| 23 | 0 | 72 |
| 下载次数 | 被引频次 | 阅读次数 |
【目的】燃煤电厂输煤系统中传统的人力运维效率低下、作业风险高，且难以满足高精度、及时性的严苛要求，亟需研发一套融合多模态感知技术的全自动化智能运维系统，以实现对关键设备状态的实时监测与精准诊断。【方法】本文以燃煤电厂输煤系统为研究对象，构建了一套包含访问层、访问控制层、应用集成层、子系统层的智能巡检系统逻辑架构，以图像增强算法为核心，融合物联网感知、机器视觉分析、AI算法等前沿技术，对从输煤系统采集的2020—2024年共128万条有效样本进行模型验证，并在“黑灯车间”模式下进行了现场验证。【结果】研究表明，基于多模态感知技术的智能运维系统在有效样本的模拟实验中，模型故障诊断能力达到了84.5%。在燃煤电厂输煤系统的“黑灯车间”模式下，智能运维系统对设备故障、作业人员违规、洒落煤识别这3类事件的平均报警响应时间均≤0.5 s，且误报率≤2%。智能运维系统构建的“功能安全+信息安全+冗余备用”三位一体安全防护体系能够确保输煤系统的全生命周期安全可控。【结论】该技术有效提升了输煤系统巡检的智能化水平与运维可靠性，显著降低了人工干预强度与现场作业安全风险，为火电厂智慧化转型提供了可行技术路径。同时，研究成果可为同类燃煤发电企业燃料系统智能化改造提供参考示范，对推动电力行业安全高效、绿色低碳发展具有重要工程应用价值。
Abstract：[Objective] The traditional human operation and maintenance system in coal-fired power plants has low efficiency and high operation risk, and it is difficult to meet the strict requirements of high precision and timeliness. It is urgent to develop a fully automated intelligent operation and maintenance system that integrates multi-modal sensing technology to realize real-time monitoring and accurate diagnosis of key equipment status. [Methods] This paper takes the coal conveying system of coal-fired power plant as the research object, and constructs a set of logical architecture of intelligent inspection system including access layer, access control layer, application integration layer and subsystem layer. Taking image enhancement algorithm as the core, and integrating advanced technologies such as Internet of Things perception, machine vision analysis and AI algorithm, the model verification of 128 million valid samples collected from the coal conveying system from 2020 to 2024 is carried out, and the field verification is carried out in the ' black light workshop ' mode. [Results] The research shows that the intelligent operation and maintenance system based on multi-modal perception technology has a model fault diagnosis ability of 84.5% in the simulation experiment of effective samples. In the 'lights-out workshop' mode of coal handling system in coal-fired power plant, the average response time of the intelligent operation and maintenance system to the three types of alarm time of fault alarm, operator violation and coal spill identification is less than or equal to 0.5 s, and the false alarm rate is less than or equal to 2%. The three-in-one security protection system of 'functional security + information security + redundant reserve' constructed by the intelligent operation and maintenance system can ensure the safety and controllability of the whole life cycle of the coal transportation system. [Conclusion] The proposed technology effectively improves the intelligence level and operation reliability of inspection in the coal handling system, significantly reduces manual intervention intensity and on-site operation safety risks, and provides a feasible technical path for the intelligent transformation of coal-fired power plants. Meanwhile, the research results can provide a reference for the intelligent transformation of fuel systems in similar coal-fired power enterprises, and possess important engineering application value for promoting the safe, efficient, green and low-carbon development of the electric power industry.
[1]LIU J, DU Z, WANG Q, et al. Critical review of intelligent coalfired power technologies and applications[J]. Engineering, 2026,59:42-59.
[2]XIA X D, YANG Y, SUN Y N, et al. Research status of dust reduction technology in coal handling system of thermal power plant[J]. Journal of Physics:Conference Series, 2021, 2005(1):012143.
[3]张冬练,成佳慧,沈纳昊,等.火电厂智慧煤场管控系统技术研究[J].电力科技与环保, 2020, 36(5):49-53.ZHANG Donglian, CHENG Jiahui, SHEN Nahao, et al. Research on intelligent management and control system of coal yard in thermal power plant[J]. Electric Power Technology and Environmental Protection, 2020, 36(5):49-53.
[4]LIU Z F, WANG C Y, FAN J L, et al. Enhancing the flexibility and stability of coal-fired power plants by optimizing control schemes of throttling high-pressure extraction steam[J]. Energy, 2024, 288:129756.
[5]ISLAM M S, ISLAM M I, MOZUMDER A Q, et al. A conceptual framework for sustainable AI-ERP integration in dark factories:Synthesising TOE, TAM, and IS Success Models for autonomous industrial environments[J]. Sustainability, 2025, 17(20):9234.
