| 引用本文: | 符杨,黄路遥,刘璐洁,魏书荣,任浩瀚,王毅,唐庚培.基于状态自适应评估的海上风电机组预防性维护策略[J].电力自动化设备,2022,42(1): |
| FU Yang,HUANG Luyao,LIU Lujie,WEI Shurong,REN Haohan,WANG Yi,TANG Gengpei.Preventive maintenance strategy for offshore wind turbine based on state adaptive assessment[J].Electric Power Automation Equipment,2022,42(1): |
|
| 摘要: |
| 针对海上风电机组对故障特征的增量式学习及主动维护的问题，提出了一种基于状态自适应评估的海上风电机组预防性维护策略。首先，采用非正态总体假设检验量化机组实时状态与典型状态的信息差异，通过增量字典学习捕捉机组典型状态特征，基于支持向量机构建了机组状态自适应评估模型。然后，结合部件有效役龄，以机组状态概率向量为决策依据、单次维护费用最小为目标，优化部件维护策略。同时计入部件成组时由于提前或延迟维护的损失，以维护总费用最小为目标、日维护时长为约束，建立了基于状态自适应评估的海上风电机组预防性维护模型。最后，以某海上风电机组为例，验证了所提维护策略的有效性，分析了维护次数、可及性对维护策略的影响。 |
| 关键词: 海上风电机组 状态空间划分 增量字典学习 状态自适应评估 预防性维护 |
| DOI：10.16081/j.epae.202110009 |
| 分类号:TM614;TK83 |
| 基金项目:上海绿色能源并网工程技术研究中心项目 (13DZ2251900)；上海市科技创新行动计划研究项目 (18DZ1202400) |
|
| Preventive maintenance strategy for offshore wind turbine based on state adaptive assessment |
|
FU Yang1, HUANG Luyao1, LIU Lujie1, WEI Shurong1, REN Haohan2, WANG Yi2, TANG Gengpei2
|
|
1.Department of Electrical Engineering, Shanghai University of Electric Power, Shanghai 200090, China;2.Shanghai Green Environmental Protection Energy Co.,Ltd.,Shanghai 200433, China
|
| Abstract: |
| In order to solve the problems of incremental learning of fault characteristics and active maintenance for OWT(Offshore Wind Turbine),a preventive maintenance strategy for OWT based on state adaptive assessment is proposed. Firstly, the non-normal population hypothesis is used to test and quantify the information difference between the real-time state and the typical state of OWT, and the typical state characteristics of OWT are captured by the incremental dictionary learning. Then, an adaptive state assessment model of OWT is built based on support vector machine. Secondly, combined with the effective service life of the components, the component maintenance strategy is optimized with the probability vector of state as the decision constraints and the minimum single maintenance cost as the objective function. At the same time, the preventive maintenance model of OWT based on state adaptive assessment is established with the minimum total maintenance cost as the objective function and the daily maintenance time as the constraints, taking into account the loss caused by early or delayed maintenance when the components are grouped. Finally, an offshore wind turbine is taken as an example to verify the effectiveness of the proposed maintenance strategy, and the influence of maintenance times and accessibility on the maintenance strategy is analyzed. |
| Key words: offshore wind turbine state space partition incremental dictionary learning state adaptive assessment preventive maintenance |
