本文已被：浏览 610次 下载 0次
中文摘要: 在对称重设备数字化改造的过程中，有些研究人员提出了对某一特定传感器的故障诊断方法，但对于非指定传感器或者两个传感器同时发生故障的情况却没有检测方法。为此，本文提出了一种基于径向基神经网络预测的任意一个或两个称重传感器的故障检测方法。本文首先建立单个传感器的预测模型和任意两个传感器的预测模型，然后通过这两个预测模型计算出任意一个称重传感器的预测值和任意两个传感器的预测值，根据预测值与实际值之间的差值判断称重传感器故障个数、位置、类型等信息。实验表明，当称重传感器的输出误差大于0.3t（吨）时使用此方法可以准确检测出称重传感器的故障信息。
Abstract:In the process of digitalization of weighing equipment, some researchers have proposed some fault diagnosis methods for a particular sensor, but for non-specific sensor or two fault sensors’ situation these methods are not applicable. So, this paper presents a method for any one sensor or two sensors based on RBFNN. Firstly, this paper establish the prediction model of any single sensor and prediction model of any two sensors, and then calculate any one weighing sensor’s predictive value and any two sensors’ predictions, judging the fault weighing sensor’s number, location, type through the difference between predicted value and actual value. Experiments show that this method can accurately detect the fault information of the sensor when the error of the sensor is above 0.3 tons.
keywords: weighing sensor fault detection fault type identification radial basis function neural network
文章编号：cg16001230 中图分类号： 文献标志码：
基金项目:江苏省科技支撑重点项目；江苏省自然科学基金
引用文本：