基于贝叶斯与数据驱动的智能电表状态感知技术研究

为更准确地对智能电能表进行状态评估,文中将加速退化试验数据与现场检测状态数据相结合。文章基于加速退化试验(Accelerated Degradation Test,ADT)数据,成立了线性Wiener过程退化以及综合湿、温度加速模型,以贝叶斯理论对模型进行参数预测,利用外场检测状态数据修正退化模型中的参数,最终给出了智能电能表在运行状态下的状态评估结果。该方法同时解决了仅基于加速退化试验数据的在线运行状态评估和通过外场条件获得的状态数据预测模型的两类不准确性问题,对智能电能表数据融合方法的研究具有一定的参考价值。

Research on the state perception technology of intelligent meters based on Bayesian and data-driven

In this paper,the accelerated degradation test(ADT)data is combined with the on-site detection status data to accurately evaluate the status of the smart meter.Based on the accelerated degradation test data,the degradation of linear Wiener process and the acceleration model of comprehensive humidity and temperature are established.Bayesian theory is used to estimate the parameters of the model.The external field detection state data is used to correct the parameters in the degradation model.The state evaluation result of the intelligent meter in the running state is given.The method solves the inaccuracy problem of the accelerated degradation test data for the intelligent power online operation state evaluation and the state data prediction model obtained by the external field condition,and has certain reference value for the research of the smart meter data fusion method.