3 Min. path 1 > 3 > 5 5 > 3 > 1 2 > 3 3 > 2 2 > 3 > 5 5 > 3 > 2 3 > 5 5 > 3
Step 2: k = 1. Consider the zone pair 1 - 3. Step 3: Step 4: Since , set back to Step 3. Step 3: Step 4: Since , set k = 3 and select zone pair as the next pair and go back to Step 3. ; the rest of the remain zero.
remain zero.
In this manner, Steps 3 and 4 are repeated till all the zone pairs are chosen (i.e., k = 16). Finally, the following assignment is obtained. = 450 = 650 = 0 = 450 = 650 = 550 = = = = = = 300 0 500 300 0 0
User-equilibrium model
The user-equilibrium model of traffic assignment is based on the fact that humans choose a route so as to minimize his / her travel time and on the assumption that such a behaviour on the individual level creates an equilibrium at the system (or network) level. Flows on links (whose travel times are assumed to vary with flow) are said to be in equilibrium when no trip-maker can improve his/her travel time by unilaterally shifting to another route. This notion of equilibrium flows is generally referred to as Wardrop's principle. Before presenting the model which can determine such equilibrium flows on a
network, the idea of equilibrium flows or the concept of user-equilibrium needs to be explained further. Consider, the example network shown in Figure 7(a). Figure 7(b) gives the travel time function for each of the three single link routes between the origin O and destination D. The total demand from O to D is 110.
Figure 7 : Example problem on user-equilibrium assignment technique. In the example, obviously if demand is less than or equal to 100, everybody will travel using Route 2 since the travel time offered by Route 2 (between 30 minutes and 40 minutes) will be less than any other route. However, when demand is in excess of 100, if everybody travels on Route 2, then the travel time on Route 2 will become more than 40 min, implying that Route 1 will become more lucrative (offering about 40 min of travel time, since the volume on Route 1 is zero at this stage) and some can benefit by unilaterally shifting to Route 1. On doing so, if the the volume carried by Route 2 falls below 100 then again Route 2 will become lucrative and some will shift to Route 2 from Route 1. Given this, and the fact the total demand is 110, it can be said that if 100 use Route 2 and 10 use Route 1 then each one will face a travel time of about 40 min and none can improve his/her travel time by unilaterally shifting from one route to another. This then will be the equilibrium flow to which the network has to eventually converge given that every traveller always tries to minimize his/her travel time. Note that under the equilibrium condition, Route 3 is not used as it entails more than 40 min of travel time. The above example, in addition to clarifying the concept of equilibrium, also illustrates the fact that at equilibrium all routes which are used between a given origin and destination offer the same travel time which is less than all routes which are not used. Another interesting feature is that at equilibrium flow, the sum of the areas under the travel time versus flow plots is the least. This feature can be easily seen by visually inspecting the sum of the areas shown in Figure 7. Figure 7 (a) corresponds to the equilibrium flow, and the sum of the areas is equal to area ACDE + area ABFG. Figure 8 (b), corresponds to an instance of non-equilibrium flow conditions with Route 2 carrying all the flow (i.e., 110); in this case the relevant sum of the areas is simply equal to area HIJK. Figure 8(c) represents another instance of non-equilibrium flow where
Route 2 carries a flow of 90 and the rest 20 are carried by Route 1; in this case the relevant sum of areas is area LMNO + area LPQR. Finally, Figure 8 (d) represents another instance of non-equilibrium flow conditions with Routes 1, 2, and 3 carrying flows of 30, 70, and 10, respectively; the relevant sum of the areas in this case is area TUVW + area TXYZ + area T. Note that the sum of the areas is the least in (a) -- the case representing equilibrium flow. The interested reader may try to verify this feature by constructing other such examples or by proving them mathematically.
As the name suggests, shortest path algorithm find the shortest path between two nodes in a network. In the case of transportation networks, the nodes may represent points (or locations) from which traffic is produced or to which traffic is attracted. The nodes also represent intersections. The links represents roads or movements. The above description of a transportation network is a very simplistic viewpoint. Interested readers may refer to books on transportation network analysis to gain a insight of how a real-world transportation facility may be represented as a network. Figure 1 shows a typical network representation of a transportation facility.
Figure1: A typical transportation network. Two shortest path algorithms are most commonly used in transportation network analysis. One is Dijkstra's algorithm and the other is Floyd's algorithm. In the following these algorithms are described in details. The descriptions are largely adopted from Dusan Teodorovic's book on "Transportation Networks: Aqualitaive treatment". DIJKSTRA'S ALGORITHM This algorithm is one attempt to find the shortest path from one node to all other nodes in the network. It assumes that the link lengths are always nonnegative. In this method, every node is assigned a label with two components ( x, y). A label could either be temporary or permanent. The algorithm stops when all labels are permanent. As will soon become apparent, after completion, the labels give information on the shortest distances as well as the shortest paths from a particular node to all the other nodes. Also a node is referred to being in the open state if its associated label is temporary; it is to be in the closed state if the label is permanent. Before proceeding further, some of the notation used here are presented. l(i, j) : length of the link joining node i to node j. a : node for which we are investigating the shortest paths to all other nodes.
