See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/370733933

Energy Optimization Model Using Linear Programming for Process Industry: A

Case Study of Textile Manufacturing Plant in Kenya

Article · May 2023

DOI: 10.5923/j.ijee.20190902.03

CITATIONS READS

13 1,360

3 authors, including:

Isaiah K Kimutai Augustine Makokha

KIPKABUS TECHNICAL & VOCATIONAL COLLEGE Moi University
7 PUBLICATIONS 30 CITATIONS 9 PUBLICATIONS 117 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Isaiah K Kimutai on 13 May 2023.

The user has requested enhancement of the downloaded file.

International Journal of Energy Engineering 2019, 9(2): 45-52
DOI: 10.5923/j.ijee.20190902.03

Energy Optimization Model Using Linear Programming

for Process Industry: A Case Study of Textile
Manufacturing Plant in Kenya
Isaiah Kimutai1,*, Paul Maina2,3, Augustine Makokha1,3

1
Department of Energy Engineering, Moi University, Eldoret, Kenya
2
Department of Mechanical & Production Engineering, Moi University, Eldoret, Kenya
3
Africa Center of Excellence in Phytochemicals, Textile and Renewable Energy, Moi University, Kenya

Abstract The increasing cost of energy has caused the energy intensive industries to examine means of reducing energy
consumption in processing in order to remain competitive both in local and global markets. This paper presents a method for
modeling and optimizing energy use in textile manufacturing using linear programming (LP). A linear programming model
has been developed which meets the finished product requirements at a minimum cost of energy used in the process subject to
different operational constraints. To develop the model, data required were collected through energy audit of the plant. It is
through energy audit that constraints and products manufactured were identified and used in the development of the model.
Constraints come in the form of system material balance equations and output production demands. Material balance
equations and energy balance are the main features of the model. The model will determine optimum values for the process
design variables, so as to achieve minimum cost. Sensitivity analysis of the model determines how the optimal production
values (optimal solution) are affected by changes in prices of the resources (objective function coefficients).
Keywords Optimization model, Linear programming, Process industry, Energy audit, Material balance, Energy balance,
Sensitivity analysis, Kenya

expended to achieve worthwhile aims [1]. Therefore, energy

1. Introduction management modeling may be defined as an investigation of
the allocation of energy resources over time by determining
Increasing energy demand, high energy costs and scarce the cheapest and most efficient way of meeting final
availability of resources demands energy conservation and demands with available and potential resources and
optimal utilization of energy inputs across all energy technologies.
consuming sectors in order to reduce energy consumption Energy systems are usually complex systems within
and related energy costs. It is through adoption of energy which elements and subsystems are highly correlated, and
management that industries can address energy matters in their behavior is hardly predictable without application of
order to remain profitable and competitive both locally and exact mathematical methods. Under the conditions of limited
globally especially for energy intensive industries such as energy sources, the optimal economic effects of an energy
textile manufacturing. A good and sound energy process can be achieved only if optimal relationships of all
management ensures optimum utilization of energy inputs in components of the system are kept permanent. In addition,
a plant such as electricity and fuel. under the conditions of increased prices of energy systems,
Energy management applies to resources as well as to the optimal operation of energy systems tend to be very relevant
supply, conversion and utilization of energy. Essentially it [2].
involves monitoring, measuring, recording, analyzing, Textile manufacturing is complex because of the wide
critically examining, controlling and redirecting energy variety of substrates, processes, machinery and apparatus
and material flows through systems so that least power is used, and finishing steps undertaken. Different types of yarns,
methods of fabric production, and finishing processes
* Corresponding author: (preparation, printing, dyeing, chemical/mechanical
isaiahkimutai@yahoo.com (Isaiah Kimutai)
finishing, and coating), all correlate in producing a finished
Published online at http://journal.sapub.org/ijee
Copyright © 2019 The Author(s). Published by Scientific & Academic Publishing fabric. Textile manufacturing is an energy intensive process.
This work is licensed under the Creative Commons Attribution International Especially in times of high energy price volatility, improving
License (CC BY). http://creativecommons.org/licenses/by/4.0/ energy-efficiency should be a primary concern for textile
46 Isaiah Kimutai et al.: Energy Optimization Model Using Linear Programming
for Process Industry: A Case Study of Textile Manufacturing Plant in Kenya

