基于PNN的航空铅酸蓄电池容量预测

Aviation Lead-Acid Battery Capacity Prediction Based on Probabilistic Neural Network

针对航空铅酸蓄电池容量预测的复杂性和非线性等问题,提出了一种基于概率神经网络(PNN,probabilistic neural network)的航空铅酸蓄电池容量预测模型。阐述了PNN的基本理论,分析了影响航空铅酸蓄电池容量的因素,并合理地选取了PNN的输入量。在中国民航飞行学院各个分院采集样本数据并进行了验证,验证结果表明,基于PNN的航空铅酸蓄电池容量预测方法不但具有泛化能力好、学习速度快、预测精度高等优点,而且可以有效地减少由满容量放电造成的电池老化,延长航空铅酸蓄电池的使用寿命,具有良好的应用前景。

In view of the features of complexity and non-linear for aviation lead-acid battery capacity prediction,a new model based on probabilistic neural network（PNN） for aviation lead-acid battery capacity prediction is presented.The basic theory of PNN is formulated,the factors that affect the aviation lead-acid battery capacity are analyzed,and the input variables of PNN are selected reasonably.To measure the effectiveness of the proposed model,data sets are acquired from Civil Aviation Flight University of China.Results indicate that the capacity prediction model based on PNN has some advantages such as good generalization ability,fast learning speed and high prediction precision.What＇ s more,it can effectively avoid the battery aging which results from full-capacity discharge to extend the battery service life,and has a good application prospect.