dai : the shortest known path from node a to node i found in the network, so far. qi : the immediate predecessor node of node i on the shortest known path from node a to node i found so far. c : the last node to have moved to being in the closed state. x : x = dai y : y = qi The Dijkstra algorithm is comprised of the following 5 steps: Step 1:The process starts from node. Since the length of the shortest path from node a to node a is 0, then daa = 0. The immediate predecessor node of node a will be denoted by the symbol + so that qa = +. Since the lengths of the shortest paths from node a to all other nodes on the shortest path are unknown, we put qi = - for all . The only node which is now in a closed state is node a. Therefore we write that c = a. Step 2: In order to transform some of the temporarly labels into permanent labels, we examine all branches (c, i) which exit from the last node which is in a closed state (node c). If node i is also in a closed state, we pass the examination on to the next node. If node i is in an open state we obtain its first label dai based on equation:
in which the left side of the equation is the new label of node i. We should note that dai appearing on the right side of the equation is the old label for node i. Step 3: In order to determine which node will be the next to go from an open to a closed state, we compare value dai for all nodes which are in an open state and choose the node with the smallest dai. Let this be sdme node j. Node j passes from an open to a closed state since there is no path from a to j shorter than daj. The path through any other node would be longer. Step 4: We have ascertained that j is the next node to pass from an open state to a closed one. We then determine the immediate predecessor node of node j and the shortest path which leads from node a to node j. We examine the length of all branches (i, j) which lead from closed state nodes to node j until we establish that the following equation is satisfied :
Let this equation be satisfied for some node t. This means that node t is the immediate predecessor of node j on the shortest path which leads from node a to node j. Therefore, we can write that qj = t. Step 5: Node j is in a closed state. When all nodes in the network are in a closed state, we have completed the process of finding the shortest path. Should any node still be in an open state, we return to step 2.
The algorithm described above can also be ussed to find the shortest path between two specific nodes. In this case, the algorithm is completed when both nodes are in a closed state. Example: Using the Dijkstra algorithm, calculate the shortest path from node a to all other nodes on the transportation network shown in Figure Y1. Figure Y1
The number next to the network branches shown on Figure Y1 signify the length of each branch. (We should mention that the length of a branch in a network could also, in addition to actual length, stand for travel time or travel expenses or numerous other values). We start the process of finding the shortest path at node a. Since the length of the shortest path from node a to node a is 0, then daa = 0. The immediate predecessor node of starting node a is denoted by the symbol + so that qa = +. The length of all other shortest paths from node a to all other nodes are for the present unexamined, so for all other nodes we put dai = . Since the immediate predecessor nodes of nodes on the shortest path are unknown, we put qi = - for all . The only node which is now in a closed state is node a. Therefore, c = a. Next to the node a symbol we put the label (0 , +) and add the symbol ' to emphasize that node a is in a closed state. Ths completes the first step of the algorithm. After the first step, the transportation network looks like Figure Y2.Figure Y2.
We now move to the second step of the algorithm. By examining the length of all branches leaving node a which are in a closed state we can write:
In step 3 we determine which node will be next in line to pass from an open to a closed state. Since dab< dad node b passes from an open to a closed state. In the same manner, since:
we can conclude in step 4 that node a is the immediate predecessor of node b on the shortest path, i.e. qb = a. Now in step 5 we note that there are still many nodes which are in an open state. Therefore, we have to return to step 2. The transportation network looks like Figure Y3 after going through all five steps of the algorithm for the first time:
Figure Y3
Let us now return to the second step of the algorithm. The last node to go from an open to a closed state was node b. This means that c = b. When we examine all branches leaving node b going towards nodes which are in an open state, we have:
Since dac< daf node c is the next to switch from an open to a closed state. We then determine the immediate predecessor node of node c. Since :
then node b is the immediate predecessor of node c and qc = b. The network still contains many nodes in an open state, so we must once again return to step 2 of the algorithm. After the second time through all 5 steps of the algorithm, the transportation network looks like Figure Y4.
Since dad< dae node d is the next node to switch to a closed state. Since:
then node a is the immediate predecessor of node d on the shortest path, so qd = a. And now c = d. The transportation network looks like Figure Y6 after the third time through the algorithm: Figur Y5
Table 2 shows the calculations leading to the shortest paths and determination of the immediate predecessor nodes after going through the algorithm 11 times.
Table 2: Calculating the shortest paths and predecessor nodes. Times Last node Branches from last closed node to open Next node in through in closed nodes closed state algorithm state dai = min [dai, dac + 1 (c, i)] 4 d dae = min [8, 12] = 8 dag = min [, 14] = 14 e Predecessor node qe = c
5 6
e f
daf = min [13, 15] = 13 dah = min [, 23] = 23 dai = min [, 22] = 22 dag = min [14, 17] = 14
f g
qf = b qg = d
7 8
g i
dai = min [22, 21] = 21 dah = min [23, 27] = 23 daj = min [, 29] = 29 dah = min [, 27] = 27
i h
qi = g qh = f
qj = h
10
dak = min [27, 30] = 27 dal = min [, 30] = 30 dal = min [30, 30] = 30
qk = i
11
ql = j
The shortest paths from node a to all other nodes in the network are shown in Figure Y7 with immediate predecessor nodes noted after going through the algorithm 11 times. Figur Y7
Note that given all the predecessor nodes the shortest path from node a to any particular node can be easily reconstructed. For example, the shortest path from a to j can be reconstructed in the following way: From Figure Y7 it can be seen that the immediate predecessor node of node j is node h. Hence "h to j" is a part of the shortest path to j. Similarly "f to h" is a part of the shortest path to h from a. Hence, "f to h to j" is a part of the shortest path to j. Proceeding similarly we can see a to b to f to h to j is the shortest path from a to j. Using similar logic for all other nodes the shortest paths from a to all other nodes can be diagramatically represented as shown in Figure Y8. Figure Y8
Figure 8: Areas under travel-time versus flow plots under various flow distributions. The user-equilibrium model formulation is based on this area-related observation about equilibrium flows and the fact that the travel time on a route is simply a linear sum of the travel times on the constituent links. The formulation, which is a non-linear programming problem is given in Equation 28.
subject to =
(28)
where w is a variable denoting flow. Nonlinear programming problems of this kind (with linear constraints and a nonlinear objective function) can be effectively solved using convex combination algorithms. The interested reader may refer to Taha [225] for more details on convex combination algorithms. Here, only the principle is given so that the latter discussions on solving user-equilibrium problems are more meaningful to the reader. Principles of convex combination methods for solving optimization problems with nonlinear objective function and linear constraints The convex combination algorithm determines the optimum point by proceeding in the following iterative manner.