plants [3]. Fig. 1 shows product flow diagram depicting the technologies and services in accordance with the executive
various textile processes that are involved in converting raw goals. Linear programming is one of several mathematical
materials in to a finished product. techniques that attempt to solve problems by minimizing or
Modelling of complex problems can lead to better maximizing a function of several independent variables. The
decisions by providing decision makers with more objective function may be profit, cost, production capacity or
information about the possible consequences of their choices. any other measure of effectiveness, which is to be obtained
Hence, the use of energy models as tools for dealing with in the best possible or optimal manner. It concerns the
complicated problems can help decision makers to overcome optimum allocation of scarce or limited resources among
this difficulty. Energy models are useful mathematical tools competing activities, under a set of constraints imposed by
based on the system approach and the best model should be the nature of the problem being studied [7]. These constraints
determined based on the problem that decision makers could reflect financial, technological, marketing,
endeavour to solve [4]. organizational, or many other considerations.
Optimization models try to define the optimal set of It was in 1947 that George Dantzig and his associates
technology choices to achieve a specific target at minimized found out a technique for solving military planning problems
costs under certain constraints leaving prices and quantity while they were working on a project for U.S. Air Force.
demanded fixed in its equilibrium. Every modelling This technique consisted of representing the various
approach abstracts to a certain degree from reality using activities of an organization as a Linear Programming model
stylized facts, statistical average figures, past trends as well and arriving at the optimal programme by minimizing a
as other assumptions. Consequently, energy models linear objective function. Afterwards, Dantzig suggested this
represent a more or less simplified picture of the real energy approach for solving business and industrial problems. He
system and at best they provide a good approximation of also developed the most powerful mathematical tool known
today’s reality [5]. An optimization model defines the as ‘Simplex Method’ to solve Linear Programming problems
required input data, the desired output, and the mathematical [8,9].
relationships in a precise manner. Programming problems deal with the efficient use or
There are many types of optimization models such as allocation of limited resources to optimize certain desired
linear programming, nonlinear programming, objectives. Often, for each problem there exist a large
multi-objective programming, and bi-level programming. number of solutions that satisfy the basic conditions. The
Linear programming has a tremendous number of best solution to a problem is selected based upon some
application fields. It has been used extensively in business overall objective that is included in the statement of the
and engineering, in the areas of transportation, energy, problem. The optimum solution is the one that satisfies the
telecommunications, and manufacturing. It has been proved constraints of the problem and maximizes/minimizes the
to be useful in modeling diverse types of problems in given objective [10].
planning, routing, scheduling, assignment, and design.
Linear programming is also heavily used in various 2.1. Mathematical Formulation of Linear Programming
management activities, either to maximize the profit or Problem
minimize the cost. It is also the key technique of other There are four basic components of an LPP:
optimization problems. This model is defined by an 1. Decision variables -The quantities that need to be
objective function and one or more constraints which have determined in order to solve the LPP are called
linear formats [6]. decision variables.
This paper develops a textile manufacturing plant 2. Objective function -The linear function of the decision
industrial model which is designed to analyze and minimize variables, which is to be maximized or minimized, is
energy costs for a system involving the manufacture of yarn, called the objective function.
grey fabric and finished fabric with equipments using 3. Constraints -A constraint is something that plays the
electricity, wood fuel and fuel oil. The most common part of a physical, social or financial restriction such as
problem in these industries involves allocation of limited labor, machine, raw material, space, money, etc. These
resources among competing activities in the best possible limits are the degrees to which an objective can be
(optimal) way. The objective is to develop mathematical achieved.
models that meet the finished product requirements at a A linear programming problem (LPP) is an optimization
minimum cost of energy used in the process subject to problem in which
different operational constraints.
(i) The linear objective function is to be maximized (or
minimized);
(ii) The values of the decision variables must satisfy a set
2. Linear Programming of constraints where each constraint must be a linear
In recent decades, several models have been proposed in equation or linear inequality;
the field of energy management systems. These models have (iii) A sign restriction must be associated with each
been used widely for optimal allocation of energy resources, decision variable.
 International Journal of Energy Engineering 2019, 9(2): 45-52 47

