基于神经网络的再制造自动变速箱检测及试验研究

摘要

Performance test and reliability control of remanufactured automatic transmission should be an important step in process of transmission remanufacturing, therefore, a vibration detecting system for the remanufactured transmission is established based on neural net. On the base of analysis processing of vibration data, it extracts typical time -domain characteristics from the signals as input vector for BP neural network to execute network training setting forth a new way to detect vibrations in remanufactured transmission.%再制造自动变速箱的性能检测和可靠性控制是变速箱再制造过程中的重要步骤,建立了一套基于神经网络的再制造变速箱振动检测系统.在对其振动数据进行分析处理的基础上,提取信号的典型时域特征作为BP神经网络的输入向量进行网络训练,提出了一种再制造变速箱振动检测的新方法.