First, a feasible solution (i.e., a solution which satisfies all the constraints), say , is determined. Next, another feasible solution, say , is determined. Since all the constraints are linear (thereby giving rise to a convex solution space), it can be shown that any point on a straight line joining and (note ; ; where ) is also feasible. The convex combination algorithm then determines the value of which gives the best value for the objective function. Note that determination of the best value of , say *, requires an unconstrained optimization of the objective function written in terms of . Once the value of * is determined, the point is precisely identified by using the value of * in place of in the earlier equation for . Next is set equal to ; also another feasible point is determined and is replaced with the new feasible point. The process goes back two steps and continues repeatedly till the objective function value changes negligibly from one iteration to the next. The optimal solution is obtained in the last iteration.
Solving the mathematical programming formulation of the user-equilibrium model As stated earlier, the convex combination algorithm is used to solve the mathematical programming formulation of the user-equilibrium model. Here the procedure is presented step-wise. For a better understanding of the procedure, parallels are drawn with the steps presented in the earlier section. Step 1: Set iteration counter . Determine a feasible solution by performing all-ornothing assignment with as the travel time on link a. Note that the feasibility conditions of the programming formulation will be met by performing the all-ornothing assignment. Let this solution vector be . Step 2: Update the travel time of the links using the link volumes obtained in the . Step 3:
With the updated travel time information, perform an all-or-nothing assignment to obtain another feasible solution. Let this solution vector be . Step 4: Determine by finding that value of which minimizes the following quantity
where
, indicates the flow value for link ' a' as per the solution
vector . Step 5: Obtain the new feasible point, . Step 6: Check whether the value of the objective function (given in Equation 28 or in Step 4 here) has changed significantly. If it has not, then stop and report the solution. Otherwise, set n = n + 1 and continue from Step 2 onwards. Given example illustrate how the solution procedure suggested here works . Example For the network shown in Figure 9, and 1000 trips per day from Node A to Node B determine the link flows using the user-equilibrium assignment technique. The link travel times, , are given by . The link number, the value, and the value for a particular link are mentioned as on the links. Note that in this case each route has one link. as
Figure 9: Road network for the example on user-equilibrium method of traffic assignment. Solution Table 6 provides the output of every step at each iteration. As per the table the flow on link a, , is: , , and . Note that all links which are used have approximately the same travel time and this is less than the travel time offerred by the unused link (in this case Link 1). Table 6: Solution of the example on user equilibrium model of traffic assignment. n Step Links 1 2 3 4 Objective function value
1 1
35 0
10 1000 947 0
20 0 20 1000
25 0 25 0 1975.0
2 3 4 5 6 2 2 3 4 5 6 3 2 3 4 5 6 4 2 3 4 5 6 5 2 3 4 5 6 *, n *, n *, n *, n *, n
35 0
As the name suggests, shortest path algorithm find the shortest path between two nodes in a network. In the case of transportation networks, the nodes may represent points (or locations) from which traffic is produced or to which traffic is attracted. The nodes also represent intersections. The links represents roads or movements. The above description of a transportation network is a very simplistic viewpoint. Interested readers may refer to books on transportation network analysis to gain a insight of how a real-world transportation facility may be represented as a network. Figure 1 shows a typical network representation of a transportation facility.
Figure1: A typical transportation network. Two shortest path algorithms are most commonly used in transportation network analysis. One is Dijkstra's algorithm and the other is Floyd's algorithm. In the following these algorithms are described in details. The descriptions are largely adopted from Dusan Teodorovic's book on "Transportation Networks: Aqualitaive treatment". DIJKSTRA'S ALGORITHM This algorithm is one attempt to find the shortest path from one node to all other nodes in the network. It assumes that the link lengths are always nonnegative. In this method, every node is assigned a label with two components ( x, y). A label could either be temporary or permanent. The algorithm stops when all labels are permanent. As will soon become apparent, after completion, the labels give information on the shortest distances as well as the shortest paths from a particular node to all the other nodes. Also a node is referred to being in the open state if its associated label is temporary; it is to be in the closed state if the label is permanent. Before proceeding further, some of the notation used here are presented.
