【摘要】 工业设备故障预测是智能制造领域的重要课题，在工业生产中被广泛应用。设备故障预测能够提前对设备进行故障预判和诊断，有利于降低设备的运营风险并提高设备的利用率。由于传统预测方法相对简单，对记录数据利用不充分，使得故障预测准确率较低，并且对于保外期的设备故障预测没有系统的解决方案。为解决上述问题，文章设计了一种基于模型集成的保内保外设备故障预测方法。通过分析挖掘预测性维护数据(包含设备的运行状态数据、发生异常和故障数据以及历史维修数据等),构建高维设备特征属性，使用灰色关联度分析提取主要特征，并结合支持向量机模型与XGBoost模型建立保内保外故障预测协同机制，实现了高、低故障风险设备在保修期内外的协同预测。实验结果表明，该方法在多预测周期、多特征空间、多模型对比下的预测准确率均有提升。更多还原

【Abstract】 The fault prediction of industrial equipment is an important subject in the field of intelligent manufacturing, which is widely used in industrial production. Equipment fault prediction can predict and diagnose equipment failures in advance, which is conducive to reducing equipment operation risk and improving equipment utilization. Because the traditional prediction method is relatively simple and the recorded data is not fully utilized, the accuracy of fault prediction is low, and there is no systematic solution for equipment fault prediction outside the warranty period. To solve the above problems, this paper designs a fault prediction method for equipment within and outside the warranty based on model integration. By analyzing and mining Predictive Maintenance data(Telemetry Time Series data, Error and Failure data, Maintenance data, etc.), constructing high dimensional equipment feature attributes, using gray correlation analysis to extract main features, and combining the support vector machine model and the XGBoost model, a cooperative mechanism for fault prediction within and outside the warranty is established. It realizes synergistic forecasts within and outside the warranty between high and low failure risk equipment. The experimental results show that the prediction accuracy of this method is improved under the comparison of multiple prediction periods, multiple feature spaces and multiple models.更多还原