张文龙,朱培毅,朱宪然,焦宏波,杨磊,周亚男,王熙
ZHANG Wenlong,ZHU Peiyi,ZHU Xianran,JIAO Hongbo,YANG Lei,ZHOU Yanan,WANG Xi

• 1:阳城国际发电有限责任公司
• 2:华北电力大学动力工程系
• 3:中国大唐集团科学技术研究总院有限公司华北电力试验研究院

摘要(Abstract)：

以掌握凝汽器传热性能为目标，采用基于CNN-BiGRU-AT（基于注意力机制的卷积神经网络和双向门控循环单元）的模型对凝汽器传热系数进行预测。以某电厂1 000 MW机组的凝汽器作为研究对象，首先分析凝汽器特性，构建凝汽器传热系数计算的理论模型；然后通过CNN（卷积神经网络）对传热特性影响因素的数据特征进行提取，并利用BiGRU（双向门控循环单元）提取数据间长短期依赖关系；最后使用AT（注意力机制）来突出特征中的重要部分，由此构建了基于CNN-BiGRU-AT的凝汽器传热系数预测模型。分析计算结果表明，与传统的理论计算及BP（反向传播）网络、GRU（门控循环单元）网络相比，采用CNNBiGRU-AT模型的计算结果更加精确且规律性好。
To master the heat transfer performance of the condensers, a model based on convolutional neural network and bidirectional gated recurrent unit with attention mechanism(CNN-BiGRU-AT) is adopted to predict their heat transfer coefficients. Taking the condenser of a 1,000 MW unit in a power plant as the research subject, the condenser characteristics are first analyzed, and a theoretical model for calculating its heat transfer coefficients is constructed. Then, CNN is employed to extract the data features of factors influencing the heat transfer characteristics, while BiGRU is utilized to capture the long and short-term dependencies between the data. Finally, AT is used to highlight the important parts of the features, thereby constructing a prediction model for heat transfer coefficients based on CNN-BiGRU-AT. The analysis of the computational results indicates that compared to traditional theoretical calculations and BP(backpropagation) networks, as well as GRU networks, the computational results obtained using the CNN-BiGRU-AT model are more accurate and exhibit better regularity.

关键词(KeyWords)： 传热系数;卷积神经网络;双向门控循环单元;凝汽器
heat transfer coefficient;CNN;BiGRU;condenser

Abstract:

Keywords:

基金项目(Foundation): 河北省自然科学基金(E2018502059)

作者(Author): 张文龙,朱培毅,朱宪然,焦宏波,杨磊,周亚男,王熙
ZHANG Wenlong,ZHU Peiyi,ZHU Xianran,JIAO Hongbo,YANG Lei,ZHOU Yanan,WANG Xi

