Multi-objective Optimization Research Of Electromagnetic Induction Heating Coil

With the development of society in science and technology,Eco-friendly lifestyle becomes more and more popular.In contrast to traditional heating methods,induction heating technology is much safer and more efficient,so it is widely applied to industrial,household and medical equipment.Heating efficiency and temperature uniformity are two important metrics when it comes to evaluating induction heating performance.Induction cooker is one of the typical applications of induction heating technology.The magnetic field effect of heating system is studied in the first place,which helps analyze main factors that affect heating system.Afterwards,a 3D simulation model of induction cooker is built based on finite element simulation software.Using the variable-controlling method,the following three influence factors on the coupling between the pot and the inductor are studied by simulation:the material of pot,the distance between the pot and the inductor,and the shape of the ferrite.In addition,Magnetothermal coupling analysis of induction heating system is carried out in this paper in order to obtain the temperature distribution on pot’s surface heated by two typical induction cooker coils.Moreover,the relationship between heating efficiency and uniformity is analyzed quantitatively and qualitatively.In order to decrease the number of iterations in multi-objective optimization of engineering problems,this paper proposes a comprehensive use of adaptive grid method,roulette method and mutation strategies from genetic algorithm,which achieves an improved multi-objective particle swarm algorithm(IMMOPSO)with faster convergence.Since the multi-objective particle swarm algorithm falls into the local optimum easily and lacks particle diversity,this algorithm uses the adaptive grid method to label the particles and then uses roulette method to extract the global optimum and delete particle in dense regions.Thanks to these approaches,the diversity of particle populations is increased and the particle distribution on the Pareto front becomes more uniform.The IMMOPSO algorithm also adopts the mutation strategy from the genetic algorithm to accelerate the particles to overstep the local extremum,which further improves the overall performance.Compared with traditional MOPSO algorithm and classic NSGA-II algorithm,IMMOPSO algorithm is demonstrated to be more effective through various experiments.The proposed IMMOPSO algorithm realizes the multi-objective optimization of induction heating system.Engineers are enabled to select appropriate parameters of coils in the Pareto solution by changing the weight of the evaluation metrics according to the requirement of practical application.In other words,the proposed algorithm meets the rich customization needs of users.