为了合理降维绝缘故障特征值，精准自适应诊断绝缘故障，提出船舶电气系统绝缘故障自适应诊断技术。利用核主成分分析法降维处理绝缘故障特征值，输入支持向量机；以边际最大化为原则，建立支持向量机的最优划分超平面，通过引入拉格朗日函数求解最优划分超平面，获取绝缘故障自适应诊断结果；采用粒子群算法优化支持向量机的参数，提升故障自适应诊断效果。实验结果证明：该技术可有效降低绝缘故障特征值维度，精准自适应诊断各种绝缘故障；在不同样本维度及绝缘故障规模下，该技术绝缘故障诊断的马修斯系数较高，即绝缘故障诊断精度较高。
In order to reasonably reduce the characteristic value of insulation fault and accurately and adaptively diagnose insulation fault, an adaptive diagnosis technology of insulation fault of ship electrical system is proposed. The kernel principal component analysis method is used to reduce the dimension of insulation fault eigenvalues and input them into support vector machine. Based on the principle of marginal maximization, the optimal partition hyperplane of support vector machine is established, and the optimal partition hyperplane is solved by introducing Lagrange function to obtain the adaptive diagnosis results of insulation fault. Particle swarm optimization algorithm is used to optimize the parameters of support vector machine to improve the effect of fault adaptive diagnosis. The experimental results show that this technology can effectively reduce the eigenvalue dimension of insulation fault and accurately and adaptively diagnose various insulation faults. Under different sample dimensions and insulation fault scale, the Matthews coefficient of insulation fault diagnosis of this technology is higher, that is, the accuracy of insulation fault diagnosis is higher.
2022,44(14): 123-126 收稿日期：2022-01-28
DOI：10.3404/j.issn.1672-7649.2022.14.026
分类号：TK421
作者简介：林航(1984-)，男，工程师，研究方向为船舶电气
参考文献：
[1] 薛征宇, 郑新潮, 邱翔, 等. 基于支持向量机的船舶感应电机轴承故障在线诊断方法[J]. 船海工程, 2020, 49(5): 1–5+9
[2] 徐鹏, 杨海燕, 程宁, 等. 基于优化BP神经网络的船舶动力系统故障诊断[J]. 中国舰船研究, 2021, 16(S1): 106–113
[3] 王涤, 马爱军, 归宇, 等. 基于P-CNN的局部放电绝缘故障融合诊断[J]. 高电压技术, 2020, 46(8): 2897–2905
[4] 吴建波, 王春艳, 洪华军, 等. 基于极限学习机的船舶柴油机故障诊断[J]. 计算机工程与应用, 2019, 55(15): 147–152
[5] 臧旭, 马宏忠, 吴金利, 等. 基于改进集总经验模态近似熵的GIS放电故障诊断[J]. 高压电器, 2020, 56(6): 129–137
[6] 张育炜, 石琦, 武恩光. 基于支持向量机GIS局放小波包能量谱故障诊断[J]. 电气传动, 2020, 50(9): 99–105+114
