为了解决现有水轮机异常检测技术参数量较大的问题,本文提出了一种基于改进时延神经网络的水轮机异常检测算法.首先将实时采集到的水轮机的音频数据进行基本的信号处理,然后计算Fbank特征,将其作为模型的输入数据,使用简单无参数注意力模块(simple parameter-free attention module,SimAM)优化后的ECAPA-TDNN(emphasized channel attention, propagation and aggregation in TDNN)算法对其进行训练,得到输出特征向量,然后对样本数据采用同样的处理方式,最终得到实时数据和样本数据的输出特征向量,然后计算对应的欧几里得相似度,并依据初始设定的相似度阈值来判断是否异常,如有异常,则与样本集中的典型异常状态进行比较,用于对实时监测到的异常状态进行分类.与主流的卷积神经网络(convolutional neural network, CNN)、ResNet以及单独的ECAPA-TDNN等模型的对比结果可知,本文所建立的模型在参数量、算法准确率和等误率性能上均存在较大的优势和提升,参数量为19.7 M,模型准确率达到99.12%,模型等误率低至0.9%,可以很好地满足水轮机异常检测的需求.
Aiming at the problem of excessive parameters in existing anomaly detection techniques for hydraulic turbines, this paper proposes an anomaly detection algorithm for hydraulic turbines based on an improved time-delay neural network. Firstly, the real-time collected audio data of hydraulic turbines are subjected to basic signal processing, and then Fbank features are calculated as the input data of the model. The ECAPA-TDNN(Emphasized Channel Attention, Propagation and Aggregation in TDNN)algorithm- 引用格式:何梦腾,潘浩文,邓洪兴,等.基于点云多尺度方向一致性的羊只体尺测量方法［J］.河北大学学报(自然科学版),2026,46(2):113-127.引用格式:尹宁,王小伟,魏鹏程,等.基于改进时延神经网络的水轮机异常声纹检测算法［J］.河北大学学报(自然科学版),2026,46(2):215-224.DOI:10.3969/j.issn.1000-1565.2026.02.010基于改进时延神经网络的水轮机异常声纹检测算法尹宁1,王小伟1,魏鹏程1,周九道1,施琳长弘1,廖勇2(1.贵州乌江水电开发有限责任公司 构皮滩发电厂,贵州 遵义 564408;2.重庆大学 微电子与通信工程学院,重庆 400044)摘 要:为了解决现有水轮机异常检测技术参数量较大的问题,本文提出了一种基于改进时延神经网络的水轮机异常检测算法.首先将实时采集到的水轮机的音频数据进行基本的信号处理,然后计算Fbank特征,将其作为模型的输入数据,使用简单无参数注意力模块(simple parameter-free attention module,SimAM)优化后的ECAPA-TDNN(emphasized channel attention, propagation and aggregation in TDNN)算法对其进行训练,得到输出特征向量,然后对样本数据采用同样的处理方式,最终得到实时数据和样本数据的输出特征向量,然后计算对应的欧几里得相似度,并依据初始设定的相似度阈值来判断是否异常,如有异常,则与样本集中的典型异常状态进行比较,用于对实时监测到的异常状态进行分类.与主流的卷积神经网络(convolutional neural network, CNN)、ResNet以及单独的ECAPA-TDNN等模型的对比结果可知,本文所建立的模型在参数量、算法准确率和等误率性能上均存在较大的优势和提升,参数量为19.7 M,模型准确率达到99.12%,模型等误率低至0.9%,可以很好地满足水轮机异常检测的需求.关键词:水电设备;异常检测;注意力模块;时延神经网络;深度学习中图分类号:TP391.41;TU714 文献标志码:A 文章编号:1000-1565(2026)02-0215-10DOI:10.3969/j.issn.1000-1565.2026.02.010Anomaly acoustic detection algorithm for hydraulic turbine based on improved time-delay neural networkYIN Ning1, WANG Xiaowei1, WEI Pengcheng1, ZHOU Jiudao1, SHILIN Changhong1, LIAO Yong2(1.Goupitan Power Plant, Guizhou Wujiang Hydropower Development Co.,Ltd.,Zunyi 564408,China;2.School of Microelectronics and Communication Engineering, Chongqing University, Chongqing 400044, China)Abstract: