基于数据驱动方法的集装箱龙门起重机能源系统健康状态预测

Health prediction of energy system of gantry container cranes based on data-driven method

为实现起重机节能环保和延长电池组的寿命,以电池的剩余容量作为集装箱龙门起重机能源系统健康状态的评价标准,建立BP神经网络和最小二乘支持向量机(least squares support vector machine,LSSVM)两种电池剩余容量预测模型。分别采用梯度下降算法和标准粒子群优化算法对两种预测模型中的参数进行优化。利用训练好的模型进行电池剩余容量预测。将两种模型的预测值与实测值进行对比分析,结果表明这两种模型都具有高的预测精度,而LSSVM模型是更合适的预测模型。

In order to realize the energy saving and environmental protection of cranes and prolong the life of battery pack,the remaining battery capacity is used as the evaluation standard for the health of the energy system of gantry container cranes.Two remaining capacity prediction models of BP neural network and the least squares support vector machine(LSSVM)are established.The parameters of the two prediction models are optimized by the gradient descent algorithm and the standard particle swarm optimization algorithm.The remaining battery capacity is predicted by the trained models.Comparing the predicted values of the two models with the measured values,it shows that the two models are both of high prediction accuracy,and the LSSVM model is more appropriate.