基于贝叶斯数据融合的轨道车辆Pt100温度检测系统设计

Design of Pt100 Temperature Measurement System for Rail Vehicles Based on Bayesian Data Fusion

Pt100温度传感器是轨道车辆温度数据实时监测的常用传感器,其性能的好坏影响着人们对轨道车辆运行状态的判断。为了准确判断出存在故障的传感器,检测系统应包含精准的信号采集系统和有效的数据融合处理方法。首先对轨道车辆Pt100温度传感器信号采集系统的放大电路、A/D转换电路等进行设计,采集系统采用分段非线性多项式拟合算法,得到不同温度区间上的标度变换表达式;将一种基于贝叶斯估计算法的多传感器测量数据融合方法应用于采集信号的处理,判断轨道车辆Pt100温度传感器是否存在故障。研究结果表明,采用上述过程测量温度精准,融合误差小,能够有效筛选出发生故障的传感器。

Pt100 temperature sensor as a common type of sensors of the whole monitoring system in railway vehicles,the performance of Pt100 temperature sensor will affect people's judgment about the running state of railway vehicles. In order to accurately judge the faulty sensor,the detection system should include the accurate signal acquisition system and the effective data fusion processing method. Firstly,the signal acquisition system of the Pt100 temperature sensor is designed,including the amplifier circuit,A/D conversion circuit and so on. Piecewise nonlinear polynomial fitting algorithm is used in the acquisition system,and the scaling transformation expressions on different temperature ranges are obtained. Then Bayesian estimation algorithm is used for data fusion of the measured signal,to judge whether there is a fault in the Pt100 temperature sensor. The results show that the process of measuring temperature is accurate,the fusion error is small,and faulty sensors are effectively screened out.