Aiming at the problem of excessive parameters in existing anomaly detection techniques for hydraulic turbines, this paper proposes an anomaly detection algorithm for hydraulic turbines based on an improved time-delay neural network. Firstly, the real-time collected audio data of hydraulic turbines are subjected to basic signal processing, and then Fbank features are calculated as the input data of the model. The ECAPA-TDNN(Emphasized Channel Attention, Propagation and Aggregation in TDNN)algorithm- 收稿日期:2025-05-12;修回日期:2025-06-11 基金项目:中国华电集团科技项目(GPTDC/2024-1209) 第一作者:尹宁(1991—),男,贵州乌江水电开发有限责任公司高级工程师,主要从事水轮发电机机械检修方向研究.E-mail:1009683793@99.com 通信作者:廖勇(1982—),男,重庆大学副研究员,博士,主要从事移动通信、人工智能及其应用方向研究.E-mail:liaoy@cqu.edu.cn 第2期尹宁等:基于改进时延神经网络的水轮机异常声纹检测算法河北大学学报(自然科学版) 第46卷optimized by the Simple Parameter-free Attention Module(SimAM)is employed for model training to obtain the output feature vectors. The same processing procedure is applied to the sample data. Subsequently, the Euclidean similarity is calculated between the feature vectors of real-time data and sample data, and anomaly judgment is conducted according to a preset similarity threshold. If an anomaly is detected, it is further classified by comparing with typical abnormal states in the sample set. Compared with mainstream models such as Convolutional Neural Network(CNN), ResNet and the standalone ECAPA-TDNN, the proposed model shows significant advantages and improvements in terms of parameter quantity, algorithm accuracy and equal error rate(EER). The model achieves 19.7M parameters, 99.12% accuracy and 0.9% EER, which can well meet the requirements of anomaly detection for hydraulic turbines.
[1] 陈敏,刘卫国,徐勇晖.智能化技术在水电站中的应用研究[J].电子元器件与信息技术,2024,8(6):88-89.
[2] 杨博瀚,王子元,史宏博.人工智能在水电设备故障预测与维护中的应用效果分析[J].智能建筑与智慧城市,2025,(增刊2):404-406.
[3] 李林, 陶春华, 张鹏程, 等. 基于知识图谱与深度学习的水电设备智能诊断[J]. 微型电脑应用, 2025, 41(11): 138-142. DOI:10.3969/j.issn.1007-757X.2025.11.032.
[4] 吴宝富, 张海库, 耿清华, 等. 水电领域现状及设备挑战与更新[J]. 水电与新能源, 2025, 39(9): 1-5. DOI:10.13622/j.cnki.cn42-1800/tv.1671-3354.2025.09.001.
[5] 熊中浩. 基于模型的水电设备工况及趋势预警系统研究与应用[J]. 云南水力发电, 2025, 41(8): 112-115. DOI:10.3969/j.issn.1006-3951.2025.08.032.
[6] 汲晓飞. 基于人工智能的水利水电设备状态监测与预测技术研究[J]. 中国高新科技, 2024(10): 31-33. DOI:10.13535/j.cnki.10-1507/n.2024.10.07.
[7] 周革.大数据技术下水电站设备运行状态监测研究[J].电力设备管理,2024(21):148-150.
[8] 蒲黔辉,张子怡,肖图刚,等.基于时间序列压缩分割的监测数据异常识别算法研究[J].桥梁建设,2024,54(3):15-23.DOI:10.20051/j.issn.1003-4722.2024.03.003.
