为全面整合船舶故障相关的各种知识，为检修人员智能推荐便于理解的故障检测结果，设计基于人工智能的船舶故障检测结果智能推荐系统。知识图谱模块依据船舶故障维修日志建立船舶故障知识图谱；实体抽取模块利用人工智能的长短时记忆网络，在船舶故障描述文本内抽取故障实体；实体识别匹配模块，利用基于实体识别的文本匹配技术，计算抽取的故障实体与知识图谱内故障实体间的匹配得分，以最高匹配得分对应的故障实体为船舶故障检测智能推荐结果。实验证明，该系统可有效构建检查故障知识图谱；该系统可有效抽取船舶故障实体，完成船舶故障检测结果智能推荐。
In order to integrate all kinds of knowledge related to ship faults and to intelligently recommend fault detection results for maintenance personnel, an intelligent recommendation system for ship fault detection results based on artificial intelligence is designed. Knowledge map module builds ship fault knowledge map according to ship fault maintenance log. Entity extraction module uses artificial intelligence long and short time memory network to extract fault entities from ship fault description text. The entity recognition matching module uses text matching technology based on entity recognition to calculate the matching score between the extracted fault entity and the fault entity in the knowledge graph, and the fault entity corresponding to the highest matching score is the intelligent recommendation result of ship fault detection. The experiment shows that the system can construct the fault knowledge map effectively. The system can effectively extract ship fault entities and complete intelligent recommendation of ship fault detection results.
2024,46(11): 173-176 收稿日期：2023-11-29
DOI：10.3404/j.issn.1672-7649.2024.11.032
分类号：TP313.5
基金项目：教育部产学合作协同育人项目（230801813213629）
作者简介：涂芳(1983-),女,硕士,副研究馆员,研究方向为计算机软件及计算机应用、人工智能、信息素养教育
参考文献：
[1] 宋庭新, 韩国晨. 基于预防性维修的船舶装备等级修理决策系统研究[J]. 中国机械工程, 2022, 33(4): 496-503.
[2] 黄鹤, 肖飞, 杨国润, 等. 基于开关模态频率特征的船舶储能变流器故障在线检测方法[J]. 电机与控制学报, 2022, 26(2): 24-31.
[3] 卢月, 李维波, 李巍, 等. 船舶电站控制系统的双CPU混成式故障检测技术[J]. 中国船舶研究, 2021, 16(3): 200-206.
[4] 操江能, 尚前明, 杨安声, 等. 基于优化孤立森林的船舶柴油机故障监测[J]. 船舶工程, 2021, 43(11): 125-132.
[5] 王泷德, 曹辉, 魏来. 不平衡数据下船舶主机在线故障诊断研究[J]. 中国船舶研究, 2023, 18(5): 269-275.
[6] 邱其清, 廖志强. 基于高斯混合和概率神经网络的船舶柴油机故障诊断方法[J]. 船舶工程, 2022, 44(9): 101-106+113.
