Fault Pattern Recognition of Diesel Rotor Vibration Based on Rough Set and Neural Network
亮 桑 · Computer Science and Application · 2018
提出了一种基于粗糙集和神经网络的柴油机转子振动故障模式识别方法。针对传感器采集数据多、计算量大的缺点，提出了一种确定数据代表点集的方法。以代表点集表示原始数据，同时剔除了异常点。计算代表点集的小波包，提取振动信号能量特征。在对不同振动信号能量特征数据离散的基础上，对数据离散方法进行改进，针对数据结构特征确定了等频离散法与公平尺度法相结合的离散方法。最后以获取的简化规则为输入，利用BP神经网络进行故障诊断。将本文建立的故障诊断方法应用到柴油机转子振动系统中去，表明了该种方法的有效性。 A method of diesel rotor vibration fault pattern recognition based on rough set and neural network was proposed. Aiming at huge of collected data and computation, representative point set determination method was proposed, representative point set demonstrated original point, an abnormal point was ticked off. Wavelet packet of representative point set was calculated, energy feature of vibration signal was extracted. On the basis of different data discrete methods, the discrete method was improved; a combined discrete method was determined by data construction. Finally, faults were diagnosed by BP network and simplified rule. The fault diagnosis method was applied to diesel, which is feasible.