关键词: 桥式抓斗卸船机, 故障预测, 关联规则, 权重系数, 预测函数

Abstract: In order to ensure the safe and stable operation of the bridge grab ship unloaders, problems such as low utilization ratio of monitoring data and not timely failure diagnosis were put forward. To address these problems, a new ship unloader equipment fault prediction model method was proposed based on the improved interest association rule analysis method. The sensor monitoring method and time domain analysis method were used to obtain the operating parameter spaces of the ship unloaders. By using the clustering discrete algorithm, the monitoring data were divided into nonlinear clustering intervals according to their attribute ranges, and the association rule group of ship unloaders was obtained, and the directional feature constraint function of the state monitoring data association rules was constructed to extract the weight coefficients of state data of correlation dimensions, and fault prediction was realized through the association rule changes in the prediction model. The experimental results show that this method may effectively reflect the associated internal feature informations of the ship unloader running state monitoring, and realize the prediction of the ship unloader fault categories, and it is of realistic significance to reduce the frequency of faults.

Key words: bridge type grab ship unloader, fault prediction, association rule, weight coefficient, predictive function