船舶机电设备各状态量关系复杂，单一或少数状态量难以准确反映其真实性。决策树模型能借节点和分支清晰呈现不同状态量组合及交互作用，给出更全面准确的评估结果。因此，提出决策树模型下船舶机电设备可靠性评估方法。计算各状态量与船舶机电设备可靠性的关联度，筛选高关联度状态量作为重点监测评估对象。利用决策树融合这些高关联度状态量数据样本特征信息，通过设定根、叶节点，以当前可靠性状态数据特征子集确定基尼指数，据此分裂可靠性等级，生成直观树形结构，明确映射关系，增强评估解释性。并用悲观剪枝法对构建的决策树评估模型剪枝，简化模型以提升评估效果。实验结果表明，所提方法能够提升船舶机电设备运行可靠性29%以上，可以保障船舶机电设备的安全运行。
The relationship between various state variables of ship electromechanical equipment is complex, and a single or few state variables are difficult to accurately reflect their true reliability. The decision tree model can clearly present different combinations and interactions of state variables through nodes and branches, providing more comprehensive and accurate evaluation results. Therefore, a reliability evaluation method for ship electromechanical equipment under the decision tree model is proposed. Calculate the correlation between various state variables and the reliability of ship electromechanical equipment, and select high correlation state variables as key monitoring and evaluation objects. By using decision trees to fuse the feature information of these highly correlated state data samples, and by setting root and leaf nodes, the Gini index is determined based on the current reliability state data feature subset. Based on this, the reliability level is split to generate an intuitive tree structure, clarify the mapping relationship, and enhance the interpretability of the evaluation. And use pessimistic pruning method to prune the constructed decision tree evaluation model, simplify the model to improve the evaluation effect. The experimental results show that the proposed method can improve the operational reliability of ship electromechanical equipment by more than 29%, ensuring the safe operation of ship electromechanical equipment.
2025,47(9): 134-138 收稿日期：2025-1-13
DOI：10.3404/j.issn.1672-7649.2025.09.023
分类号：TM72
基金项目：福建省企事业单位委托科技项目（H202409016）
作者简介：李烈熊（1984-），男，硕士，讲师，研究方向为机电系统辨识及故障监测
参考文献：
[1] 李金辉, 孙嘉徽, 万军, 等. 船舶机电设备可靠性试验与评估技术研究综述[J]. 船舶工程, 2023, 45(12): 84-93.
[2] 贾小平, 贾宝柱. 基于双向正规化投影和改进TOPSIS技术的船舶设备可靠性评估模型[J]. 中国航海, 2024, 47(3): 46-54.
[3] 董正琼, 李晨阳, 唐少康, 等. 基于多状态故障树的舰船装备可靠性分析方法[J]. 火力与指挥控制, 2023, 48(4): 59-64.
[4] HOSSEINI M , SIAVASHI M , SHIRBANI M , et al. Reliability assessment of the Lattice-Boltzmann method for modeling and quantification of hydrological attributes of porous media from microtomography images[J]. Advances in Water Resources, 2023, 171: 104351.1-104351.21.
[5] 王敏, 邹婕, 王惠琳, 等. 基于改进的AHP-CRITIC-MARCOS配电网设备风险评估方法[J]. 电力系统保护与控制, 2023, 51(3): 164-172.
[6] 甄永赞, 阮程. 基于代价敏感支持向量机和多变量决策树的分级自适应暂态电压稳定评估[J]. 电网技术, 2024, 48(2): 778-791.
[7] 崔海花, 赵英凯. 基于热红外图像的船舶电气设备状态异常检测研究[J]. 舰船科学技术, 2024, 46(3): 147-150.
[8] 吴涛, 王占海, 陈奇, 等. 基于C4.5决策树的航空器机翼积冰风险监测仿真[J]. 计算机仿真, 2023, 40(7): 44-48.
