柴油机状态变化趋势的组合预测方法研究

Study on Integrated Prediction Method for Engine State Trend

以某型装甲车辆用12V150型柴油机为对象进行了大量实车试验,测取了能够表征柴油机技术状态变化的缸内压缩压力峰值、加速时间、减速时间、供油提前角、燃油消耗量、振动功率6个特征量,并对其进行归一化和特征融合处理,从而确定了柴油机寿命预测特征量;研究了AR时序预测模型、灰色预测模型和人工神经网络预测模型的建模方法,并对其组合构建了组合预测模型,组合权值由灰色关联度求解。对一定使用期的车辆用柴油机进行了技术状态变化趋势的预测。预测结果表明,利用组合预测方法对柴油机进行技术状态变化趋势预测,可以提高预测稳定性以及预测精度。

Based on numbers of experiments on 12150L diesel engine, the characteristic parameters which characterize the engine state like the cylinder maximum compression pressure, the acceleration time, the deceleration time, the fuel supply advance angle, fuel consumption and vibration power. Through parameter normalization and integration, the characterized parameter to predict engine lifetime is determined. The modeling approaches of AR time model, the grey model and the artificial neural network model are studied, and an integrated prediction model. Combination weights are calculated by using the grey correlation analysis. State trend prediction of a vehicle engine for a certain service period is made. Predicted results show that the integrated prediction model can increase the prediction stability and accuracy.