l(i, j) : length of the link joining node i to node j. a : node for which we are investigating the shortest paths to all other nodes. dai : the shortest known path from node a to node i found in the network, so far. qi : the immediate predecessor node of node i on the shortest known path from node a to node i found so far. c : the last node to have moved to being in the closed state. x : x = dai y : y = qi The Dijkstra algorithm is comprised of the following 5 steps: Step 1:The process starts from node. Since the length of the shortest path from node a to node a is 0, then daa = 0. The immediate predecessor node of node a will be denoted by the symbol + so that qa = +. Since the lengths of the shortest paths from node a to all other nodes on the shortest path are unknown, we put qi = - for all . The only node which is now in a closed state is node a. Therefore we write that c = a. Step 2: In order to transform some of the temporarly labels into permanent labels, we examine all branches (c, i) which exit from the last node which is in a closed state (node c). If node i is also in a closed state, we pass the examination on to the next node. If node i is in an open state we obtain its first label dai based on equation:
in which the left side of the equation is the new label of node i. We should note that dai appearing on the right side of the equation is the old label for node i. Step 3: In order to determine which node will be the next to go from an open to a closed state, we compare value dai for all nodes which are in an open state and choose the node with the smallest dai. Let this be sdme node j. Node j passes from an open to a closed state since there is no path from a to j shorter than daj. The path through any other node would be longer. Step 4: We have ascertained that j is the next node to pass from an open state to a closed one. We then determine the immediate predecessor node of node j and the shortest path which leads from node a to node j. We examine the length of all branches (i, j) which lead from closed state nodes to node j until we establish that the following equation is satisfied :
Let this equation be satisfied for some node t. This means that node t is the immediate predecessor of node j on the shortest path which leads from node a to node j. Therefore, we can write that qj = t.
Step 5: Node j is in a closed state. When all nodes in the network are in a closed state, we have completed the process of finding the shortest path. Should any node still be in an open state, we return to step 2. The algorithm described above can also be ussed to find the shortest path between two specific nodes. In this case, the algorithm is completed when both nodes are in a closed state. Example: Using the Dijkstra algorithm, calculate the shortest path from node a to all other nodes on the transportation network shown in Figure Y1. Figure Y1
The number next to the network branches shown on Figure Y1 signify the length of each branch. (We should mention that the length of a branch in a network could also, in addition to actual length, stand for travel time or travel expenses or numerous other values). We start the process of finding the shortest path at node a. Since the length of the shortest path from node a to node a is 0, then daa = 0. The immediate predecessor node of starting node a is denoted by the symbol + so that qa = +. The length of all other shortest paths from node a to all other nodes are for the present unexamined, so for all other nodes we put dai = . Since the immediate predecessor nodes of nodes on the shortest path are unknown, we put qi = - for all . The only node which is now in a closed state is node a. Therefore, c = a. Next to the node a symbol we put the label (0 , +) and add the symbol ' to emphasize that node a is in a closed state. Ths completes the first step of the algorithm.
After the first step, the transportation network looks like Figure Y2. Figure Y2.
We now move to the second step of the algorithm. By examining the length of all branches leaving node a which are in a closed state we can write:
In step 3 we determine which node will be next in line to pass from an open to a closed state. Since dab< dad node b passes from an open to a closed state. In the same manner, since:
we can conclude in step 4 that node a is the immediate predecessor of node b on the shortest path, i.e. qb = a. Now in step 5 we note that there are still many nodes which are in an open state. Therefore, we have to return to step 2. The transportation network looks like Figure Y3 after going through all five steps of the algorithm for the first time: Figure Y3
Let us now return to the second step of the algorithm. The last node to go from an open to a closed state was node b. This means that c = b. When we examine all branches leaving node b going towards nodes which are in an open state, we have:
Since dac< daf node c is the next to switch from an open to a closed state. We then determine the immediate predecessor node of node c. Since :
then node b is the immediate predecessor of node c and qc = b. The network still contains many nodes in an open state, so we must once again return to step 2 of the algorithm. After the second time through all 5 steps of the algorithm, the transportation network looks like Figure Y4.
Since dad< dae node d is the next node to switch to a closed state. Since:
then node a is the immediate predecessor of node d on the shortest path, so qd = a. And now c = d. The transportation network looks like Figure Y6 after the third time through the algorithm: Figur Y5
Table 2 shows the calculations leading to the shortest paths and determination of the immediate predecessor nodes after going through the algorithm 11 times.
Table 2: Calculating the shortest paths and predecessor nodes. Times Last node Branches from last closed node to open Next node through in closed nodes closed state algorithm state dai = min [dai, dac + 1 (c, i)] 4 d dae = min [8, 12] = 8 dag = min [, 14] = 14 e in Predecessor node qe = c
5 6
e f
daf = min [13, 15] = 13 dah = min [, 23] = 23 dai = min [, 22] = 22 dag = min [14, 17] = 14
f g
qf = b qg = d
7 8
g i
dai = min [22, 21] = 21 dah = min [23, 27] = 23 daj = min [, 29] = 29 dah = min [, 27] = 27
i h
qi = g qh = f
qj = h
10
dak = min [27, 30] = 27 dal = min [, 30] = 30 dal = min [30, 30] = 30
qk = i
11
ql = j
The shortest paths from node a to all other nodes in the network are shown in Figure Y7 with immediate predecessor nodes noted after going through the algorithm 11 times. Figur Y7
