Research On State Monitoring Algorithm For VRLA Battery With Substation

Valve-Regulated Lead Acid(VRLA) battery is an important part of the power system of substation, which its reliability concerns that related electric power equipment of transformer substation can run safely and stably. However,the battery itself is a complex electrochemical system and many factors can affect the State of Health(SOH), online and accurately estimating battery SOH gradually becomes research hotspot and difficulty.We study SOH online monitoring algorithms of VRLA battery working in discharge and floating state in this paper. Paper main research content includes:from the analysis of battery failure mechanism, we sum up several major factors affecting the battery SOH and compare the advantages and disadvantages of three kinds of SOH estimation methods which include based on fuzzy neural network, normalized discharge voltage curve and the remaining capacity methods; for the SOH estimation problem with battery working in discharge state, we design a SOH prediction scheme, which include against variables is difficult to measure in Shepherd equivalent circuit model, we deduce the approximate calculation mathematical model applying to engineering calculation, and in view of the five moving average method has poor quality dealing with singular value in battery data, we propose the improved data preprocessing algorithm, about problem that discharge data is used to classify healthy cells and failure cells in the battery back, we design failure battery filter module under discharge state, for battery SOH prediction problem, we design respectively SOH prediction module with failure battery and healthy battery based on recursive least square algorithm; problem focused on SOH estimation problem of the battery work in floating state, we design SOH prediction scheme,about problem that float data is used to classify healthy cells and failure cells in the battery back, we design failure battery filter module under float state, for battery SOH prediction problem, we design respectively SOH prediction module with failure battery and healthy battery based on Weibull distribution. After a review in the battery status monitoring and control system, results suggest thatbattery SOH prediction algorithm proposed in this paper meeting the precision requirement of practical engineering.