针对涡轮增压器转速实车测量中存在安装困难、精度低的问题,提出了利用RBF神经网络和BP神经网络估算涡轮增压器转速的方法。分别建立了基于RBF神经网络和BP神经网络增压器转速估算模型,通过与台架试验测试样本比较,模型误差分别2.25%和2.27%。同时,RBF神经网络较BP神经网络,具有训练次数少,收敛速度快、结果稳定的优点,更适合实车涡轮增压器转速估算。 Aiming at the problems of difficult installation and low accuracy in actual measurement of turbocharger speed, a method of estimating turbocharger rotation speed by using RBF neural network and BP neural network is proposed. The speed estimation model of supercharger based on RBF neural network and BP neural network was established. The model errors were 2.25% and 2.27% respectively compared with the bench test samples. At the same time, compared with the BP neural network, the RBF neural network has the advantages of fewer training times, faster convergence speed and stable result, and is more suitable for the estimation of the actual speed of a turbocharger.