基于数据融合方法的智能电能表运行剩余寿命预测

本文融合加速退化试验数据和外场检测退化数据对智能电能表进行在线运行的剩余寿命预测。首先，基于加速退化试验（ADT）数据建立非线性Wiener过程退化模型和温湿综合加速模型，利用贝叶斯理论估计模型参数。其次，利用外场检测的退化数据对退化模型中参数进行不断更新，采用粒子滤波算法实现这一更新过程。最终，给出智能电能表在外场状态检测时刻开始的剩余寿命预测结果。该方法解决了两个问题，一是解决了仅仅利用ADT数据对智能电表在线运行状态评估不准确的问题；二是解决了仅仅利用外场使用条件下的数据量建立预测模型不准确的问题。不仅如此，使用粒子滤波（PF）算法对参数更新的精确度也很高。因此，本文对于智能电能表数据融合方法的研究有着一定的参考价值。

Remaining life prediction of smart electricity meter based on data fusion method

This paper combines the accelerated degradation test data and the external field detection degradation data to predict the remaining life of the smart energy meter online operation. Firstly, a nonlinear Wiener process degradation model and a temperature-humidity comprehensive acceleration model are established based on accelerated degradation test (ADT) data, and Bayesian theory is used to estimate the model parameters. Secondly, using the degradation data of the external field detection to continuously update the parameters in the degradation model, the particle filter algorithm is used to realize this update process. Finally, the remaining life prediction result of the smart electricity meter at the time of detecting the external field state is given. The method solves two problems. One is to solve the problem that the evaluation of the on-line operation status of the smart electricity meter is inaccurate by using only ADT data; the second is to solve the problem that the prediction model is inaccurate only by using the data volume under the use of the external field. Not only that, the accuracy of parameter updates using the Particle Filter (PF) algorithm is also high. Therefore, this paper has certain reference value for the research of smart electricity meter data fusion method.