摘要
铅酸蓄电池是目前广泛使用的一种二次电池。在胶体电解质铅酸蓄电池的生产中,灌注的胶体电解质量不够的铅酸蓄电池必须在化成结束后重新补充胶体电解质。一般而言,判断铅酸蓄电池是否需要补充电解质是依据其化成后的电池容量和电解液体积。很明显这是一种耗时且不利于胶体电解质铅酸蓄电池配组的方法。文章提出了一种基于支持向量机的铅酸蓄电池补胶分类的方法,通过铅酸蓄电池化成过程中间步骤四个时间点的测试电压判断铅酸蓄电池是否需要补充胶体电解质。研究结果表明,该方法优于基于学习向量量化神经网络的分类方法,可以有效地缩短胶体电解质铅酸蓄电池生产时间。
In lead-acid battery manufacturing, batteries with improper electrolyte quantity must be selected from normal batteries at the end of formation according to their capacity and electrolyte volume, which apparently is not beneficial for battery assemblage and is a time cost process. In this paper a new method called support vector machine （SVM） has been used to select batteries with improper electrolyte quantity based on battery discharge voltage at four different time points at the middle of formation. The SVM gives satisfactory prediction accuracy for lead-acid battery classification, which is better than results obtained from learning vector quantization （LVQ） neural network, This method can reduce lead-acid battery manufacturing time on the base of the database constructed by previous battery test result.
出处
《电源技术》
CAS
CSCD
北大核心
2006年第9期757-760,共4页
Chinese Journal of Power Sources
