基于自适应综合权重的混流式水轮机健康状态评估模型

Health assessment model of francis turbine based on adaptive comprehensive weights

【目的】目前,水电行业对机组状态检修的需求十分强烈,因此作为状态检修关键环节的状态评估受到了广泛关注。然而,当前健康状态评估方法存在指导检修项目优化难,确定权重时具有主观盲从性与异常指标容易被淹没的问题。因此,研究提出了基于自适应综合权重的混流式水轮机健康状态评估模型。【方法】首先,建立符合工程实际的混流式水轮机运行状态多重指标体系,分别为设备层、部件层和指标层;其次,根据混流式水轮机不同评价指标的类型,将打分值、监测值与试验值转化为状态量化值;最后,采用S-AHP(Self-adaptive Analytic Hierarchy Process)确定初始相对权重,并引入客观修正因子α来得到综合权重β,计算各个部件以及整个设备的健康值。【结果】选取某水电站3号水轮机组2021年某日的监测数据与检修数据,代入文中模型和恒权重模型,结果显示:采用模型得到当下该混流式水轮机的健康值为0.730 1,设备处于轻微病态,其中水导设备处于中度病态,相比于恒权重模型结果更符合实际。【结论】结果表明:对于混流式水轮机的健康状态评估,模型评价结果比恒权重模型更符合实际,且可以指导具体检修项目的优化,有很大的推广应用价值,为混流式水轮机的状态检修提供参考。

[Objective]At present,the demand for condition-based maintenance of units in the hydropower industry is very strong.So condition assessment,as a key part in condition-based maintenance,is paid widespread attention to.However,the current method of health assessment face problems.It has difficulty in guiding the optimization of maintenance projects,subjective blind obedience in determining weights,and the negligence of abnormal indicators to be overwhelmed.Therefore,this article proposes a health assessment model of Francis turbine based on adaptive comprehensive weights.[Methods]Firstly,based on actual situation,a multi-indicator system of Francis turbine is established.This system is divided into equipment layer,component layer and indicator layer.Secondly,based on the types of indicators for Francis turbines,the scoring values,monitoring values,and test values are respectively converted into quantitative values for health assessment;Finally,S-AHP(Self-adaptive Analytic Hierarchy Process,S-AHP) is used to determine the initial relative weights and introduce objective correction factors α to obtain comprehensive weights β.The health values of entire equipment and their components are calculated.[Results]Based on the unit 3 of a certain hydropower station,its monitoring and maintenance data on a certain day in 2021 were substituted into the health assessment model and the constant weight model.The result show that the health value of the Francis turbine is 0.730 1 by using the model in this paper.The result indicates the equipment and its turbine guide bearing is in a mild ill state.Besides,it is more realistic than the constant weight model result.[Conclusion]In conclusion,the evaluation result of the health assessment model in this paper is more realistic than the constant weight model for Francis turbines.And the result can guide the optimization of maintenance projects.This model is worth promoting and applying due to the valuable reference for the condition-based maintenance of Francis turbines.