【作者】 庞玲玲；

【导师】 汪友华；

【作者基本信息】 河北工业大学 ， 电气工程， 2011， 博士

【摘要】 行波感应加热除具有传统纵向磁通感应加热和横向磁通感应加热同样的优点外,还能产生更均匀的温度分布,而且具有振动和工业噪音低的显著优点。行波感应加热多物理场耦合场问题的求解是计算电磁学的技术难点,对其研究具有重要的理论意义。并且从实用上看,行波感应加热涉及重要的工程技术,其研究成果必然会对我国金属制品加工设备水平的提高具有积极的推动作用。本课题在合理简化计算模型的基础上,建立行波感应加热耦合分析模型,对带材工件内磁场、涡流场进行研究。由于行波感应加热装置结构复杂,因而将整个场域依据材料特性与结构特点划分成不同的子域,依据子域采用适当的位函数,之后在公共子域和公共交接面处实现整体合成。各场量的耦合采用间接耦合方法。在这一问题上借鉴了横向磁通感应加热的研究成果,采用有限元方法进行研究。本文主要工作可以归纳为如下几点:(1)提出无槽模型可以在一定程度上克服狭槽效应并进行了二维磁场和涡流场的分析。在传统有槽模型中,槽的正下方与齿的正下方因有效气隙长度不同.轭铁金属磁导率远大于空气磁导率,则带材中轭铁齿位置下方磁感应强度比槽位置下方磁感应强度大的多。二维磁场分布仿真结果显示在传统模型中一部分磁力线未穿越带材区,这将影响能量的传输效率。(2)进行了交叉行波感应加热系统和传统行波感应加热系统三维磁场的计算。从行波感应加热对称模型和交叉模型三维磁场分布情况可见,交叉模型的能量传输效率要比传统模型高。并且,因为上下感应器的位置偏移,交叉结构的磁场分布范围较对称结构时要广。上下感应器线圈产生的磁场更有效地弥补了彼此的弱磁区,带材表面磁场分布更加均匀。(3)为克服此早熟收敛现象,本文借鉴人类社会在进化过程中愈来愈合理的联姻策略,提出了一种新的混合粒子群优化算法。新算法一方面通过部落联姻,避免了粒子基因过于接近,另一方面采用交叉遗传策略,在一定程度上克服了过度破坏个体适应度。在优化Rastrigin函数时,改进的算法显现出了它的优越性。(4)本文从分析粒子群算法的早熟现象的原因入手,受生物进化过程中突变和灾难现象的启发,在标准粒子群优化算法的基础上,本文提出一种群体消亡粒子群优化算法。该算法将微粒分成大小相同的几个种群,在粒子群算法运行的适当时机,依一定方式使群体的适当子群体消亡,并随机补充新个体,以维持群体的适当规模和多样性。对Schaffer’s f6函数的仿真结果表明,该算法易于找到全局最优解。最后,论文对所做工作进行了总结,并提出了进一步研究的方向。更多还原

【Abstract】 Besides the advantages of the traditional longitudinal induction heating and the transverse flux induction heating, the traveling wave induction heating (TWIH) has further advantages, for example, the reduction of the alternate component of the electromagnetic forces affecting the system elements and producing vibration and noise. Especially, it can bring more uniform temperature field distribution. The coupled multi-fields computation in traveling wave induction heating is a difficult and important problem of electromagnetics, so that its research has very significant theoretical value. For its application, traveling wave induction heating refers to very important engineering technology, so its research results could greatly promote the metal processing equipments’ development.This dissertation simplified the theoretical model and then established the reasonable coupled analysis model, so as to study the magnetic field and eddy current field distribution in the metal strip. Because of the complex structure of the traveling wave induction heating equipment, the whole research area has been divided into several sub parts based on its material characteristics. Each subfield adopts its own potential function and then the integration can be taken in the common subfield and the public interface. Every field vector coupled indirectly. This computation method has borrowed ideas from the research of transverse flux induction heating, and the finite element method was adopted here.Main achievements of this dissertation include:(1) The slotless model can overcome slots effects in a certain degree. The 2D magnetic field and eddy current field have been analyzed. In the traditional traveling wave induction heating system, the efficient airgap between the coil and load is longer than the airgap between the tooth and load. The magnetic permeability of the metal yoke is much bigger than that of the air, so the induced magnetic flux density below the tooth position will be bigger than the air position in the strip. The 2D magnetic field analysis shows that some of the magnetic lines do not pass through the load, and this will put a dent in the efficiency of the power transferred between coils and load.(2) The 3D magnetic field computation of the traditional and the crossed traveling wave induction heating systems has been carried out. The simulation results reveal that the energy transmission efficiency of the crossed TWIH system is higher than that in the traditional TWIH system. Due to the offset of the two inductors, the induced magnetic field could distribute in a larger efficient area on the workpiece. Flux densities in the upper and lower inductors make up for the weak areas of each other more efficiently, so as to obtain a more uniform magnetic field distribution on the metal strip surface.(3) A new particle swarm optimization (PSO) with allied strategy formed in the evolutionary process of human society has been introduced to avoid the premature problem in traditional PSO. The modified PSO algorithm decreases the gene similarity by the allied strategy and protects the individual fitness through the intragenic crossing-over. The superiority of new PSO has been verified by applied to optimize the Rastrigin function.(4) Based on the analysis of reason for the premature convergence causing in PSO, and enlightened by the phenomenon of mutation and disaster in the biology evolution procession, a population disappearance particle swarm optimization was presented. The particles are divided into several swarms in the same size, the same numbers of individuals disappear after several iterations, and same numbers of new individuals are added to these swarms to maintain the suitable swarm size. Simulation results of Schaffer’s f6 function show that the algorithm can easily find the global optimum solution.Finally, the work of this dissertation is summarized and the prospective of future research is discussed.更多还原