基于人工神经网络的铅酸蓄电池参数辨识-控制理论与控制工程专业论文.docxVIP

华中科技 大学硕士 学 华 中 科 技 大 学 硕 士 学 位 论 文 II II ABSTRACT The state-of-charge (SOC) of lead-acid batteries is an important parameter of the performance indicators of the batteries, with crucial impact on the output performance and battery life. It is an important part in the field of application of the battery monitoring, to access to the SOC of lead-acid batteries with computer technology and intelligent control theory, and to provide optimal decision-making control theory to the battery in use. This thesis have determined the key parameters of impacting SOC and a neural network based on BP‘s battery performance parameters identification method, by analyzing the electrochemical characteristics of lead-acid batteries and the SOC detection methods commonly used. Through accessing the information on lead-acid batteries and analysing the electrochemical properties of the batteries, it can be concluded that the electrolyte density is the key parameter in the impact on SOC, and the positive and negative electrode potential and the density meet Nernst formula. The thesis has studied the measurement methods with the reference electrode as a potential sensor, and the Ag/AgSO4 reference electrode was produced, and the relecant experiments was designed to obtain positive and negative electrode potential. Identify the density with a three-layer neural network, and the result meet the training requirements of error, and the density could be obtained better according to the parameters of under-voltage, current and so on.the SOC of the batteries could be estimated through the identified value of the electrolyte density. To further improve the forecast accuracy of the life of the batteries, the impact of other parameters could be considered. Through theoretical analysis and experimental results, it shows that: the method to access the electrolyte density indirectly, researched by this thesis, is feasible, and could be used to the density and electrolyte SOC in real time, accurate monitor