摘要
为对水电机组运行状态进行准确的评估,从而保障水电机组安全运行,文章提出一种基于DAE–GBDT健康模型的水电机组劣化程度评估方法。利用深度自编码器(DAE)对工况参数中的关键信息进行压缩凝练。用梯度提升决策树(GBDT)建立健康模型,学习振摆值与工况参数之间的潜在关系。根据所构造健康模型和机组运行数据得到机组的劣化情况。通过某抽水蓄能机组实例,验证了所提模型具有更高精度,能生成可靠的劣化趋势。
To accurately assess the operational status of hydropower units and ensure their safe operation,this paper proposes a method for evaluating the degradation degree of hydropower units based on a health model using Deep Autoencoder(DAE)and Gradient Boosting Decision Trees(GBDT).The deep autoencoder(DAE)is utilized to compress and refine critical information from operating parameters.Subsequently,a health model is established using gradient boosting decision trees(GBDT)to learn the potential relationship between vibration values and operating parameters.Finally,the degradation condition of the unit is obtained based on the constructed health model and operational data.The proposed model is validated using a case study of a pumped storage unit,demonstrating higher accuracy and the ability to generate reliable degradation trends.
作者
林峰平
陈亦真
LIN Fengping;CHEN Yizhen
出处
《电力系统装备》
2024年第10期56-57,66,共3页
Electric Power System Equipment
基金
湖北省重点研发计划项目(2023BAB209)。
