针对化工过程具有规模大、复杂性高、变量多的特点,本文提出了改进FNN(KFNN)的化工过程故障诊断方法。传统的FNN方法存在运算复杂度高、灵敏度低的问题,将峰度引入到FNN方法中,对数据进行降维处理,使运算复杂度明显降低,提高了故障诊断的精度和灵敏度。将KFNN方法应用到一个实际酮苯脱蜡的化工过程中,仿真结果表明此方法能够及时有效地检测酮苯脱蜡生产过程中存在的故障。 In view of the characteristics of large scale chemical chemical process, high complexity and many variables, this paper presents a method of chemical process fault diagnosis based on improved FNN (KFNN). The traditional FNN method has the problems of high computational complexity and low sensitivity. The introduction of kurtosis into the FNN method reduces the dimensionality of the data and significantly reduces the computational complexity and improves the accuracy and sensitivity of fault diagnosis. The KFNN method was applied to a real chemical process of ketone-benzene dewaxing. The simulation results show that this method can detect the defects in the process of ketone-benzene dewaxing timely and effectively.