摘要
针对目前仪器设备损坏率预测方法存在的预测结果与实际情况吻合度低、预测能耗高的问题,提出基于灰色马尔科夫的物理实验室高能耗仪器设备损坏率预测方法。对给定的设备信号数据集中信号进行傅里叶变换,获取信号的频谱向量,将得到的频谱向量实行形态滤波操作,得到频谱轮廓。接着通过对频谱轮廓实行峰值检测操作,将峰值依照从大到小排列,将前若干个特征向量当作物理实验室高能耗设备最具代表性的历史数据。基于仪器设备历史数据的采集,给出仪器设备信号数据频谱轮廓峰值幅值序列残差,并利用C均值聚类法对残差的状态进行分析。根据分析结果对仪器设备损坏的转移概率进行计算,得到未来不同时刻残差状态,以此对残差序列预测值进行计算,完成物理实验室高能耗仪器设备损坏率预测。实验结果表明,所提方法预测结果与实际情况吻合度高,且预测能耗低。上述方法可高效解决当前方法存在的问题,可信度高于当前方法。
In this paper,we focus on a method to predict the failure rate of high energy consumption instrument and equipment in physics laboratory based on Grey Markov model.First of all,we conducted Fourier transform on signals in given device signal data set,and then we obtained frequency vector of signal.Afterwards,we performed morphological filtering on the obtained spectrum vector and obtained the spectrum contour.Then,we conducted the peak detection operation on the spectrum contour and arranged the peak values from big to small.Moreover,we used the first several feature vectors as the most representative historical data of high energy consumption equipment in the physics laboratory.Based on the collection of historical data of instrument and equipment,we gave the residual error of the peak amplitude sequence of spectrum contour of instrument signal data.Meanwhile,we used the C means clustering method to analyze status of residual.According to the analysis result,we calculated the transfer probability of equipment damage to get the residual state of different moments in future.Finally,we calculated the prediction value of residual sequence.Thus,we completed the failure rate of high energy consumption instrument and equipment in physical laboratory.Simulation results prove that the prediction result of proposed method is highly coincident with the actual situation.Meanwhile,the predicted energy consumption is low.The proposed method can effectively solve the problems existing in the current method.The reliability is higher than the current method.
作者
林福忠
LIN Fu-zhong(Longyan University,Shanghang Fujian 364200,China)
出处
《计算机仿真》
北大核心
2019年第3期235-238,306,共5页
Computer Simulation
关键词
物理实验室
仪器设备
损坏率
预测
Physical laboratory
Instrument and equipment
Failure rate
Prediction
作者简介
林福忠(1971-),男(汉族),福建上杭人,教授,研究方向:物理学及基础物理实验教学。
