基于RBF神经网络区域划分-分块补偿的储油罐用温度传感器非线性补偿

摘要

由于储油罐温度场分布规律复杂，因此采用温度传感器进行储油罐温度数据测量时需要进行误差补偿，提出了一种基于RBF神经网络区域划分-分块补偿方法。根据储油罐温度场分布规律及温度传感器安装布置，将罐内空间划分为若干个区域。利用RBF神经网络对各个区域内的温度传感器分别构建相互独立的补偿模型进行非线性补偿。实验表明，与多种补偿方法相比，该种方法模型结构简单，补偿后的储油罐用温度传感器误差大幅减少。%The distribution of oil tank’s temperature is complicated, so nonlinear compensation for oil tank temperature sensor is needed. In this paper, an error compensation method based on RBF neural network area division-separate compensation was proposed. According to the distribution of oil tank’s temperature and the layout of oil tank temperature sensors, the tank’s internal space was divided into several areas; in each area, independent nonlinear compensation model for separate compensation was built by using RBF neural network. The experiments show that the structure of this method is simple, the measuring errors of oil tank temperature sensor decreases largely.