现代机器学习算法的船舶配电网运行状态评估  

Operation state evaluation of ship distribution network based on modern machine learning algorithm

作　　者：刘继军[1] 白玉新[2] 赫英凤 LIU Ji-jun;BAI Yu-xin;HE Ying-feng(Taiyuan Institute of Technology,Taiyuan 030008,China;State Grid Shanxi Electric Power Company,Taiyuan 030001,China)

机构地区：[1]太原工业学院,山西太原030008 [2]国网山西省电力公司,山西太原030001

出　　处：《舰船科学技术》2021年第20期106-108,共3页Ship Science and Technology

基　　金：山西省高等学校科技创新项目(2020L0672)

摘　　要：为保证评估全面性,指标选取会很多,但与此同时会导致运算量过大,使得船舶耗能增大。针对上述问题,提出现代机器学习算法的船舶配电网运行状态评估方法。该研究利用机器学习算法中的加权聚类算法计算指标到聚类中心的距离,根据距离选择评估指标,以选取的指标作为神经网络评估模型的输入变量,在隐含层计算指标与权值乘积分值,根据分值比对配电网运行状态等级划分表,在输出层得出配电网运行状态。结果表明:所研究方法应用下,评估结果与实时检测结果一致,且运算量更小(156),达到研究目标,为配电网故障处理提供可靠依据。In order to ensure the comprehensiveness of the assessment,many indicators will be selected,but at the same time,it will lead to too much computation and increase the energy consumption of ships.To solve the above problems,a modern machine learning algorithm for ship distribution network operation state evaluation method is proposed.In this study,the weighted clustering algorithm in machine learning algorithm is used to calculate the distance from the index to the cluster center,the evaluation index is selected according to the distance,the selected index is used as the input variable of the neural network evaluation model,the integral value of the index and weight is calculated in the hidden layer,the distribution network operation state is obtained in the output layer according to the score comparison.The results show that under the application of the research method,the evaluation results are consistent with the real-time detection results,and the amount of computation is less(156),reaching the research goal.This study is expected to provide a reliable basis for distribution network fault treatment.

关 键 词：现代机器学习算法 神经网络 船舶配电网 运行状态评估 

分 类 号：TP241.2[自动化与计算机技术—检测技术与自动化装置]