CAJ | 학술논문

考虑到矿用空压机在长期运行过程中容易由多种因素复合共同作用而出现各种故障，且产生故障的原因和故障之间表现出非线性关系难以用数学模型表达等问题，建立基于改进型自适应神经模糊推理系统的故障诊断系统。该系统采用附加动量算法不断修正自适应神经模糊推理系统中的前题参数以避免采用梯度下降算法时易陷入局部极小，训练速度较慢等缺点，提高系统的忽略网络中微小变化的能力。为了验证该故障诊断系统的性能，将其与基于BP神经网络的故障诊断系统相比较。分析与实验结果表明，改进型ANFIS模型的诊断输出与实际情况完全相符，最大误差为13．7％，最小误差为0．17％，其诊断准确度达到95．85％，在训练速度、误差精度以及收敛性等方面，其性能优于BP神经网络。
고필도광용공압궤재장기운행과정중용역유다충인소복합공동작용이출현각충고장，차산생고장적원인화고장지간표현출비선성관계난이용수학모형표체등문제，건립기우개진형자괄응신경모호추리계통적고장진단계통。해계통채용부가동량산법불단수정자괄응신경모호추리계통중적전제삼수이피면채용제도하강산법시역함입국부겁소，훈련속도교만등결점，제고계통적홀략망락중미소변화적능력。위료험증해고장진단계통적성능，장기여기우BP신경망락적고장진단계통상비교。분석여실험결과표명，개진형ANFIS모형적진단수출여실제정황완전상부，최대오차위13．7％，최소오차위0．17％，기진단준학도체도95．85％，재훈련속도、오차정도이급수렴성등방면，기성능우우BP신경망락。
Considering the problem that various faults have happened frequently by a variety of factors combined ac-tion in the long term operation of mine air compressor and it is difficult to use mathematical model to express the non-linear relation between reasons and faults, a fault diagnosis system based on improved ANFIS is established.This sys-tem uses the additional momentum method to constantly modify premise parameter in ANFIS in order to avoid the shortcoming that it easy to fall into local minimum point and slow training speed when using the gradient descent algo-rithm, and improve the capacity of ignoring tiny changes in the network.In order to verify the performance of the fault diagnosis system, it is compared with the fault diagnosis system based on BP neural network.The analysis and experimental results show that the improved ANFIS model diagnostic output is fully consistent with the actual situa-tion.Its maximum error and minimum error is 13.7% and 0.17%, and the diagnostic accuracy has reached 95.85%.Its performance is better than BP neural network in the terms of training speed, accuracy, convergence and so on.