|
| 基于DEGAN和SHAP的水电机组异常检测研究 |
| Research on anomaly detection of hydropower units based on DEGAN and SHAP |
| 投稿时间：2025-04-07 修订日期：2025-05-12 |
| DOI： |
| 中文关键词: 水电机组 异常检测 DEGAN SHAP 受油器摆度 |
| 英文关键词:Hydropower unit anomaly detection DEGAN SHAP oil head swing amplitude |
| 基金项目:中国水利水电科学研究院基本科研项目(AU0145C012024、AU0145B012021) |
|
| 摘要点击次数: 43 |
| 全文下载次数: 0 |
| 中文摘要: |
| 针对水电站监控系统中固定阈值法在复杂工况下异常检测灵敏度低、难以实现有效预警的问题，本文提出了一种基于生成对抗网络判别器与密度估计的特征增强异常检测模型（FEAD-DEGAN）。该模型将卷积与全局平均池化方法用于优化判别器结构，强化时序特征提取，并结合动态阈值策略与核密度估计算法，提升检测灵敏度。以轴流转浆式机组受油器摆度异常数据样本为研究对象，通过与孤立森林和自编码器方法进行对比验证，所提方法在异常检测成功率与误报率方面表现更优。进一步采用SHAP解释方法对监测指标的异常贡献度进行量化，提升了模型的可解释性，有助于故障溯因和检修策略优化，有效辅助了水电机组设备故障诊断和智慧化运维。 |
| 英文摘要: |
| In hydropower station monitoring systems, fixed threshold methods are commonly used for over-limit alarms, but they exhibit low sensitivity in complex conditions, making early warnings difficult. This paper proposes a feature-enhanced anomaly detection (FEAD-DEGAN) model based on generative adversarial network discriminator and density estimation (DEGAN). Convolution and global average pooling methods optimize the discriminator structure, enhancing time-series feature extraction. The dynamic threshold strategy and kernel density estimation improve detection sensitivity. The model is validated with abnormal oil head swing amplitude data from an axial-flow pump-turbine unit. Compared with Isolation Forest and Autoencoder, the proposed approach shows better performance in anomaly detection success rate and false alarm rate. SHAP quantifies the contribution of monitoring indicators to anomalies, identifying key factors that influence abnormal behavior and enhancing process interpretability. This supports root cause analysis and facilitates the optimization of maintenance strategies, thereby contributing to more effective fault diagnosis and intelligent maintenance. |
| 查看/发表评论 下载PDF阅读器 |
| 关闭 |