[6]YANG Y L, CHEN K F, WU Z Y, et al. Design and application of intelligent unmanned spot inspection system in thermal power plant[J]. Journal of Physics:Conference Series, 2022, 2254(1):012017.
[7]杨志刚.浅析智慧热电厂架构及建设方案[J].工业控制计算机,2024, 37(8):126-128.YANG Zhigang. Analysis on architecture and construction plan of smart thermal power plant[J]. Industrial Control Computer, 2024, 37(8):126-128.
[8]王志敏,黄骞,王可轩,等.智慧电厂大数据云平台的架构建设与模型开发研究[J].动力工程学报, 2025, 45(2):282-291.WANG Zhimin, HUANG Qian, WANG Kexuan, et al. Architecture construction and model development of big-data cloud platform for intelligent coal-fired power plants[J]. Journal of Chinese Society of Power Engineering, 2025, 45(2):282-291.
[9]庄茗贺.基于物联网、大数据、人工智能技术的智能电厂平台建设应用[J].节能技术, 2025, 43(3):256-264.ZHUANG Minghe. Intelligent power plant platform construction application based on internet of things, big data, artificial intelligence technology[J]. Energy Conservation Technology, 2025,43(3):256-264.
[10]ZHANG Z L, NIU C, ZHAO Z B, et al. Small object few-shot segmentation for vision-based industrial inspection[J]. IEEE Transactions on Industrial Informatics, 2025, 21(6):4388-4399.
[11]ZHU X K, DONG L Y, CHEN X P, et al. Confidence guided semisupervised cross-modality person re-identification[J]. Pattern Recognition, 2025, 165:111669.
[12]HAN S, KANG Y, PARK H, et al. Multimodal transformer for property prediction in polymers[J]. ACS Applied Materials&Interfaces, 2024, 16(13):16853-16860.
[13]COSTANZINO A, RAMIREZ P Z, LISANTI G, et al. Multimodal industrial anomaly detection by crossmodal feature mapping[C]//2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR). June 16-22, 2024, Seattle, WA, USA.IEEE, 2024:17234-17243.
[14]ZHAO Y H, LIU Q, SU H, et al. Attention-based multiscale feature fusion for efficient surface defect detection[J]. IEEE Transactions on Instrumentation and Measurement, 2024, 73:5013310.
[15]ZHANG M, SUN Z L, ZHANG Y F. A lightweight YOLOv7-based algorithm for detecting foreign objects on coal conveyors[J].Energy Science&Engineering, 2025, 13(11):5495-5509.
[16]DU G F, ZHANG H Z, LUO Y P, et al. Lightweight method of foreign matter detection in coal conveying based on improved you only look once version 8 and embedded equipment[J].Engineering Applications of Artificial Intelligence, 2026, 172:114374.
[17]ZHANG Y F. Application of dry fog dust suppression technology in the coal conveyer belt system of thermal power plants[J].Advanced Materials Research, 2014, 1073-1076:733-737.
[18]ZHONG Y Y, LUO P X, YAN Y F, et al. PowerGPT:A multimodal foundation model for power inspection[J]. Applied Soft Computing,2026, 186:113939.
[19]谢启源.一种皮带智能巡检系统在国内某大型矿山的应用[J].现代矿业, 2025, 41(9):20-24.XIE Qiyuan. Application of a belt intelligent inspection system in a large mine in China[J]. Modern Mining, 2025, 41(9):20-24.
[20]ZHANG R, FENG X C, YANG L X, et al. Global sparse gradient guided variational Retinex model for image enhancement[J]. Signal Processing:Image Communication, 2017, 58:270-281.
[21]LIU M Z, CHEN J Y, HAN X F. Research on Retinex Algorithm Combining with Attention Mechanism for Image Enhancement[J].Electronics, 2022, 11(22):3695.
[22]KANG Y X. Retinex algorithm and mathematical methods based texture detail enhancement method for panoramic images[J].Mathematical Problems in Engineering, 2022, 2022:1-8.
基本信息:
DOI：10.19944/j.eptep.1674-8069.2026.03.016
中图分类号:TM621
引用信息:
[1]余斌,崔怀明,王文义,等.基于“黑灯车间”模式的输煤系统智能运维技术研发与应用[J].电力科技与环保,2026,42(03):506-516.DOI:10.19944/j.eptep.1674-8069.2026.03.016.
基金信息:
国家能源集团江西电力有限公司项目（SHJ-24-KJ-04）
2026-01-09
2026
2026-06-24
2026
1
2026-06-15
2026-06-15