[9] 林泊锟,丁勇,李登华.一种基于深度学习与图像局部特征提取的边坡异常监测技术[J].测绘通报,2024(4):23-28. DOI:10.13474/j.cnki.11-2246.2024.0405.
[10] 殷鹏程,谭曼丽莎,曹阳梅,等.基于迁移学习图像识别的桥梁监测数据异常检测方法[J].大连交通大学学报,2024,45(3):106-113.DOI:10.13291/j.cnki.djdxac.2024.03.017.
[11] 孙政杰,丁勇,李登华.基于Prophet-GMM的大坝监测数据异常检测算法[J].人民黄河,2024,46(3):132-135. DOI:10.3969/j.issn.1000-1379.2024.03.024.
[12] 黎祎,赵二峰,何菁.大坝变形监测异常数据识别和重构模型[J].水利水电科技进展,2023,43(2):109-114. DOI:10.3880/j.issn.1006-7647.2023.02.017.
[13] 李若峰,付卫宁.一种基于音频数据特征提取和单分类器的电机异常监测方法[J].电工技术,2024(3):183-187.DOI:10.19768/j.cnki.dgjs.2024.03.050.
[14] 杜文彬,黄东,纪宏彦.基于最小二乘支持向量机的水电机组设备异常监测告警方法[J].电气技术与经济,2024(12):77-80.DOI:10.3969/j.issn.1673-8845.2024.12.023.
[15] 杨伟业,陈漾.基于流形学习的多模态水轮机温度时间序列异常监测[J].科学技术创新,2023(26):13-16. DOI: 10.3969/j.issn.1673-1328.2023.26.005.
[16] 徐韧, 苏怀智, 杨立夫. 基于GP-XGBoost的大坝变形预测模型[J]. 水利水电科技进展, 2021,41(5): 41-46. DOI:10.3880/j.issn.1006-7647.2021.05.007.
[17] Yu Z, Wei Y, Niu B, et al. Automatic condition monitoring and fault diagnosis system for power transformers based on voiceprint recognition[J]. IEEE Trans Instrum Meas, 2024,73: 9600411. DOI:10.1109/tim.2024.3384551.
[18] Miao Y, Zhang B, Li C, et al. Feature mode decomposition: New decomposition theory for rotating machinery fault diagnosis[J]. IEEE Trans Electron, 2022, 70(2): 1949-1960. DOI:10.2139/ssrn.5104568.
[19] 张航,张广东,刘康,等.声纹检测技术在变压器故障诊断中的应用及展望[J].机械研究与应用,2022,35(5):243-246. DOI:10.16576/j.ISSN.1007-4414.2022.05.063.
[20] Waibel A, Hanazawa T, Hinton G, et al. Phoneme recognition using time-delay neural networks[M].Backpropagation. Psychology Press, 2013: 35-61. DOI: 10.1109/ijcnn.1989.118586.
[21] Desplanques B, Thienpondt J, Demuynck K. Ecapa-Tdnn: Emphasized channel attention, propagation and aggregation in TDNN based speaker verification[EB/OL]. arXiv preprint arXiv:2005.07143, 2020.DOI:10.21437/interspeech.2020-2650.
[22] Yang L, Zhang R Y, Li L, et al. SimAM: A simple, parameter-free attention module for convolutional neural networks[C] //International conference on machine learning, PMLR, 2021: 11863-11874.
[23] 王茂,何勇.基于FBank特征与改进CNN的声纹识别研究[J].智能计算机与应用,2024,14(8):40-47. DOI:10.20169/j.issn.2095-2163.240807.
[24] He K, Zhang X, Ren S, et al. Deep residual learning for image recognition[C] //Proceedings of the IEEE conference on computer vision and pattern recognition. 2016: 770-778. DOI:10.1109/cvpr.2016.90.
[25] 万玫汐,王华朋,闫道申,等.基于改进ECAPA-TDNN的法庭自动说话人识别[J].科学技术与工程,2024,24(27):11763-11773. DOI:10.12404/j.issn.1671-1815.2307966. (
