基于WOA优化FNN-PID的单晶硅加热炉炉温控制  

Temperature Control of Monocrystalline Silicon Heating Furnace Based on WOA Optimized FNN-PID

作　　者：周佳凯 张洪 Zhou Jiakai;Zhang Hong(School of Mechanical Engineering,Jiangnan University,Wuxi 214000,China;Jiangsu Provincial Key Laboratory of Adanced Food Manufacturing Equipment Technology,Wuxi 214000,China)

机构地区：[1]江南大学机械工程学院,江苏无锡214000 [2]江苏省食品先进制造装备技术重点实验室,江苏无锡易214000

出　　处：《半导体技术》2025年第1期86-94,共9页Semiconductor Technology

摘　　要：针对单晶硅加热炉炉温控制的大惯性、强耦合、长调节时间等问题,提出了基于鲸鱼优化算法(WOA)的优化模糊神经网络(FNN)比例-积分-微分(PID)算法。通过测试实验装置的温度推算出模型表达式,采用WOA进行选代寻优,得到合适的PID参数,利用FNN对PID参数进行实时调整,以实现动态解耦。通过仿真软件进行仿真验证,并在搭建的模型上分别进行阶跃响应实验和信号跟随实验。仿真结果表明,相较于传统的PID算法和FNN-PID算法,基于WOA的优化FNN-PID算法有效提升了系统的升温速度且无超调。对加热炉进行升温实验,结果表明温度超调量最高为0.9℃,恒温区温控精度保持在±0.3℃,表明该方法可有效提升系统升温速度和稳定性。To address the problems of large inertia,strong coupling and long adjustment time in the temperature control of monocrystalline silicon heating furnace,an optimized fuzzy neural network(FNN)proportional-integral-differential(PID)algorithm based on whale optimization algorithm(WOA)was proposed.By testing the temperature of the experimental setups,a model expression was derived.The WOA was used to iteratively optimize and find appropriate PID parameters.The FNN was used to adjust PID parameters in real time to achieve dynamic decoupling.Simulation verification was carried out by simulation software,and step response experiment and signal tracking experiment were conducted on the constructed model,respectively.The simulation results show that compared with traditional PID algorithm and FNN-PID algorithm,the optimized FNN-PID algorithm based on WOA effectively improves the heating speed of the system without overshoot.The heating experiment results of the heating furnace show that the maximum temperature overshoot is O.9℃,and the temperature control precision in the constant temperature zone is maintained at±O.3℃.These results indicate that this method can effectively improve the heating speed and stability of the system.

关 键 词：多温区温度控制 鲸鱼优化算法(WOA) 模糊神经网络(FNN) 比例-积分-微分(PID) 单晶硅加热炉 

分 类 号：TN304.05[电子电信—物理电子学]