Article
Full entry |
PDF
(0.4 MB)
Feedback
PDF
(0.4 MB)
Feedback
Keywords:
heat radiation; global optimization; genetic algorithms; parallel programming
heat radiation; global optimization; genetic algorithms; parallel programming
Summary:
This article focuses on the practical possibilities of a suitable use of parallel programming during the computational processing of heat radiation intensity optimization across the surface of an aluminium or nickel mould. In practice, an aluminium or nickel mould is first preheated by infrared heaters located above the outer mould surface. Then the inner mould surface is sprinkled with a special PVC powder and the outer mould surface is continually warmed by infrared heaters. This is an energy-efficient way to produce artificial leathers in the car industry (e.g., the artificial leather on a car dashboard). It is necessary to optimize the location of the heaters to approximately ensure the same heat radiation intensity across the whole outer mould surface during the warming of the mould (to obtain a uniform material structure and color tone of the artificial leather). The problem of optimization is complicated (moulds used in production are often very rugged, during the process of optimization we avoid possible collisions of two heaters as well as a heater and the mould surface). Using of gradient methods is not suitable for solving the problem (minimized function contains many local extremes). A genetic algorithm is used to optimize the location of the heaters. The optimization computation procedure is demanding in terms of the number of numerical operations (especially when the mould volume is large and the number of used infrared heaters is higher). In this article practical results of parallel programming during the calculation process of the evaluation function of every created individual (one possible solution of optimization problem using genetic algorithm) to define its fitness are given. The numerical calculations were performed by a Matlab code written by the authors. Numerical experiments are focused exclusively on the opportunities to use parallel programming to accelerate the optimization procedure.
Search
Partner of