Note that given all the predecessor nodes the shortest path from node a to any particular node can be easily reconstructed. For example, the shortest path from a to j can be reconstructed in the following way: From Figure Y7 it can be seen that the immediate predecessor node of node j is node h. Hence "h to j" is a part of the shortest path to j. Similarly "f to h" is a part of the shortest path to h from a. Hence, "f to h to j" is a part of the shortest path to j. Proceeding similarly we can see a to b to f to h to j is the shortest path from a to j. Using similar logic for all other nodes the shortest paths from a to all other nodes can be diagramatically represented as shown in Figure Y8. Figure Y8
When solving practical transportation problems, it is often necessary to calculate the shortest path between all pairs of nodes in a transportation network. If the network has n nodes, then the Dijkstra algorithm must be applied n times, taking a different node each time as trhe starting node. This is a lengthy process. Hence, Dijkstra's algorithm is rarely used to determine the shortest path between all pairs of nodes; instead Floyd's algorithm is used. The algorithm works by updating two matrices, namely Dk and Qk, n times for a n node network. The matrix Dk, in any iteration k, gives the value of the shortest distance (time) between all pairs of nodes (i, j) as obtained till the kth iteration. The matrix Qk has as its elements. The value of gives the immediate predecessor node from node i to node j on the shortest path as determined by the kth iteration. Do an Qo give the starting matrices and Dn and Qn give the final matrices for an n-node system. The first task is to determine Do and Qo. Do is taken up first. The element dij of matrix Do are defined as follows: If a link (branch) exists between nodes i and j the length of the shortest path between these nodes equals length l (i, j) of branch (i, j) which connects them. Should there be several branches between nodes i and node j, the length of the shortest path must equal the length of the shortest branch, i.e.:
whre m is the number of branches between node i and node j. It is clear that = 0 when i = j. In the case when there is no direct link between node i and node j, we have no information at the beginning concerning the length of the shortest path between these two nodes so we treat them as though they were infinitely far from each other, that is, =
Elements
First, we assume that = i, for , i.e. that for every pair of nodes (i, j) for , the immediate predecessor of node j on the shortest path leading from node i to node j is actually node i. After defining Doand Qo the following steps are used repeatedly to determine Dn and Qn. Step 1 : Let k = 1 Step 2 : We calculate elements of the shortest path length matrix found after the k-th passage through algorithm Dk using the following equation:
Step 3: Elements of predecessor matrix Qk found after the k-th passage through the algorithm are calculated as follows:
Step 4: If k = n, the algorithm is finished. If k < n, increase k by 1, i.e. K = k+1 and return to step 2. Let us now look at the algorithm in a little more detail. In step 2, each time we go through the algorithms we are checking as to whether a shorther path exists between node i and j other than the path we already know about which was established during one of the earlier passages through the algorithm. If we establish that , i.e. if we establish during the k -th passage through between nodes i and j
is less than the length of the shortest path known previous to the k-th passage, we have to change the immediate predecessor node to node j. Since the length of the new shortest path is:
it is clear that in this case node k is the new immediate predecessor node to j, and therefore:
This is actually done in the third algorithmic step. It is also clear that the immediate predecessor node to node j does not change if, at the end of step 2, we have established that no other new, shorter path exists. This means that: for
When we go through the algorithm n times (n is the number of nodes in the transportation network), elements of final matrix Dn will constitute the shortest path going from node i to node j. Example: Determine the shortest paths between all pairs of nodes on transportation network shown in Figure z1. Branch lengths are shown on the figure.
All elements along the main diagonal of matrix Do equal zero since by definition . We note element of matrix Do. This element equals 8 since
the length of the branch connecting nodes 1 and 2 is 8. Element equals infinity since the network has no branch which is oriented from node 3 to node 1. Element of matrix Do equals infinity as well since there is direct branch linking nodes 5 and 1. Starting matrix Qo is as follows:
First we note that node i is the immediate predecessor of node j on the shortest path leading from node i to node j (for ) . For this reason we have, for example:
We now go to the first algorithmic step. Let k = 1. As an illustration of Step 2 we calculate the elements of the first three rows of matrix D1. Calculations for other rows are left as an exercise.
Matrix D1 is as follows:
Matrix elements which changed values compared to the values they had in matrix DO are circled.
So, for example, the shortest distance between nodes 2 and 4 is 13 after the first algorithic step. In starting matrix DO this distance was . Since
then node 1 is the new immediate predecessor of node 4 on the shortest path from node 2 to node 4. After passing through the algorithm the first time, Q1 looks like this:
After the second, third, fourth and fifth passages through the algorithm, matrices D2, Q2, D3, Q3, D4, Q4 and D5, Q5 are as follows:
Matrices D5 and Q5 furnish us with complete information on the lengths of the shortest paths and the nodes on those paths between all pairs of nodes in the transportation network. For example the shortest paths from, node 5 to node 4 has a length of 10. Also by studying the Q5 matrix one can obtain the shortest paths from, node 5 to node 4. As per Q5, the immediate predecessor node to node 5 for the shortest path from 5 to 4 is . Hence 3 to 4 forms the final link on the shortest path from 5 to 4. Next the predecessor node to node 3 on the shortest path from 5 to 3 is Hence 2 to 3 to 4 forms a .
part of the shortest path from 5 to 4. Proceeding similarly, we see, Hence 5 to 2 to 3 to 4 is the shortest path.