DOI: 10.19585/j.zjdl.202409013

参考文献(References)：

• [1]葛晓霞，肖洪闯，嵇卫，等.基于果蝇算法优化广义回归神经网络的凝汽器真空预测[J].汽轮机技术，2018,60(3):208-212.GE Xiaoxia,XIAO Hongchuang,JI Wei,et al. Vacuum prediction of condensers based on generalized regression neural network optimized by Fruit fly algorithm[J].Steam Turbine Technology,2018,60(3):208-212.
• [2] SAARI J,KAIKKO J,VAKKILAINEN E,et al. Comparison of power plant steam condenser heat transfer models for on-line condition monitoring[J]. Applied Thermal Engineering,2014,62(1):37-47.
• [3] MEDICA-VIOLA V,PAVKOVI?B,MRZLJAK V.Numerical model for on-condition monitoring of condenser in coal-fired power plants[J]. International Journal of Heat and Mass Transfer,2018,117:912-923.
• [4] XIA M,WANG K,SONG W,et al. Non-intrusive load disaggregation based on composite deep long short-term memory network[J]. Expert Systems with Applications,2020,160:113669.
• [5] XIAO Z,GANG W,YUAN J,et al.Cooling load disaggregation using a NILM method based on random forest for smart buildings[J]. Sustainable Cities and Society,2021,74:103202.
• [6] CHEN C,XIE D M,XIONG Y H,et al.Optimization of turbine cold-end system based on BP neural network and genetic algorithm[J]. Frontiers in Energy,2014,8(4):459-463.
• [7] AZIZ N L A A,YAP K S,BUNYAMIN M A.A hybrid fuzzy logic and extreme learning machine for improving efficiency of circulating water systems in power generation plant[J].IOP Conference Series:Earth and Environmental Science,2013,16(1):012102-012102.
• [8]葛晓霞，赵舒莹，肖洪闯，等.基于改进果蝇算法优化SVM的凝汽器真空预测[J].热能动力工程，2020,35(11):39-45.GE Xiaoxia,ZHAO Shuying,XIAO Hongchuang,et al.Prediction of condenser vacuum based on SVM improved by IFOA[J].Thermal Power Engineering,2020,35(11):39-45.
• [9]吴伟，冯林魁，王平，等.基于PSO-LSSVM的双压凝汽器真空建模[J].机械研究与应用，2021,34(4):128-130.WU Wei,FENG Linkui,WANG Ping,et al. Vacuum modeling for dual-pressure condensers based on PSOLSSVM model[J].Mechanical Research and Application,2021,34(4):128-130.
• [10]徐乃华.660 MW机组凝汽器真空偏低原因诊断分析[J].电站系统工程，2020,36(2):59-61.XU Naihua.Causes diagnosis analysis of low vacuum for condenser of 660 MW unit[J].Power Plant System Engineering,2020,36(2):59-61.
• [11]王建刚.火电机组冷端系统运行优化的研究[D].上海：上海电力学院，2016.WANG Jiangang. Research on operation optimization of cold end system operation in thermal power units[D].Shanghai Electric Power University,2016.
• [12]张娜.1 000 MW机组凝汽器真空提高方法研究[D].北京：华北电力大学，2017.ZHANG Na. Research on vacuum improvement of condenser system in 1 000 MW unit[D].Beijing:North China Electric Power University,2017.
• [13]石磊，王毅，成颖，等.自然语言处理中的注意力机制研究综述[J].数据分析与知识发现，2020,4(5):1-14.SHI Lei,WANG Yi,CHENG Ying,et al.Review of attention mechanisms in natural language processing[J]. Data Analysis and Knowledge Discovery,2020,4(5):1-14.
• [14]魏健，赵红涛，刘敦楠，等.基于注意力机制的CNNLSTM短期电力负荷预测方法[J].华北电力大学学报（自然科学版），2021,48(1):42-47.WEI Jian,ZHAO Hongtao,LIU Dunnan,et al.Short-term power load forecasting method by attention-based CNNLSTM[J].Journal of North China Electric Power University(Natural Science Edition),2021,48(1):42-47.
• [15]姚程文，杨苹，刘泽健.基于CNN-GRU混合神经网络的负荷预测方法[J].电网技术，2020,44(9):3416-3424.YAO Chengwen,YANG Ping,LIU Zejian.Load forecasting method based on CNN-GRU hybrid neural network[J].Power Grid Technology,2020,44(9):3416-3424.
• [16]张宸嘉，朱磊，俞璐.卷积神经网络中的注意力机制综述[J].计算机工程与应用，2021,57(20):64-72.ZHANG Chenjia,ZHU Lei,YU Lu.Review of attention mechanisms in convolutional neural networks[J]. Computer Engineering and Applications,2021,57(20):64-72.
• [17]李杨，井新经，王宏武，等.基于效能-传热单元数的回转式空预器换热性能计算方法[J].热力发电，2019,48(1):73-76.LI Yang,JING Xinjing,WANG Hongwu,et al. Calculation method for heat transfer efficiency of rotary air preheaters based on effectiveness-NTU[J]. Thermal Power Generation,2019,48(1):73-76.
• [18]王玮，曾德良，杨婷婷，等.基于凝汽器压力估计算法的循环水泵最优运行[J].中国电机工程学报，2010,30(14):7-12.WANG Wei,ZENG Deliang,YANG Tingting,et al.Optimal operation of circulating water pumps based on condenser pressure estimation algorithm[J]. Chinese Journal of Electrical Engineering,2010,30(14):7-12.