湿敏电容器的温度补偿方法研究  

Study on temperature compensation approaches of humidity sensitive capacitor

作　　者：叶小岭[1] 廖俊玲[1] 孙宁[1] 

机构地区：[1]南京信息工程大学信息与控制学院,江苏南京210044

出　　处：《传感器与微系统》2013年第6期70-72,76,共4页Transducer and Microsystem Technologies

基　　金：公益性行业(气象)科研专项资助项目(GYHY201106040);江苏省高校优势学科建设工程资助项目;江苏省农业科技自主创新资金项目(SCX(12)3137);江苏省产学研联合创新资金-前瞻性联合研究资助项目(BY2011111);南京市产学研资金资助项目(2012T026)

摘　　要：为了精确辨识湿敏电容器的温度补偿模型,减小系统测量误差,分析了湿敏电容器的温度补偿原理,提出了一种基于改进的遗传算法(GA)和支持向量机(SVM)相结合的补偿算法。基于湿敏电容器,建立了GA-SVM补偿模型,并与BP神经网络方法进行了比较。共144个样本点,经过补偿后GA-SVM和BP模型的相对误差绝对值小于3%的个数分别为143,110;最大分别为0.34%,19.68%。结果表明:改进的GA-SVM算法有效地补偿了温度影响,提高了湿敏电容器的测量精度,同时该算法的逼近能力和泛化能力均要优于BP神经网络。因此,该方法用于湿度传感器温度补偿是有效可行的。To accurately identify the temperature compensation model of humidity sensitive capacitor,and reduce error of system measurement,the principle of temperature compensation of humidity sensitive capacitor is analyzed.A kind of improved compensation algorithm combined genetic algorithm（GA） with support vector machine（SVM） is proposed.The GA-SVM compensation model is established based on the humidity sensitive capacitor,and compared with BP neural network method.A total of 144 sample point,after compensation,the number relative error absolute value of GA-SVM and BP model less than 3 % is 143,110.The maximum value is 0.34 %,19.68 %.The results show that the improved method can compensate temperature influence effectively,the measurement precision of humidity sensitive capacitor is improved.At the same time,the approximation ability and generalization ability of this algorithm is superior to BP neural network.Therefore,using this method for temperature compensation of humidity sensitive capacitor sensor is effective and feasible.

关 键 词：湿敏电容器 支持向量机 遗传算法 温度补偿 

分 类 号：TP302[自动化与计算机技术—计算机系统结构]