Optimal Predictive Control Strategy Of SOC On Hybrid Vehicle Lithium Ion

With the quantity increased of automobile, energy shortage and environmentpollution problems become more and more serious. Developing one automobile withnew energy is imminent. Because of low discharge and small energy consumption,hybrid vehicle is by the attention of people from different filed. As the key powersupply of hybrid vehicle, dynamic lithium-ion battery, its technology has been one ofblocks for the development of electric car.SOC value is the most important parameters in electric vehicles operationprocess, also is the main basis for energy control strategy of hybrid vehicle. Underdetailed analysis of characteristics of the lithium-ion battery, find out the influentfactors of SOC value, put forward a new method to estimate SOC value moreaccurately. The primary works and innovation of this thesis are as follows:(1) Analyze relationship between electric current, voltage, temperature, internalresistance, etc in the operation process (the charging and discharging) and SOC valuethrough the working principle and general characteristics of the lithium ion battery, soas to lay a certain foundation for accurately predict SOC value.(2) Analyze four key factors-the electric current, voltage, battery temperature andbattery internal resistance by using gray correction analysis to find out the primaryand secondary order of influence factors on SOC value.(3) Combined with the advantages of least squares support vector machine andfuzzy thought, create a new fuzzy least squares support vector machine (FLSSVM) tocalculate SOC value. In order to consistent with the actual situation, this paperadopted the data collected from battery in variable current charge and discharge stateunder the condition of Prius vehicles in10-15typical working, establish estimationmodel with battery working voltage, electric current and temperature as input and thebattery SOC value as output. Finally conclude that the estimate of maximum is lessthan1%, it is a kind of effective improvement way to estimate the real-time SOCvalue.