摘要
反面选择算法是行之有效的用于检测各种变化的方法。改进的这种算法成功用到了故障的精确诊断之中。在此基础上 ,引入经验值 ,对算法做进一步改进 ,使之既能对已经或正在发生的设备故障进行精确诊断 ,又能预测未来可能发生的故障。试验台以及现场机组测试表明 ,只要参数选择合理 ,该套算法能够对设备中诸如转子弯曲、转子不对中、转子不平衡、油膜振荡、旋转失效、转子碰摩、转子支撑系统联结松动等故障进行高准确率的诊断和预测 ,尤其适用于传统诊断和预测方法所无法提取故障特征的复杂系统。同时 ,该方法操作简单且无变换误差 ,无人为干预因素 。
The negative-selection algorithm is a valid way used to detect various kinds of changes. Being improved, this algorithm has been used successfully for accurately diagnosing faults. By introducing an experiential value, the algorithm is further improved, thus it not only can diagnose precisely any fault happened or being happened, but also can predict any possible fault in the future. Having been tested on the test bed and by field compressor set, it has been proved that, given reasonable parameters, this algorithm program can contribute a high veracity diagnosis and prediction of such equipment faults as rotor bended or offset or unbalanced, oil film surged, rotating missed, rotor abraded, and the connection of the rotor supporting system loosed. And it is especially suitable for complicated system in which traditional diagnosis and prediction methods are hardly to pick up fault features. Besides, the algorithm is simple to use without any transforming error and artificial interferential factor, so it provides an accessible way for field technicians.
出处
《石油机械》
北大核心
2002年第6期33-35,0,共3页
China Petroleum Machinery
关键词
反面选择算法
转动设备
故障预测
应用
故障诊断
fault diagnosis prediction rotary equipment negative-selection algorithm