Two of the most basic concepts associated with LP are you can use it for studying/teaching linear
feasible region and optimal solution. programming.
Feasible region –The feasible region for an LPP is the set  LiPS gives sensitivity analysis procedure-s, which
of all points that satisfy all the constraints and sign allow us to study the behaviour of the model when you
restrictions. change its parameters, including: analysis of changes
Optimal solution -For a maximization problem, an optimal in the right sides of constraints, analysis of changes in
solution is a point in the feasible region with the largest value the coefficients of the objective function, analysis of
of the objective function. Similarly, for a minimization changes in the column/row of the technology matrix.
problem, an optimal solution is a point in the feasible region Such information may be extremely useful for the
with the smallest value of the objective function. practical application of LP Models.
 LiPS provide methods of goal programming,
2.2. General Linear Programming Problem including lexicographic and weighted GP methods,
A general linear programming problem can be which are oriented on multi-objective optimization.
mathematically represented as follows [10]:
Maximize (or Minimize)
3. Production Process
Z = C1X1 +C2X2+…+CnXn
Subject to, Textile manufacturing consists of five main stages namely
spinning, sizing/preparation, rewinding, weaving and wet
a11x1 + a1jxj+…+ a1nxn (≤, =, ≥) b1
processing (finishing). Each stage is made up of a group of
a21x1 + a2jxj+…+ a2nxn (≤, =, ≥) b2
machines used for different purposes. Textile process starts
……………………………………
with yarn making in spinning and woven fabric (grey fabric)
ai1x1 + aijxj+…+ ainxn (≤, =, ≥) bi
in weaving through to finished fabric in wet processing.
……………………………………
The energy inputs used in textile processing are electricity
am1x1 + amjxj+…+ amnxn (≤, =, ≥) bm
as a common power source for machinery, wood fuel and oil
and x1, x2,…,xn ≥ 0
fuel as a fuel for boilers which generate steam and heat. The
The problem is to find the values of xj’s that optimize first, third and fourth stages consumes only electricity to
(maximize or minimize) the objective function. The values operate the mechanical devices while the second and fifth
of xj’s must satisfy the constraints and non-negativity stage consume not only electricity but also fuel (wood and oil)
restrictions. Here, the coefficients cj’s are referred to as cost to generate steam and heat energy. Figure 1 shows the five
coefficients and aij’s as technological coefficients; aij main stages involved in textile manufacturing, energy inputs
represents the amount of the ith resource consumed per unit and products processed in each stage.
variable xj and bi, the total availability of the ith resource. Analysis of each of the five processes with respect to
An LP problem can also be written in a compact form as energy, products and technologies are described below.
follows [11]:
Stage 1-Spinning Process
Maximize (or minimize) the objective function
Spinning is the first stage of textile manufacturing
Z = CX
where raw materials (cotton, viscose and polyester) are
Subject to m constraints processed into yarn. Raw materials are the main inputs of
AX ≤ b, X ≥ 0 stage 1. The main processed product of stage 1 is yarn which
The values in the X vector are called decision variables is represented by X1 in figure 1 below. Depending on
(the unknowns), and the values in the b vector are often customer orders or forecast demands provided by marketing
called right-hand sides (RHS). The technique finds broad use department part of yarn manufactured in stage 1 are sent to
in operations research. store as finished product represented by P1 and the rest sent
Linear Program Solver (LiPS v1.11.1 software.) was used to stage 2 for further processing.
to solve the energy problem. Linear Program Solver (LiPS) Spinning process is made up of several machines which
is an optimization package oriented on solving linear, integer include bale breaker, step cleaner, draw frame, carding and
and goal programming problems. It takes the optimization open end machine. Spinning process electricity consumption
equation, variables and constraints as the input and generates represents 48% of the total consumption in the plant and is
the optimum solution as the output. It solves both the highest consumer of electricity among the five textile
minimization and maximization problems. manufacturing processes. Average monthly electricity
The main features of LiPS are: consumption in stage 1 is 43,942 kW.
 LiPS is based on the efficient implement-ation of the Stage 2-Sizing/Preparation
modified simplex method that solves large scale The input material of stage 2 is sent from stage 1 plus
problems. recycled material from stage 3. The product processed in
 LiPS provide not just an answer, but a detailed stage 2 is illustrated as X2 in figure 1. 3% of the products
solution process as a sequence of simplex tables, so processed in stage 2 are sent to stage 3 as wastes for
48 Isaiah Kimutai et al.: Energy Optimization Model Using Linear Programming
for Process Industry: A Case Study of Textile Manufacturing Plant in Kenya