Transportation Economics Definition of parameters Objectives The objective of this module is to introduce the principles of economic evaluation and financing of a highway project. This module also discusses briefly the present road scenario, its future growth and financing strategies for highway projects with reference to Indian context. Introduction Highway is a major Civil Engineering construction, which involves large sum of investment. Construction of a six-lane expressway may cost Rupees 5 to 10 crore per kilometer. A well designed rural road with 3.75m carriageway, shoulder and crossdrainage works may cost between Rs.6 lakhs to 25 lakhs, depending on the construction conditions (PMGSY 2001). A proposed road project is expected to have various alternatives in terms of its way of implementation. The overall expenditure involved, and the expenditure required at various stages of project execution would be different for different alternatives. A judicious selection of the best possible strategy, among various alternatives, necessitates understanding in economics of highway projects. Study of highway economics and finance involves understanding of various cost components of highway projects the economic feasibility of alternative highway projects decision on scheme of investment on a project at its various stages funding source and policies for road projects. In this module, brief discussions have been placed in six sections. The first section contains a general discussion on the scenario of Indian roads. Some of parameters used in economic analysis are defined in the second section. Various cost and benefit components of transportation infrastructure has been discussed in the third and the fourth section respectively. The fifth section discusses different approaches of economic evaluation of highway projects. Lastly a brief idea on highway financing, with reference Indian context, has been placed in the sixth section. India has 60.8 km of road length per 100 sq km, in USA it is 66.5km/100 sqkm. However, out of total road length of over 1.8 million-km (1996), only about 45% is surfaced (Bhatia 1996). Total road length in India comprised of Expressways, National Highways, State Highways, Major District Roads and Rural Roads is 33 lakhs kilometer (NHAI 2005), out of which National Highways covers a length of 65,569 km (NHAI 2005). On the existing road network National Highways carry 40% of the total traffic but constitute less than 2% of the total road length (NHAI 2005). At the time of independence, India's road network was rudimentary and poor even for the small traffic plying on it. At present, the length of road network has increased from 4 lakh kms in 1951 to 33.4 lakh-km in 2005, making India the second largest road network
in the world (NHAI 2005, MORT&H 2005). During this period, number of trucks and vehicles has increased 10 to 40 times, with consequent increase in share of passengers (26 to 85%) and goods (11 to 65%) traffic carried by road (MORT&H 2005). The number of automobiles has grown from 0.3 million in 1951 to 48.5 million by the year 2001 (Narain 1996). The freight traffic has also increased from 6 billion ton-km in 1951 to 400 billion ton-km to 1200 billion ton-km (2000). The present traffic growth on roads is 7 to 10% per year, and vehicle population growth is about 12% per year (MORT&H 2005) The newly conceived Golden Quadrilateral project involves 4-laning of road linking Delhi-Kolkata-Chennai-Mumbai , and the North-South and East-West corridors involves road linking of Srinagar-Kanyakumari and Silchar-Porbandar respectively (MORT&H 2005). The lengths and estimated dates of completion of National Highway Development Projects (NHDP) are presented in Table-1. Table-1 (MORT&H 2005) Project NHDP Phase-I (i) Quadrilateral Golden 5846 1133 (ii) Port Connectivity & others NHDP Phase-II (i) N-S Corridor (ii) E-W Corridor NHDP Phase-III Dec., 2005 Dec., 2007 Length (in km) Target date completion of National Highway Development Projects (NHDP)
7300 10,000
In order to provide road connectivity to the villages, Government of India has launched (in December, 2000) the Pradhan Mantri Gram Sadak Yojana (PMGSY), where all villages having population more than 1000 persons are provided connectivity through all-weather-roads. Connectivity will be further improved to the villages having population more than 500 persons by the end of year 2007 (PMGSY 2005). The estimated cost of this time bound project is approximately Rs.60,000 crore (PMGSY 2005) Time Horizon The investment for highway construction, maintenance, and its benefits are spread over a time span, and is called as the time horizon of economic assessment. This is generally selected as twenty to thirty years for a highway project, depending on policy or type of road. Time horizon, basically, is the economic analysis period. Interest Rate Money earns its interest intrinsically. Interest rate is the return obtained after the end of the year as percentage of the capital invested at the beginning of the year. Interest rate could be simple or a compound one. This interest rate is an extra cost charged to the highway project and is payable to the source where from the investment has been generated. This is because of the reason, that the money, which has been invested in highway project, could have been invested elsewhere to earn the same interest (Nash
1997). The interest rate is used to calculate what would be the amount of money at a future date. And the discount rate, on the other hand, is used to calculate the present equivalent amount of money, of the amount which will be actually invested in future (IRC:SP-30 1993). Inflation Construction of major highway project takes a number of years, and meanwhile the cost of material, labour, equipment etc. undergoes price escalation due to inflation. At the same time, due to inflation, the Vehicle Operating Cost (VOC) increases, thereby reducing the benefit. Thus in the benefit-cost considerations, the effect of inflation also needs to be considered (IRC:SP-30 1993) in all the cost and benefit components. Salvage value Salvage value ( S ) is the worth of the structure at the end of the analysis period. This value is carried over to the next analysis period. There could be different basis of calculating the salvage value. If, after the expiry of the first analysis period, it is assumed that the pavement materials would be recycled, then the cost of existing pavement materials (to be used for recycling) are considered in computation of salvage value. Alternatively, if the pavement life is extended further by putting overlay, in the next analysis period, the salvage value, S , could be calculated in the following way (MS-1 1999):
.(1) where, Y is the number of years between the last overlay (which is done in the year n m ) and the analysis period for which was the cost incurred and X is the number of years it is expected to actually serve. This is based on the assumption that the service life of the last resurfacing overshoots beyond the analysis period, and accordingly a proportionate salvage value is assigned. Present worth Present worth Present worth is the total cost of the project, when investments in various years (during the analysis period) are brought back to the equivalent worth of present year. The present worth can be expressed in the form of the following equation (Equation 2)
.(2) where, present C is the cost of construction, n 1 is the 1st year in which major maintenance (say, overlay) is done, n m last year within the analysis period in which maintenance job is carried out, O k is the cost of maintenance in the k th year, S is the salvage value and r is the discount rate. Capital recovery factor
The concept of capital recovery factor is used when only the recurring investments at different periods of time, is brought back to the equivalent investment put at the beginning of the project. This could be made clear by taking the example of the middle term of above equation (Equation 2), using the following simplifications: maintenance is done periodically in each year the maintenance expenditure is same always, say x , and the total maintenance expenditure calculated as the 1st year as the base year is equivalent to an one-time expenditure of y . Then,
.(4)
As a routine maintenance work, Rs.20,00,000/- each is to be spent on a particular stretch of a highway during the 3rd year, 5th year and the 7th year. Calculate the total present worth of these expenditures, if the annual discount rate is 12% (compound). Solution The present worth of the maintenance investment is
= Rs. 34,63,080/Example The expenditure of construction of new highway was estimated as Rs. 200 crore. It was decided that this money will be raised from loans. What is the equal amount of money to be paid each year, such that the loan is repaid by 15 years? Assume compound rate of interest as 10%.
