【摘要】 论述一种基于人工神经网络（ＡＮＮ）的知识学习方法，采用改进的ＢＰ算法训练ＡＮＮ，用于船用柴油机故障诊断，以解决一般专家系统在知识获取过程中的“瓶颈”问题。还讨论了ＢＰ算法的改进与参数优化，并给出了在故障诊断中的应用实例。更多还原

【Abstract】 This paper describes a knowledge learning method based on artificialneural network(ANN). An improved BP algorithm is used to train the ANN, and its application to the fault diagnosis of marine diesel engine helps to solve the problem of a narrow bottle neck in knowledge acquisition of the common expert systems. In this paper the improvement of BP algorithm and the choice of the optimal parameter are discussed,and some practical training results and faultdiagnosing examples are given.更多还原