recycling while 97% is sent to stage 4 for further processing. Stage 4-Weaving Process
Sizing is the process of applying the size material on yarn. Weaving is the fourth stage of textile manufacturing
Size is a gelatinous film forming substance in solution or where treated yarn produced in stage 2 is woven through
dispersion form, applied normally to warp yarns. Starch, weaving machines to produce grey fabric represented by X4
gelatin, oil, wax, and manufactured polymers such as in figure 1. Depending on customer orders part of products
polyvinyl alcohol, polystyrene, polyacrylic acid, and processed at this stage are sent to store as finished product
polyacetates are applied to warp yarn to bind the fiber represented by P2 and the rest sent to stage 5. The main
together and stiffen the yarn to provide abrasion resistance weaving machines used include Projectile, Rapier, Air-Jet
during weaving [12]. Other objectives of sizing include; and Water-Jet depending on the grey fabric required. The
1. To improve the breaking strength of the yarn energy input of stage 4 is electricity and the average monthly
2. To increase smoothness of yarn electricity consumption is 14,428 kW.
3. To protect the yarn from abrasion Stage 5 Finishing Process (wet processing)
4. To increase yarn elasticity
This stage is the last stage of textile manufacturing. The
5. To decrease the generation of static electricity
main processes in this stage include singeing, washing,
6. To decrease hairiness
de-sizing, scouring, bleaching, mercerizing, stentering,
The process involves the use of wood fuel in fire-tube curing and dying (printing). The grey fabric from stage 4 is
boiler to generate steam used for sizing yarn, that is passed through various processes to achieve the ultimate
treatment using starch solution and the use of electricity to product which is finished fabric represented by X5 in figure 1.
operate warping machines. The wood fire-tube boiler 96% of products processed at this stage are sent to store as
efficiency is 17.35% with steam generation rate of 3TPH. finished products represented by P3 and approximately 4%
Average monthly electric consumption in stage 2 is 5,400 is wasted and not recycled. This last stage is also energy
kW. intensive since grey fabric passes multiple stages to meet
Stage 3-Rewinding customer demands. Thermal energy is of high demand in this
Stage 3 is a recycling unit. Waste materials from stage 2 stage especially in washing, dyeing, treatment and drying of
and stage 4 are processed to be used in stage 2 and are the fabric and is supplied by thermo boiler and fire tube
represented by X3. Waste materials are the remains of yarn boiler by burning heavy fuel oil and wood fuel. Electricity is
during the transfer of yarn from cones to warpers beam and required to operate the machines such as thermostenter,
transfer of yarn from weavers beam on weaving. This stage jigger and scouring machine.
consumes only electricity and the average monthly The efficiency of thermo boiler used is 86.1% with an
consumption is 4,147 kW. output of 1860 kW. Average monthly electricity
consumption at this stage is 20,608 kW.

