OMOptim presentation Hubert Thieriot Mines ParisTech Center for - PowerPoint PPT Presentation

OMOptim presentation Hubert Thieriot Mines ParisTech Center for Energy and Processes February 2011

Context

Two projects • Mines ParisTech : CERES : energy efficiency in industrial processes – Key parameters – Technology choices • PELAB : SSF Proviking EDOp – Dynamic optimization

What is OMOptim ?

What OMOptim intends to be ? Optimization platform designed to : – Facilitate algorithms development – Share optimization functions – Apply optimization easily and efficiently

Two main users Academics Industrial A platform to develop A user-friendly tool to and test optimization perform process methods optimizations Hubert Thieriot – OMOptim Presentation 6

Types of problems  Static parameters’ optimization  Structure optimization  Data reconciliation  Parameter estimation  Sensitivity analysis  Dynamic optimization  ... Hubert Thieriot – OMOptim Presentation 7

OMOptim Current version OMOptim 0.9 8

OMOptim – Current structure Reading/ Open modifying Modelica model structure • Heuristic Reading Corba • Linear programming Writing Optimization Modelica OMOptim model C++, C++, Qt External libs Files Simulation Reading/ modifying variables Open Modelica, Dymola

What OMOptim can do ? Static parameters optimization

e.g. Optimization of heat-pump parameters Heat- Exchanger Condenser A Valve • pressure levels • heat-exchanger area Valve P P Compressor Evaporator Compressor • Parameters are static : constant during one simulation • Simulation can still be dynamic Hubert Thieriot – OMOptim Presentation 11

What OMOptim can do ? Structure optimization

Structure optimization Cooling • Introduction of alternative options Heating’ Heating Heating’’ • Optimal choice Drying Drying’ Cooling Hubert Thieriot – OMOptim Presentation 13

Structure optimization • Current : – One compilation per configuration • Challenge : – « Dynamic » change of model structure Hubert Thieriot – OMOptim Presentation 14

What OMOptim can do ? Which objectives ?

Multi-objectives Pareto criteria allows several objectives simultaneously Investment cost Operational cost Hubert Thieriot – OMOptim Presentation 16

What does it look like ?

Optimized parameters Model structure Model Variables Optimized Objectives Hubert Thieriot – OMOptim Presentation 18

Problems Result plot Export result data .csv Solved problems 19

Future work Hubert Thieriot – OMOptim Presentation 20

To be done … • Finalize OMOptim structure • Strengthen link Simulation – Optimization – Derivative information – Structural change – Parallelization – Dynamic optimization • Organize sharability of optimization functions Hubert Thieriot – OMOptim Presentation 21

OMOptim for you… • You’re welcome to use it and to develop in ! • Version is still 0.9 • Share our ideas about structure, algorithms… Hubert Thieriot – OMOptim Presentation 22

Thanks for your attention