Data Adjustment Method for Cyber-Physical Systems Based on Unscented Particle Filter

Abstract

Cyber-physical system (CPS) in distribution networks comprehensively applies technologies including cloud computing, edge computing and big data mining to realize further utilization of massive data and promote deep integration of power flow and information flow. With the randomness and volatility of abnormal sources in CPS, the coherent combination of sensor communication network and power physical network enhances the cumulative effect of data measurement errors, causing uncertainty of data transmission, further affecting intelligent terminals' monitoring, acquisition and perception of basic data including operation environment, state and electrical quantity of distribution equipment. Consequently, an appropriate method should be adopted to adjust the measurement data in the dynamic process of CPS. In this paper, an unscented particle filter algorithm with resampling for CPS data adjustment is proposed. The unscented Kalman filter is adopted to construct the importance sampling density function of the particle filter algorithm. Resampling is adopted to solve the problem of calculation accuracy degradation due to particle dilution. Taking an actual 85 node system as an example, the simulation is carried out. Experimental results show that this method has higher estimation accuracy than traditional particle filter.