Figure 1. Textile production flow chart

 International Journal of Energy Engineering 2019, 9(2): 45-52 49

4. Mathematical Model Table 1. Energy cost coefficients for the model

Stage Energy cost per Kg

Textile manufacturing is an energy intensive process and
Electricity Wood fuel Oil fuel Total energy
the objective of this study is to develop mathematical model
(Ksh/kg) (Ksh/kg) (Ksh/kg) cost (Ksh/kg)
that meet the finished product requirements at a minimum
Stage 1 50 - - 50
cost of energy used in the process subject to different
operational constraints. The result of this process will yield a Stage 2 5 2 - 7
unique set of values that represent the quantities of products Stage 3 40 - - 40
to be processed to meet monthly demands. Multi-stage linear Stage 4 15 - - 15
programming is used in the development of the model. Stage 5 20 14 14 48
Material balance equations are the main features of
Note: The electricity cost per kWh is Ksh 20, wood fuel cost per kg is Ksh 2.20
multi-stage linear programming. Material balance equations and fuel oil per Litre is Ksh 57 as per company records year 2017.
are general expressions of material flows in the process.
The linear programming problem is solved using the Material Balance Equations
computer software called LiPS solver. To complete the P1= X1 - X2 (2)
analysis LiPS solver is used to perform a sensitivity study X3 = 0.07X4 + 0.03X2 (3)
which shows the effect of changes in the different system
X4 = 0.97X2 (4)
parameters on the current optimal solution. This study then
determines the range of a given element for which the P2 = 0.93X4- X5 (5)
solution remains optimal and, therefore, how sensitive the P3 = 0.96X5 (6)
optimum solution obtained is to the parameter setting. These constraints are the mathematical forms of the
Objective Function material flows of the process. Constraint (2) represents the
The objective function is to minimize the total energy material flow from stage 1 to finished product P1 and stage 2.
costs. The constraint (3) represents the waste material flow from
Minimize Z stage 2 and stage 4 to stage 3 to be recycled. Constraint (4)
= C1X1 +C2X2+C3X3 + C4X4+C5X5 (1) represents the material flow from stage 2 to stage 4.
Constraint (5) represents the material flow from stage 4 to
Where, finished product P2 and stage 5. Constraint (6) represents the
C1 = Energy cost per unit of product to be produced in material flow from stage 5 to finished product P3.
spinning Demand constraints
C2 = Energy cost per unit of product to be produced in
sizing/preparation P1 ≥ D 1 (7)
C3 = Energy cost per unit of product to be produced in P2 ≥ D2 (8)
rewinding P3 ≥ D 3 (9)
C4 = Energy cost per unit of product to be produced in X1, X2, X3, X4, X5 ≥ 0 (10)
weaving
C5 = Energy cost per unit of product to be produced in Where;
finishing P1 = number of units of finished product produced in stage
X1 = number of units of product processed at stage 1 in a 1 in a month
month D1 = number of units of demand for product produced in
X2 = number of units of product processed at stage 2 in a stage 1 in a month
month P2 = number of units of finished product produced in stage
X3 = number of units of product processed at stage 3 in a 4 in a month
month D2 = number of units of demand for product produced in
X4 = number of units of product processed at stage 4 in a stage 4 in a month
month P3 = number of units of finished product produced in stage
X5 = number of units of product processed at stage 5 in a 5 in a month
month D3 = number of units of demand for product produced in
The objective function coefficients are the energy cost stage 5 in a month
(electricity cost, wood fuel cost and oil fuel cost) per unit of P1= X1 - X2
product to be produced. These cost coefficients are P2= 0.93 X4 - X5
calculated by multiplying amount of energy used to produce P3= 0.96 X5
one unit of product by the cost of one unit of energy. Constraint (7, 8, and 9) ensures that the amount of finished
Calculated coefficients of energy for the five stages of textile product (P1, P2 and P3) cannot be negative. Constraint (10)
production are presented in table 1. represents that the amount of products processed in all the
stages (X1, X2, X3, X4, and X5) cannot be negative.
50 Isaiah Kimutai et al.: Energy Optimization Model Using Linear Programming
for Process Industry: A Case Study of Textile Manufacturing Plant in Kenya

