高效液相色谱仪异常故障智能诊断仿真

High-Performance liquid Chromatograph Abnormal Fault Intelligent Diagnosis Simulation

在色谱仪异常情况下,传统的故障诊断方法存在诊断所需时间长、故障位置诊断准确率低等不足。针对以上不足,提出一种高效液相色谱仪异常故障智能诊断方法。利用威布尔分布模型计算液相色谱仪故障率,并构建零部件故障率模型,通过该模型计算积累分布函数与可靠度函数。对两个函数进行互换获取威布尔模型的线性回归方程,将线性回归中的散点坐标画在威布尔色谱仪概率图,确定单一分布函数模型,根据该模型存在明显的拐点,对其进行改进获取双重威布尔分布模型。观察其模型分析威布尔概率图上若干个分散数据点,根据拟合曲线对其进行拟合,以此为基础获取故障发生时间拐点。对故障发生点进行定位,通过相邻协作的故障检测算法完成最终故障与否的判断。实验结果表明,所提诊断方法可以快速、精确地诊断出高效液相色谱仪异常故障。

In the abnormal state of chromatographic instrument,the traditional fault diagnosis method leads to long diagnosis time and low diagnosis accuracy for fault location.Therefore,an intelligent method to diagnose the fault of high-performance liquid chromatographic instrument was proposed.First of all,Weibull distribution model was used to calculate the failure rate of liquid chromatography,and then the model of failure rate of components and parts was constructed.Based on this model,the cumulative distribution function and the reliability function were calculated.Then,two functions were interchanged to find the linear regression equation of Weibull model.After that,scatter coordinates in linear regression are plotted on Weibull probability graph of chromatographic instrument,so as to determine a single distribution function model.Because there were clear inflection points in this model,it was necessary to improve the model,so that the dual Weibull distribution model was obtained.In addition,the model was observed and several scattered data points on Weibull probability graph were analyzed.According to the fitting curve,the model was fitted.On this basis,the happening time inflection point of fault was obtained.Moreover,the fault happening point as located.Finally,the adjacent cooperative fault detection algorithm was used to judge final fault.Simulation results prove that the proposed diagnostic method can quickly and accurately diagnose abnormal faults of high-performance liquid chromatographic instrument.