基于知识图谱嵌入与补全的电梯故障预测技术研究

Research on Elevator Fault Prediction Technology Based on Knowledge Graph Embedding and Completion

针对电梯故障数据量大、数据质量差、故障关系复杂等问题,提出一种基于知识图谱嵌入与补全的电梯故障预测技术。首先,以电梯故障数据为研究对象,构建电梯故障知识图谱,并采用本体理论对电梯故障数据进行描述;其次,设计了电梯故障知识图谱嵌入模型,并采用预训练模型进行特征提取和表征;然后,利用知识图谱中的公共关系和语义关系信息对嵌入模型进行补全,并进行预测模型训练;最后,利用训练好的预测模型对电梯运行过程中出现的故障进行预测。实验结果表明,本文提出的故障预测方法故障诊断准确度相对较高,可为电梯故障预测和维保提供参考。

An elevator fault prediction technology based on knowledge graph embedding and completion is proposed to address issues such as large amounts of elevator fault data,poor data quality,and complex fault relationships.Firstly,taking elevator fault data as the research object,a knowledge graph of elevator faults was constructed,and ontology theory was used to describe elevator fault data;Secondly,an embedded model of elevator fault knowledge graph was designed,and a pre trained model was used for feature extraction and representation;Then,the public relations and semantic relationship information in the knowledge graph were used to complete the embedded model and train the prediction model;Finally,the trained prediction model was used to predict faults in elevator operation.The experimental results indicate that the accuracy of the fault prediction results proposed in this paper is relatively high,and the proposed method can provide reference for elevator fault prediction and maintenance.