Using the Equation 4, therefore, x = 26.294 crore Cost Components in Transportation System
mprovement of transportation system may, reduce the cost of vehicle operation on highways, cost of traffic congestion, travel time etc., but these savings are achieved as a return benefit of construction of new highway, maintenance of roads, construction of new flyover etc. which again involves cost. The cost of transportation system could be thought of constituted with cost incurred by the two parties viz. (i) agency cost and the (ii) user cost. Agency cost The agency cost is the cost incurred by the agency (government or private) for construction and maintenance of a highway facility. Construction cost The construction cost involves, the cost of (i) surveying, planning and design, (ii) acquisition of land, (iii) construction of highway, (iv) installation of traffic control devices, (v) cost of supervision, quality control and administrative costs, (vi) installation of other transportation facilities etc. Maintenance cost The maintenance cost involves the cost of planning and implementation of various maintenance measures implemented on the in-service roads in various time phases. The maintenance costs are cost of (i) periodic repair (ii) major rehabilitation (iii) operational expenditure of traffic facility, (iv) supervision and installation charges etc. User cost Unlike the case of determination of cost of construction, the evaluation of the cost incurred by the users is sometimes difficult to estimate. In some cases, it is not possible to measure also, for example, the pain and grief caused by accidents. Vehicle Operating Cost (VOC) The cost of owning and operating vehicle due to its use on roads is called the Vehicle Operating Cost (VOC). VOC has two components, variable cost and fixed cost. Variable costs may be subdivided into two categories: Distance related components, like, (i) fuel consumption, (ii) spare parts, (iv) tyre wear (v) lubricants (vi) maintenance labour cost etc. Time related components, like, (i) depreciation (ii) value of the passengers time (iii) wages of the crew etc. Fixed cost includes the capital cost, registration fees, insurance, road permit charges, road and other taxes. Cost due to traffic congestion and restraint Any renovation on existing pavement causes disruption of traffic in terms of congestion, reduction of speed, and even complete restraint of movement. This delay in travel time is a function of road geometrics, traffic volume, time and duration of road renovation work. The effect of congestion is taken into account by applying correction factors to VOC.
Suitable correction factor, called Congestion Factor is suggested (Kadiyali 1991), which is a function of type of road and the volume capacity ratio, and is multiplied with the distance related component to VOC. The ratio between the free speed and speed during congestion is multiplied with the time component of VOC. The congestion occurs during the peak hours only. Therefore, information about the hourly traffic volume is desirable to calculate precisely the congestion factor. Cost due to accident Roads, having poor functional characteristics are prone to have more number of accidents. The cost of accident is difficult to measure. Accident involves (i) cost of human lives and human agony (ii) loss due to injury (iii) cost of hospitalization (iv) damage to the property and vehicles. Depending on the severity and degree of damage caused by accident, some values are generally suggested (Kadiyali 1991). Cost of travel time Better quality of roads, new road diversions, road over bridge, exclusive expressway, bring about reduction in the travel time. The savings in travel time is enjoyed by the passengers and the transport operators. The cost of travel time is different for different users. A person having higher salary scale may be willing to pay higher fare, for a quick and comfortable transportation, and accordingly he will select a specific mode of travel. This concept is used to determine the cost of travel time of an individual - thus, a person's wage rate may be considered to determine the value of his time. However, this approach can only give time value during the work trip and not for recreation or other non-working time. The time value of non-working hours can be calculated from the studying the behaviour of passengers in showing the preferences of choosing a particular mode of transport during the non-working hours, when various alternative modes of transportation are available (Nash 1997). Table-2 presents values of the travel time evaluated based on Indian study (Kadiyali 1991). As obvious, the cost of travel time for work trip is higher than that of non-work trip. Table-2 Value of travel time of passengers as per 1990 Road User Cost Data (IRC:SP-30 1993) Type of passenger Work trip (Rs/hr) Non-work (Rs/hr) Bus passengers on trunk routes 27 3 Bus passengers in secondary routes10 2 Car/ taxi/ two-wheeler passengers 30 4 trip
The objective of a good transportation system is to provide a efficient, quick and safe transportation to its users. And this is counted as the benefit of transportation. The various possible forms of benefits can be summarized as follows (IRC:SP-30 1993): Road user benefits: This type of benefit includes, benefit due to (i) saving in VOC, (ii) saving of travel time (iii) saving in terms of accident cost (iv) saving in the cost of maintenance etc.
