PID神经网络算法对K型热电偶非线性校正

Nonlinear correction of K-type thermocouple using PID neural network algorithm

针对火箭发射场发射效应测温系统中K型热电偶存在的非线性特性,设计中将非线性特殊点作为训练样本,采用新型动态PID神经网络算法对热电偶进行非线性校正。针对基本BP算法收敛慢、易陷入局部极值的缺点,提出利用粒子群算法来改进网络的寻优过程,并在传统算法基础上对其惯性权值的递减式子进行改进。使用Matlab建模仿真表明,改进算法在寻优过程中,收敛速度快,全局寻优能力强,有较好的控制效果。拟合出的温度电压关系呈现好的线性度,相对误差均控制在1%以内,提高了系统测试精度,满足对火箭发射时温度环境效应的监测要求。

In allusion to the nonlinear feature of K-type thermocouple in the launch effect temperature measurement system of the rocket launching site,the nonlinear special points are taken as the training samples in the design,and a new dynamic PID neural network algorithm is adopted to perform the nonlinear correction of the thermocouple. In allusion to the disadvantages of slow convergence and easy to be trapped in the local extremum of the basic BP algorithm,the particle swarm algorithm is proposed to improve the network optimization process,and its inertia weight decreasing formula is improved on the basis of the traditional algorithm. The Matlab modeling simulation shows that the improved algorithm has a fast convergence speed,a strong global optimization capability,and a good control effect during the optimization process;the fitted temperature-voltage relationship presents a good linearity;the relative errors are all controlled within 1%;the improved algorithm can improve the test accuracy of the system,and meet the monitoring requirement of temperature environment effect during rocket launch.