Research On Thermal Structure Optimization Of Forced Air Cooling Systems Based On Constrained Extremal Optimization And NSGA-?

Forced air cooling systems have been widely applied to the heat dissipation for the high-power power electronic devices and equipments due to their advantages such as good heat dissipation performance,high reliability,simple system structure,and low economic cost.In recent years,some higher requirements on multiple performance indices including low thermal resistance,miniaturization,and light weight have been raised from high power electronic systems,so the optimal design issue of thermal structure for forced air cooling systems has been attracted increasing attentions by academia and industry.This thesis firstly analyzse the main factors affecting the heat dissipation performance by finite-element simulation studies on IGBT modules,then adopts three evolutionary algorithms to the optimal design of thermal structure for forced air cooling systems.These research works have provided new research methods for the optimal thermal structure design area of forced air cooling systems used in high power density power electronic equipments.The main research works and innovations of this thesis are as follows:(1)A 3D finite-element analysis software called ANSYS Icepak has been utilized to study the 3D model and thermal simulation of real IGBT modules.The effects of material and thickness associated with base plate,solder layer and scaleboard of IGBT modules on the heat dissipation performance have been discussed.Also,thermal impedance curves of IGBT modules under different heat dissipation systems have been analyzed.The above analysis have provided some theoretical basis for both the structure design of IGBT modules and optimal thermal structure design of heat dissipation systems.(2)Based on the theory of entropy generation minimization and real engineering limited constraints,the optimal thermal structure design issue of forced air cooling systems used in power electronic equipments has been formulated as a typical constrained optimization problem.Then,a constrained population-based extremal optimization called CPEO algorithm has been proposed to solve this constrained optimization problem.The simulation studies have shown that CPEO performs better than some traditional optimization algorithms such as genetic algorithm and particle swarm optimization algorithm.(3)On the basis of research work(2),an adaptive constraints-dealing population-based extremal optimization algorithm called ATM-CPEO is developed by introducing adaptive trade-off model into CPEO algorithm.The simulation studies on multiple typical constrained optimization benchmark test problems have demonstrated that ATM-CPEO has better optimization performance than some popular constrained optimization algorithms such as genetic algorithm,stochastic ranking(SR),and simple multimembered evolution strategy(SMES).Moreover,the ATM-CPEO algorithm has been applied to the optimal thermal structure design issue of forced air cooling systems.The simulation studies have shown that ATM-CPEO algorithm is superior to CPEO algorithm in terms of thermal structure optimization performance.(4)From the perspective of multi-objective optimization,the optimal thermal structure design issue of forced air cooling systems has been firstly formulated as a typical multi-objective optimization problem,and then a popular multi-objective genetic algorithm called NSGA-? has been used to solve this problem.Furthermore,multiple different performance indices have been adopted to select the Pareto optimal set.The simulation studies have been demonstrated that NSGA-? algorithm obtains a better trade-off solution among multiple performance indices including thermal resistance,pressure drop,and total weight than ATM-CPEO algorithm.The above research work has provided a novel research method for the multi-objective optimal thermal structure design of forced air cooling systems.