This document summarizes advanced analytical techniques for structural engineering. It discusses nonlinear analysis methods, automation of analysis through scripting, modeling of composite bridge sections under construction and in service, accounting for creep and shrinkage effects, analyzing structures with seismic isolation devices, and iterative methods for analyzing nonlinear dynamic problems like seismic isolation. Key challenges and solutions are outlined for multi-step nonlinear analysis, automated processing of large models, and accounting for complex behaviors in bridges.
This document summarizes advanced analytical techniques for structural engineering. It discusses nonlinear analysis methods, automation of analysis through scripting, modeling of composite bridge sections under construction and in service, accounting for creep and shrinkage effects, analyzing structures with seismic isolation devices, and iterative methods for analyzing nonlinear dynamic problems like seismic isolation. Key challenges and solutions are outlined for multi-step nonlinear analysis, automated processing of large models, and accounting for complex behaviors in bridges.
This document summarizes advanced analytical techniques for structural engineering. It discusses nonlinear analysis methods, automation of analysis through scripting, modeling of composite bridge sections under construction and in service, accounting for creep and shrinkage effects, analyzing structures with seismic isolation devices, and iterative methods for analyzing nonlinear dynamic problems like seismic isolation. Key challenges and solutions are outlined for multi-step nonlinear analysis, automated processing of large models, and accounting for complex behaviors in bridges.
This document summarizes advanced analytical techniques for structural engineering. It discusses nonlinear analysis methods, automation of analysis through scripting, modeling of composite bridge sections under construction and in service, accounting for creep and shrinkage effects, analyzing structures with seismic isolation devices, and iterative methods for analyzing nonlinear dynamic problems like seismic isolation. Key challenges and solutions are outlined for multi-step nonlinear analysis, automated processing of large models, and accounting for complex behaviors in bridges.
ANALYTICAL TECHNIQUES Dmitry Maslov & Roman Guzeev Institute Giprostroymost Saint-Petersburg Russian Federation PART I finding a problem Issue: The program we use may seem incapable of solving a problem Reasons: Excessive demand Lack of knowledge Solutions: Find a different program Expand the limits of existing software 3 All the commercial FE software is featured with non-linear analysis Non-linear analysis is an iterative process involving the model modification according to immediate results Expand the limitsmeans a way of making any possible change in the model after results review 4 Manual result handling requires the art of mouse clicksfor large models XXI century information technologies hint at some automation 5 There is a variety of instruments for building an automation facility VBScript or JavaScript MatLab or MathCad General purpose programming languages Result postprocessing and interaction with solver functionality are the only two things to be developed 6 Scripts: Easy to master but produce very slow code Native code: Runs fast but requires programming skills to be developed Engineers are not programmers, programmers are not familiar with Structural Mechanics A balanced solution has to be found 7 ORCODEN is the new name for Expression Converter program New features : Unicode text editor with syntax highlighting, multiple undo, find and replace features, advanced clipboard facilities, popup hints, and context-sensitive help Script debug tools, breakpoints, watches, step-by-step tracing Resource-intensive GT STRUDL interaction is developed as built-in functions which users treat as integral part of script language 8 Composite cross-section bridges One of the simplest concepts causes additional problems in its analysis 9 Construction Stages: Assembling of steel beams at construction site Incremental launching using temporary and rolling piers Deck concreting at bridge spans Dismantling temporary piers Deck concreting above piers Making road surface Service Actions: Carrying temporary loads Suffering creep and shrinkage effects Possible seismic events 10 Analytical model for composite section
N M Such modeling produces a slightly incorrect diagrams 11 Construction steps for the analytical model: All deck members deactivated, pier joints declared supports, the structure carries the self weight of metal beam and the concrete of the first two divisions S q 1 C q 2 C q 12 Construction steps for the analytical model: Truss members of the first two concreting divisions activated, temporary support joints made free, the structure carries the third concreting division weight 3 C q 13 Construction steps for the analytical model: All members activated, the structure carries the remaining loads II q 14 Are the final displacements correct? 15 NO! At the second stage we had to apply inverse joint reaction in the joints made free, which would make the concrete carry the self weight 1 C q 1 R 2 R 3 R 4 R 16 Correct sequence The joint reaction compensating for the stress can only be foundafter the first stage completed Thus, the analysis after the first stage has to be suspended, reactions listed and put into the model for stages 2 and 3 ORCODEN script performs the entire routine automatically from creating the model up to displaying the results 17 Displacements doubled in the correctmodel 18 So did stresses 19 A new LOADING command would be useful L L i JOINTS COMPENSATING (FOR STRESS) LOADING FACTOR v 'a ' MEMBERS list ... list