Time constraints
P1t1+ P2t2+ P3t3 ≤ A (11)
Where,
t1= required processing time to produce unit of product P1
t2 = required processing time to produce unit of product P2
t3 = required processing time to produce unit of product P3
A = available processing time in a month
Constraint (11) ensures that total production time in a
month cannot exceed the available processing time in a
month. Table 2 shows production rate of each product and
the required processing time to produce one unit of product.
Table 2. Production rate (kg/hr) and processing time per unit product D1 = Product, P1 monthly demand
(hr/kg)
D2 = Product, P2 monthly demand
Production Processing time per unit D3= Product, P3 monthly demand
Product rate product (hr/kg) A = Available processing time in a month
(kg/hr) 1/production rate
The monthly production demands in kg for the three
Product, P1-yarn 150 0.007
products and available processing time in a month are given
Product, P2-grey fabric 75 0.013 below;
Product, P3-finished fabric 50 0.02 P1 = 400 kg
P2 = 600 kg
4.1. Mathematical Model Data P3 = 20,000 kg
The necessary for the model data include; Monthly energy A = 720 hrs
consumption (electricity, wood fuel and oil fuel) for each
5.2. Output of the Model
stage, Monthly production for each stage, Forecast for future
monthly demands, Capacity of the process and Material >> Optimal solution FOUND
availability. >> INFO: feasible solution FOUND after 5 iterations
The data required were obtained during energy audit from >> INFO: LiPS finished after 5 iterations and 0.03
historical records, production process and production seconds
manager data. The necessary input data for the model are >> Minimum energy cost (Z) = Ksh 2,813,030 per month
mainly the energy (electricity, wood fuel and oil fuel) used
Table 3. *** Results – Variables/Constraints ***
per unit of each product at each stage. The units of energy
used are kWh, Litre and Kg for electricity, oil fuel and wood Variable Value Obj. Cost Reduced Cost
fuel respectively. X1 24159.38 50 0
X2 23759.38 7 0
X3 2326.043 40 0
5. Results and Discussion X4 23046.59 15 0
Linear Program Solver (LiPS v1.11.1 software) was used X5 20833.33 48 0
to solve the energy problem. This program solved the linear Constraint Value RHS Dual Price
program problem and gave the ranges over which the cost Row1 400 400 50
coefficients and the right-hand side variables can vary Row2 0 0 -40
without changing the optimality or feasibility of the problem.
Row3 0 0 60
5.1. Input Data Row4 600 600 83.65591398
Row5 20000 20000 137.1415771
Minimize total cost (Z)
= C1X1 +C2X2+C3X3 + C4X4+C5X5 Row6 431.5724 720 0
Row7 24159.38 0 0
The terms of the linear programming objective function
consist of the product of a cost coefficient and a variable. The Row8 23759.38 0 0
cost coefficients have the units of Ksh/kg while the variable Row9 2326.043 0 0
units are Kg. Therefore, each term and the total objective Row10 23046.59 0 0
function have the units, Ksh. This means that the model Row11 20833.33 0 0
minimizes the energy cost spent during production period.
Minimize Z = 50X1 +7X2+40X3 + 15X4+48X5 The model consists of 5 variables and 11 constraints. From
Subject to; table 3 above, we see that, the optimal solution is X1 =
24,159.4, X2 = 23,759.4, X3 = 2,326.04, X4 = 23,046.6 and
 International Journal of Energy Engineering 2019, 9(2): 45-52 51

