基于SVR的电工实验装置故障预测方法研究

本文利用实验装置的数据采集装置,结合学生实验报告的数据分析,通过时间相关性分析,建立了实验装置运行状态的时间序列.采用的算法利用基于支持向量机回归的预测技术对实验装置的关键参数进行数值预测,通过分析预测值与实际值之间的残差对实验装置故障预测.最后,通过典型故障验证了该方法的有效性和实用性.

Study on Fault Prognostic Method of Experimental Equipment Based on SVR

With the data acquisition devices of experimental experiment, combined with the analysis of the experimental data in lab report, the time series of the experimental device operating status is established by time correla- tion analysis. Using numerical prediction of the key parameters of the experimental device is achieved using forecas- ting techniques based on support vector regression algorithm. Whether there is warning incident is determined through the analysis of the residuals between the predicted values and actual values. Finally, failure prediction of the experimental apparatus is achieved, and the method proves to be effective and practical through the typical failure verification.