基于神经网络的AMT离合器自适应温度补偿控制

An Adaptive Temperature Compensative Control of an Automated Mechanical Transmission Clutch Based on Neural Network

采用电液控制的机械式自动变速器 (AMT)控制品质受液压油温度的影响很大 ,克服液压系统温度对 AMT离合器工作过程的影响是 AMT产品化进程中的一项重要工作。本文应用神经网络辨识的方法建立了无脉宽调制离合器接合过程模型及离合器分离过程温度压力补偿模型 ,并在此基础上提出了基于神经网络的离合器自适应温度补偿控制方法。试验证明 。

The control quality of automated mechanical transmission (AMT) by an electronic-hydraulic controlled system is highly influenced by the temperature of hydraulic oil. Solving the problem of the temperature influence on a clutch work process is an important job in the mass-producing progress of AMT. In this paper, BPNNAS (Back Propagation Neural Network Application System), a software having the feature of Windows GUI, based on BP algorithm and running under Matlab environment, was designed and developed; the model of clutch engagement without pulse-width modulation and the temperature-pressure compensative model of clutch disengagement were established by means of the method of neural network identification using BPNNAS. Based on the models, an adaptive temperature compensative control method for a clutch working process was put forward, which could be described as: while the temperature of hydraulic oil changed, the duty rate was adjusted to control engaged speed of clutch, and the pressure was adjusted to control disengaged speed of clutch. The test results prove that this control method is practicable.