基于神经网络的柴油机特征参数评价与选择

在设备故障诊断中,正确地提取与选择特征参数对于诊断结果的有效性和准确性具有关键性的意义,在提出评价判据时样本的概率分布往往难以确定,针对模式识别中特征量的选择方法,结合人工神经网络原理,提出了利用人工神经网络进行故障特征量评价与选择的方法,实现了对柴油机特征参数的提取及选择,有效地解决了柴油机状态监测与故障诊断中测试参数多而难以优化的问题。

Evaluation and Selection of Diesel Characteristics Based on Artificial Neural Network

In fault diagnosis of equipment, extracting and selecting the characteristics correctly contribute greatly to effectiveness and accuracy of diagnosis result. Due to difficulty of determining the probability distribution of the samples in defining the valuation criterion, this paper has studied the method for the valuation and selection of the fault characteristics using artificial neural network based on principles of neural network and the method for the selection of fault characteristics, and has realized the extraction and selection of the fault characteristics of diesel engine. This method can deal with the problem effectively that the characteristics to be measured are difficultly optimized in the state monitoring and fault diagnosis of diesel engine.