基于BA-SVM的铂电阻温度传感器非线性校正方法  

Nonlinear Rectification Method of Platinum Resistance based on BA-SVM

作　　者：杨舟 李刚 蒋雯倩 YANG Zhou;LI Gang;JIANG Wen-qian(Guangxi Power Grid Company, Nanning 53002)

机构地区：[1]广西电网有限责任公司电力科学研究院,南宁530023

出　　处：《环境技术》2018年第3期49-53,共5页Environmental Technology

基　　金：广西科技项目"基于多应力模型的智能电表可靠性试验研究"(GXKJXM20160351)

摘　　要：针对铂电阻温度传感器应用中存在的非线性问题,本文提出了基于蝙蝠算法(BA)和支持向量机(SVM)相结合的BA-SVM方法对铂电阻温度传感器进行非线性补偿。补偿结果表明:BA-SVM方法补偿精度较高,收敛速度快,远远优于传统SVM支持向量机模型、PSO-SVM支持向量机模型以及基于最小二乘的多项式拟合方法,可有效提高温度传感器的可靠性,便于在可靠性试验中的应用。A method to compensate nonlinearity of platinum resistance temperature sensor is presented by using nonlinearity compensation model founded by bat algorithm optimized support vector machine（BA-SVM） to aim at its nonlinear problem. The compensation results have shown that the BA-SVM method has higher compensation precision and faster convergence speed compared with the traditional support vector machine model（SVM）, particle swarm optimization optimized support vector machine（PSO-SVM） and polynomial fitting method with least square. Therefore,the proposed method in this paper can effectively improve the reliability of humidity sensor and easy to apply in the reliability test.

关 键 词：BA-SVM 铂电阻 非线性补偿 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置]