基于Krylov子空间降阶的温控算法研究

Research of temperature control algorithm based on Krylov subspace reduction

针对当前常用的散热前馈获取方法存在耗时长、能耗大、步骤繁杂等缺点,文中提出引入模型降阶方法获取散热前馈。通过3组实验获取控制对象响应特性;对在ANSYS Workbench仿真平台建立的控制对象模型进行校正,使仿真模型特性与实际控制对象特性尽可能相同;使用Krylov子空间法对仿真模型进行降阶,降阶后的模型可以通过Matlab计算得出散热前馈;以差示扫描量热仪炉体为控制对象验证散热前馈-PID控制策略的效果。实验结果表明,在环境温度至300℃范围内,扫描模式温控误差小于±5%,恒温模式温控误差小于±0.2℃。

Aiming at the shortcomings existing in the current common heat dissipation feedforward acquisition methods,i.e.,time-consuming,energy-consumption,complex,etc.,the heat dissipation feedforward acquisition by introduced model order reduction method is proposed.The response characteristics of control object were obtained by three groups experiments.In order to make the characteristics of the simulation model and the actual control object as same as possible,the model of control object established on the ANSYS Workbench simulation platform was calibrated.The simulation model was reduced by using Krylov subspace method,and the reduced order model could be calculated to obtain the heat dissipation feedforward by Matlab.Taking the furnace of differential scanning calorimeter as the control object,the effect of heat dissipation feedforward-PID control strategy was verified.The experimental results show that the error of temperature control is less than±5%in scanning mode and the error of temperature control is less than±0.2℃in isothermal mode,between ambient temperature to 300℃.