Social benefits: This type of benefit includes, benefit due to improvement in administration, health, education, agriculture, industry, trade, environmental standards etc. Estimating the benefit component of a transportation system is difficult, and sometimes is not possible to quantify benefits for some components. For example, a good transportation system may reduce the noise and air pollution level, or, may enhance the aesthetics, which is difficult to quantify. Also, the benefits are derived from the future traffic situation, which sometimes becomes difficult to apprehend beforehand. However, the direct benefits, for example, the reduction in the VOC due to a proposed new transportation facility can be quantified As mentioned earlier, the cost as well as the benefit is spread over the time horizon. It is difficult to compare the cost/ benefit components, unless they are brought to equivalent values at a particular base year. There are various methods available for economic evaluation of highway projects, which enables an analyst to compare relative benefits which could be derived from various alternative projects. The various methods of economic evaluation of highway projects are discussed briefly in the following. Cost-Benefit Ratio Method In cost benefit ratio analysis, the costs and the benefits of individual highway projects are calculated, bringing all the expenditures to the base year for comparison purposes. The cost-benefit ratios of the various alternative highway projects are then compared. Net Present Value Method In the net present value method, the cost and the benefits of the individual years are discounted to the present value and compared across various alternatives. The Net Present Value ( NPV ) at the base year can be written as:
.(5) where, B i is the benefit of the i th year, C i is the cost of the i th year and n is the number of years. Internal Rate of Return Method (IRR) Internal rate of return is that discount rate, for which the NPV value is zero. This can be obtained by setting the value of NPV in Equation (5) as zero, and solving (by trial and error) for the value of r . If the rate of return thus calculated is more than the market interest, then the project is adjudged to be acceptable. Comparison of various methods The cost-benefit model is simple to use, but sometimes when the cost-benefit ratio of two models are close to each other, it becomes difficult to interpret, and choose the best option.
Some components whether will be treated as benefit or cost (i.e. whether it will go to the numerator or denominator), sometimes appear confusing. This is because savings in cost is benefit in other words. In the NPV or cost-benefit ratio methods, some discount rate is assumed, and various alternative projects are compared. If different discount rate is assumed instead, the order of choice among the alternatives may change. IRR method itself finds out the discount rate, and therefore inaccuracy in analysis in assuming some arbitrary discount rate (as is done in cost-benefit ratio or in NPV method) is taken care. Thus, IRR method seems to be the most preferred economic analysis tool.
An approximate estimation of the expenditure in the road sector in India, carried out for the coming ten years, amounts to Rs. 25,000crore for expressways, Rs. 120,000crore for National Highways and Rs. 70,000crore for state highways (Gupta 2000). The funding sources of this huge investment could be managed from the (i) government sector (ii) international bank loans (iii) private sectors etc. The Golden Quadrilateral and North-South and East-West corridors, for example, will involve an approximate expenditure of Rs.58,000 crore (NHAI 2005). The funding of these projects has been arranged from various organizations - approximately Rs.20,000 crore from cess (from petrol and partially from diesel), Rs.20,000 crore from the World Bank and the Asian Development Bank loans, Rs.12,000 crore from market borrowing, and Rs.6,000 crore from the private sector (NHAI 2005). For the PMGSY project, it is decided that the finance will come from 50% of diesel cess , market borrowing, and external funding agencies (PMGSY 2005). The government funding may come from various sources, like, budgetary allocation, special road development bonds, fund out of cess on diesel or petrol (Central Road Fund) etc. Toll revenue can be utilized for raising the debt finance, or, supporting the maintenance activities on that particular road stretch, and needs to be operated by government or private agency. The difficulty with the toll financing for debt realization is that the plough back time-period is long, and fluctuative. If at least the maintenance requirement of a particular stretch can be completely financed from toll, it can stay away from the competition with the requirements of the other roads in the network. Shadow tolling is another concept, where government pays the investor for each vehicle that enters the stretch. Shadow tolling works better than direct toll from vehicles. Funding for transportation projects can as well be attracted from private organizations. The following policies (NHAI 2005, MORT&H 2005) have been adopted by the Government for encouraging participation of private organizations in highway construction projects: Declaration of the road sector as an industry Provision of capital grants subsidy up to 40% of project cost on case-to-case basis Duty-free import of certain identified high quality construction plants and equipment 100% tax exemption for 5 years and 30% relief for next 5 years, which may be availed of in 20 years
Provision of encumbrance-free site for work, i.e. the Government shall meet all expenses relating to land and other pre-construction activities Foreign direct investment upto 100% in road sector Easier external commercial borrowing norms Higher concession period, up to 30 years Right to collect and retain toll There could be various models of degree of involvement of private organizations in transportation project financing, like, (i) completely owned, financed and operated by private body (ii) build, operate and transfer (BOT) approach (iii) build, transfer and operate (BTO) approach, (iv) finance, build and lease approach etc. Out of these models BOT is generally considered as the most effective form of privatization. Several variations of the BOT approach exist, like, build, own and operate (BOO), build, own, operate and sell (BOOS), built, operate, lease and transfer (BOLT), build, own, operate and transfer (BOOT) etc (Krishnan 2000). Transportation project requires large outlay of investment; therefore, all the funding possibilities may have to be explored for successful implementation of a conceived project The cost of a proposed highway construction incurred by the agency (government or private) is rather easier to estimate than estimating the equivalent costs incurred and/or benefits enjoyed by the road users. The agency cost involves cost of construction and maintenance of a highway facility. The cost (or benefit) of the road user includes vehicle operating cost, congestion cost, accident cost, travel time cost etc. Economic viability of a proposed project can be adjudged through mainly three alternative approaches, viz. cost-benefit ratio method, net present value method and internal rate of return method. The finance to highway projects are arranged from various sources, e.g., government funding, loan from financial organizations, private investment, imposition of cess to fuel, floating public bond, tolling etc.