基于遗传算法的K型热电偶非线性校正

Nonlinear Correction of K-type Thermocouple Based on Genetic Algorithm

针对直接用遗传算法对K型热电偶进行曲线拟合出现的不足,如收敛过早及局部搜索能力效果差等,采用基于遗传算法的逐步逼近来拟合。考虑到K型热电偶非线性程度不高,提出一种改进遗传算法的非线性校正。经过3次不同适应度函数,同时对遗传算法中的交叉和变异操作进行适当的改进,拟合数据参考标准K型热电偶分度表。最终拟合出电压温度关系函数。通过对比不同方法的拟合结果,表明新方法能实现更好地进行非线性校正,且拟合精度高,相对误差控制在1%以下。

Considering the genetic algorithm （GA） ＇ s premature convergence and poor local search capability in direct fitting curves of K-type thermocouple,the GA-based stepwise approaching method was employed to fit parameters; and regarding the poor nonlinearity of K-type thermocouple,a nonlinear correction method based on modified GA was proposed; having three different fitness function considered,the crossover and mutation operation properly improved in standard GA,as well as the fitting data referencing the standard K-type thermocouple data indexing table,the function between temperature and the voltage was obtained.Comparing the fitting results of different methods indicates that this new method can achieve a better non-linear correction with high fitting precision and the relative error less than 1％.