The command is supposed to calculate compensating joint forces before the user specifies STATUS FREE or INACTIVE MEMBERS/ELEMENTS for the staged analysis 20 Plate deactivation problem Concrete structural elements do not work in tension areas due tocracks Tension areas are dependent on loading Each loading has to be applied twice: in the whole model and after deck deactivation ORCODEN script performs the entire routine automatically after some modification of the correct model 21 Displacement diagram after the deck deactivation 22 Stress diagram after the deck deactivation 23 According to our experience, its quite enough to find tension areas under dead load, deactivate the deck, and proceed with live load analysis 24 Incremental launching Consequent analysis of many steps for a mechanical system with unilateral constraints with initial gaps 25 Methods for finding solution: Disregarding unilateracyand/or gaps 26 Methods for finding solution: Traditional Rabinovitchalgorithm At each iteration all supports with negative reaction are detached (declared free joints), and detached supports with negative displacement are attached (declared support joints) This algorithm may enter an endless loop and collapse at an iteration, even thoughthe solution definitely exists 27 Methods for finding solution: Modified Rabinovitchalgorithm At each iteration only the support joint with maximal negative reaction is detached and the joint with maximal negative displacement is attached The chance of looping and collapsing for this approach is way less than for traditional method 28 Methods for finding solution: Optimization problem Quadratic programming approach: in the case of incremental launching, the cost function definition involvespositive semi-definite matrix, thusnot every algorithmis suitable General optimization approach turns to either Rabinovitchalgorithms 29 Methods for finding solution: Compression-only non-linear springs Exact values for their stiffness have to be found. These cannot be orders larger than the beam stiffness or the results will be absolutely incorrect We choose the modified Rabinovitchalgorithm 30 Straight and plain models can be drawn in AutoCAD using annotative blocks, then processed in Orcoden, which runs GT STRUDL and results in DXF file with force, moment, reaction, and stress diagrams 31 Curved and sophisticated models with thousands of joints and finite elements still need GT STRUDL meshing facilities to be created Joints: 17000 Elements: 18500 Steps: 200 32 PART II offeringsome science Creep and shrinkage basic assumptions Concrete creep: Creep deformation is proportional to elastic deformation (linear creep) The ratio between creep and elastic deformation is a creep coefficient function 0 ( , ) cr t t 0 0 ( , ) ( , ) ( ) cr el t t t t t
( ) el t creep deformation elastic deformation creep coefficient function concrete age at loading application moment 0 ( , ) t t 0 t 34 Creep and shrinkage basic assumptions Concrete shrinkage: Shrinkage deformation is time dependent function ( ) ( ) ( ) sh sh t f t = shrinkage deformation final shrinkage deformation shrinkage function ( ) sh t ( ) sh
( ) f t 35 Creep coefficient and shrinkage functions for composite bridges (1 ) ( ) t k t e
= ( ) 1 ( ) ( ) t sh sh t e
= 1.6 1.8 k
final creep coefficient at t 4 ( ) 2 10 sh
= final shrinkage deformation at t
36 Elastic deformation obtaining at moment t 0 ,0 0 / ( ) el b b F E A = t=t 1 t n Incremental initial deformation calculation Member distortion calculation Bridge analysis on action of external loading andmember distortion obtained Obtaining elasticdeformationfor the next step Post-processing , , , 1 , i cr i sh i i el i sh i = + = + 1 0 i i i d d l
= + , / ( ) el i i b b F E A = Step-by-step algorithm for creep and shrinkage problem 37 Seismic isolation and energy dissipation Hydraulic damper Friction pendulum bearing Anti-seismic devices for seismic isolation and energy dissipation Rubber bearings with lead core Friction pendulum bearings Hydraulic dampers Steel hysteretic dampers Main objectivesof anti-seismic devices Increasing effectivenatural period Energy dissipation 38 Seismic isolation analysis Direct approach Artificial accelerogram generation Nonlinear time history analysis Simplified approach Response spectrum and damping scale factors are specified by code Response spectrum analysis . 0 1 2 3 4 5 6 0 0.5 1 1.5 2 2.5 3 Natural Period T, sec
Ground type II Ground type I 39 Bridge analytical model with anti-seismic devices eff(d) eff(d) eff(d) eff(d) eff(d) eff(d) d The problem isthat neither secant stiffness nor damping ratio which are dependent on amplitudecan be determined in advance. We need an iterative process for determinationof required parameters. 2 1 2 D eff eff E K d
=
- energy dissipated duringvibration cycle D E 40 Friction pendulum bearing Hydraulic damper 4 D d sd bd E N d = 1 d eff sd bd K N R d
= +
max bd F C
= v 0.15 max 4 D bd E F d eff bd F K d
max 41 The analysis of seismically isolated structure comes down tofinding the solution of non-linear equation with unknown displacements 1. Setup initial value of horizontal displacement d. 2. Initial secant stiffness and damping ratio computation 1. Eigenproblemanalysis 2. Response spectrum analysis Obtaining displacement from DBX Adjust stiffness and damping ratio 1 5% i i i d d d
> Postprocessing 42 Example of seismic isolation analysis Friction pendulum bearing Effective curvature radius R=3.2m Dynamic friction coefficient =0.055 Hydraulic damper Velocity exponent 0.15 Maximal damper force Fmax=50mton max 1 1 ab eff V F K V R d d
= + +
[ ] max 4 D ab E V F d = + Abutment Middle pier 1 1 m eff m V K V R d
= +
4 D m E V d = 2 1 2 D eff eff E K d
=
Vab, Vm support reaction caused by dead load 43 Response spectrum 10% = 2% = 5% = 10% 0.55 5% = + - damping scale factor Peak ground acceleration 0.6g 44 Results of seismic isolation analysis Effective natural period T=1.8s Design horizontal seismic displacement 430mm Acceleration of the bridge girder 0.56g<PGA=0.6g 45 Conclusions: No need to give up using the software if it seemsincapable of something There is a way of going beyond the limits Advice: Dont trust the digit until it has been checked seven times Questions: ??? 46 THANK YOU FOR YOUR ATTENTION