X5 = 20,833.3 with optimal value of Ksh 2,813,030 per Ksh 2,813,114 per month. Similarly, if monthly demand of
month. The minimum cost of product mix, when the product P3 increases to 20,001 kg the objective function
electricity cost is Ksh 20/kWh, wood fuel cost Ksh 2.20/kg value would increase by Ksh 137 (the dual price) to Ksh
and fuel oil is Ksh 57/Litre, is Ksh 2,813,030 for the planned 2,813,167 per month. Additional hour of production time
month. Results of the model provide optimum production will not affect the objective function value since the dual
mix with minimum energy cost. price corresponding to production time is zero.
Table 5 lists the range of values for which the costs of the
5.3. Sensitivity Analysis various energy sources may vary without changing the
Sensitivity Analysis deals with finding out the amount optimality of the solution though the objective function value
by which we can change the input data for the output of our may vary. Reduced cost indicates how much each cost
linear programming model to remain comparatively coefficient would have to be reduced before the activity
unchanged. This helps us in determining the sensitivity of the represented by the corresponding variable would be
data we supply for the problem. cost-effective. In this research, the reduced cost is zero for all
This includes analyzing changes in: cost coefficients, which shows no reduction is necessary.
1. An Objective Function Coefficient (OFC)
2. A Right Hand Side (RHS) value of a constraint
6. Conclusions
The sensitivity analysis for the monthly demand variables
is shown in table 4 while sensitivity analysis for cost The objective of this study was to apply the linear
coefficients is shown in table 5 below. programming techniques to optimize energy use in textile
manufacturing industry in order to minimize energy cost.
Table 4. *** RHS Range ***
For solving this problem, a linear programming model using
Constraint Current RHS Min RHS Max RHS Dual Price simplex algorithm has been developed. The Linear
Row1 400 -23759.4 41603.9 50 programming technique determined optimum values for the
Row2 0 -infinity 2326.04 -40 process design variables, so as to achieve minimum cost.
Row3 0 -23046.6 +infinity 60 From the results the following findings were drawn:
Row4 600 -20833.3 21233.7 7780/93 1) Spinning and wet processing constitute the bulk of
Row5 20000 0 33722.1 137.142 energy cost in textile manufacturing. Results show
Row6 720 431.572 +infinity 0
that spinning process and wet processing represents
43% and 35% respectively of the total energy cost
Row7 0 -infinity 24159.4 0
(optimal value).
Row8 0 -infinity 23759.4 0 2) Sensitivity analysis shows that an increase in demand
Row9 0 -infinity 2326.04 0 of yarn, woven fabric and finished fabric by 1 kg
Row10 0 -infinity 23046.6 0 causes an increase of Ksh 50, 84 and 137 respectively
Row11 0 -infinity 20833.3 0 on total energy cost.
3) It has been shown that with the application of linear
Table 5. *** COST Range *** programming model, energy cost can be reduced
Variable, Current Reduced considerably in textile manufacturing.
Min COST Max COST
Xn COST, Cn Cost Based on the findings of the research, the following
X1 50 0 +infinity 0 conclusions were drawn:
X2 7 -68.466 +infinity 0 1) Optimal energy use in textile manufacturing can be
X3 40 -730.848 +infinity 0 achieved with the application of linear programming
X4 15 -62.8 +infinity 0 model. The system of equations can be expanded or
X5 48 -7780/93 +infinity 0 reduced to accommodate any variety of system
combinations.
Table 4 lists the range of values for which the RHS 2) Energy saving and conservation measures in spinning
(monthly production demand variables) may change without process and wet processing is vital in order to
changing the optimality of the solution. Dual price is the minimize energy cost in textile manufacturing.
amount in Kenya shillings per month by which the objective 3) The research is significant in the sense that it will
function value is increased per unit increase of the assist the company in making corrective decisions
corresponding right hand side value (Kg/Month). In this well in time using the methods of linear programming.
research, if monthly demand of product P1 increases to 401 This will determine the future production patterns and
kg the objective function value would increase by Ksh 50 outlook resulting in the establishment of new
(the dual price) to Ksh 2,813,080 per month. Also, if monthly production units, while planning for energy cost
demand of product P2 increases to 601 kg the objective minimization of the company.
function value would increase by Ksh 84 (the dual price) to 4) Although this research deals with minimizing energy
 52 Isaiah Kimutai et al.: Energy Optimization Model Using Linear Programming
for Process Industry: A Case Study of Textile Manufacturing Plant in Kenya

cost in textile manufacturing industry, it can be economics and statistics, 148(2), 111-135.
applied to other process industries with similar [6] Zhang, G., Lu, J., & Gao, Y. (2015). Optimization Models. In
processes. Multi-Level Decision Making (pp. 25-46). Springer, Berlin,
Heidelberg.
[7] Akpan, N. P., & Iwok, I. A. (2016). Application of Linear
Programming for Optimal Use of Raw Materials in Bakery.
REFERENCES International journal of mathematics and statistics invention.
Vol, 4.
[1] O'Callaghan, P. W., & Probert, S. D. (1977). Energy [8] Dantzig, G. B., & Thapa, M. N. (2006). Linear programming
management. Applied Energy, 3(2), 127-138. 1: introduction. Springer Science & Business Media.
[2] Dragićević, S., & Bojić, M. (2009). Application of linear [9] Lewis, C. (2008). Linear Programming: Theory and
programming in energy management. Serbian Journal of Applications. Whitman College Mathematics Department.
Management, 4(2), 227-238.
[10] Gass, S. I. (2004). Linear programming. Encyclopedia of
[3] Hasanbeigi, A. (2010). Energy-efficiency improvement Statistical Sciences, 6.
opportunities for the textile industry.
[11] Lewis, C. (2008). Linear Programming: Theory and
[4] Neshat, N., Amin-Naseri, M. R., & Danesh, F. (2014). Energy Applications. Whitman College Mathematics Department.
models: Methods and characteristics. Journal of Energy in
Southern Africa, 25(4), 101-111. [12] Medeiros, R. W., Corrigan, E. J., Tam, T. Y. T., Salem, E. H.,
Chen, J. Y., Reynolds, M. J., & Yasnowsky, J. K. (2004). U.S.
[5] Herbst, A., Toro, F., Reitze, F., & Jochem, E. (2012). Patent No. 6,796,337. Washington, DC: U.S. Patent and
Introduction to energy systems modelling. Swiss journal of Trademark Office.

View publication stats