论文部分内容阅读
从工业生产过程实用的观点出发,鉴于小波变换能有效地对信号进行消噪的优点和支持向量机的卓越学习性能,探讨基于小波和支持向量机的故障趋势预报,并结合专家系统建立解释机制。将其用于工业精对苯二甲酸(PTA)生产过程中对二甲苯(PX)氧化反应器尾氧浓度故障预报的结果表明:该方法能准确地对尾氧浓度故障趋势进行预测,并同时给出故障产生的概率大小,为PX氧化反应器的安全平稳操作提供了保证。
In view of practical industrial process, in view of the advantages of wavelet denoising which can effectively denoise the signal and the excellent learning performance of support vector machine, this paper discusses the fault trend forecasting based on wavelet and support vector machines, and establishes the explanation mechanism by combining with expert system . The results of the prediction of the tail oxygen concentration of para-xylene (PX) oxidation reactor in industrial PTA production process show that this method can predict the trend of tail oxygen concentration accurately and at the same time Given the probability of failure, it provides a guarantee for the safe and stable operation of the PX oxidation reactor.