【摘要】 提出了一种新的、基于人工神经网的电机设计知识表示方法。该方法适用于电机设计中经验知识的表示，它与传统专家系统中的知识表示方法具有一定的互补性。同时采用了带有可调参数的多层神经网络，并对ＢＰ学习算法进行了改进。最后给出了一个用该方法表示电机设计中经验知识的实例。更多还原

【Abstract】 A new knowledge representation method in electrical machine design based on artificial neural net- works is presented. This method is suitable for the experience knowledge representation in electrical ma- chine design, and it is complementary with respect to the knowledge representation methods in tradition- al expert systems . Multilayer neural network with adjustable parameters is applied in this method, and the backpropagation learning algorithm is improved . Finally an example of experience knowledge repre- sentation in electrical machine design based on this method is presented .更多还原