【摘要】 为使铝电解生产过程中异常诊断的实时性和准确性得到提高,针对铝电解过程异常特点,提出了一种基于智能诊断和模型诊断的融合故障诊断方法。将系统辨识与参数估计和模糊神经网络异常诊断有机结合,通过异常诊断、异常分类环节,实施对铝电解生产异常类型的识别与异常预报。本文采用了由模型异常诊断模型进行异常判断,由神经网络进行异常分类,多层异常综合诊断模式,通过神经网络进行异常类型识别与优化,不仅提高了异常诊断的实时性,又提高了系统异常诊断的准确率。更多还原

【Abstract】 In order to improve the real-time and accuracy of abnormal status diagnosis,according to the characteristics of abnormal status in aluminum reduction process,a new fusion abnormal diagnosis method based on model diagnosis and intelligent diagnosis is proposed,the system identification is combined with parameter estimation and fuzzy neural network abnormal status diagnosis. Through the abnormal status diagnosis,abnormal status classification process,the objective of the identification of aluminum reduction abnormal status types and abnormal status prediction are implemented. In this paper,we use abnormal diagnosis model to judge the abnormal status and achieve abnormal status classification by neural network,multi-layer abnormal integrated diagnosis mode,through abnormal status recognition and optimization of abnormal status type by neural network,not only to improve the real-time abnormal status diagnosis,and improve the accuracy of the system abnormal diagnosis.更多还原