您好,欢迎来到login_星欧(官方)代理招商直营娱乐!

首页> 《中国测试》期刊 >本期导读>基于遗传粒子群优化的热负荷预测方法

基于遗传粒子群优化的热负荷预测方法

340    2024-06-26

¥0.50

全文售价

作者:谢文举1, 薛贵军1,2, 史彩娟3, 李水清2

作者单位:1. 华北理工大学电气工程学院,河北 唐山 063210;
2. 华北理工大学智能仪器厂,河北 唐山 063210;
3. 华北理工大学人工智能学院,河北 唐山 063210


关键词:完全噪声辅助聚合经验模态分解;样本熵;遗传粒子群;一次网;热负荷预测


摘要:

集中供热系统一次网热负荷预测对于换热站能源合理分配具有重大意义,针对换热站之间存在较强的耦合性,如何保证换热站实现节能减排的同时并保证热用户的舒适性是供热行业的首要任务,提出一种基于遗传粒子群的混合神经网络(GAPSO-CNN-BiLSTM)预测模型。首先,利用热负荷历史值、供水流量、供水温度以及回水压力构建模型输入;然后,利用完全噪声辅助聚合经验模态分解(complete ensemble empirical mode decomposition with adaptive noise,CEEMDAN)将供热负荷分解为不同子序列,以弱化供热负荷数据复杂程度,挖掘数据内部潜在特征;其次,为进一步减少模型计算时间,根据样本熵(sample entropy,SE)对子序列进行合并;最后,利用所提模型对不同子序列进行预测重构。实验表明所提模型相比LSTM、CNN-LSTM以及粒子群优化的混合神经网络(PSO-CNN-LSTM)在供热负荷预测中精度分别提高42%,32%,30%拥有更出色的特征提取能力和精度。


Heat load prediction method based on genetic particle swarm optimization
XIE Wenju1, XUE Guijun1,2, SHI Caijuan3, LI Shuiqing2
1. College of Electrical Engineering, North China University of Technology, Tangshan 063210, China;
2. Instrument Factory, North China University of Technology, Tangshan 063210, China;
3. College of Artificial Intelligence, North China University of Technology, Tangshan 063210, China
Abstract: The primary network heat load prediction of centralized heating system is of great significance for the reasonable energy distribution of heat exchange stations. In view of the strong coupling between heat exchange stations, how to ensure the energy saving and emission reduction of heat exchange stations while ensuring the comfort of heat users is the top priority of the heating industry, a hybrid neural network (GAPSO-CNN-BiLSTM) prediction model based on genetic particle swarm is proposed. First, the model input is constructed using the heat load history, water supply flow, water supply temperature, and return water pressure; then, the heat load is decomposed into different subseries using complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) to weaken the complexity of the heat load data and to explore the potential features within the data; second, to further reduce the model computation Secondly, in order to further reduce the model computation time, the subsequences are combined according to Sample Entropy (SE); finally, the proposed model is used to reconstruct the prediction of different subsequences. The experiments show that the proposed model has 42%, 32%, and 30% better feature extraction ability and accuracy than LSTM, CNN-LSTM, and hybrid neural network with particle swarm optimization (PSO-CNN-LSTM) in heating load prediction, respectively.
Keywords: CEEMDAN; SE; GAPSO; first network; heat load prediction
2024, 50(6):131-138,147 收稿日期: 2022-08-24;收到修改稿日期: 2022-11-24
基金项目: 国家自然科学基金(61502143)
作者简介: 谢文举(1998-),男,江苏苏州市人,硕士研究生,专业方向为集中供热系统的智能控制及应用。
参考文献
[1] 江亿, 唐孝炎, 倪维斗,等. 北京PM2.5与冬季采暖热源的关系及治理措施[J]. 中国能源, 2014, 36(1): 7-13+28.
JIANG Y, TANG X Y, NI W D, et al. The relationship and treatment measures between Beijing PM2.5 and winter heating sources [J].Energy of China, 2014, 36 (1): 7-13+28.
[2] VOGLER–FINCK P J C, BACHER P, MADSON H. Online short-term forecast of greenhouse heat load using a weather forecast service[J]. Applied Energy, 2017, 205: 1298-1310.
[3] 顾伟, 薛贵军, 李水清. 结合天气因素的HTS-FOA-GRNN短期热负荷预测研究[J]. 机械设计与制造, 2022(1): 81-83+87.
GU W, XUE G J, LI S Q. Research on short term heat load prediction of HTS-FOA-GRNN combined with weather factors [J].Machinery Design and Manufacture, 2022 (1): 81-83+87.
[4] 许裕栗, 姜娜, 陈卓, 等. 基于LSTM网络的居民区供热负荷短时预测[J]. 自动化仪表, 2018, 39(10): 1-5+10.
XU Y L, JIANG N, CHEN Z, et al. Short term prediction of heating load in residential areas based on LSTM network [J].Process Automation Instrumation, 2018, 39 (10): 1-5+10.
[5] KIM T Y, CHO S B. Predicting residential energy consumption using CNN-LSTM neural networks[J]. Energy, 2019, 182: 72-81.
[6] CAO J, LI Z, LI J. Financial time series forecasting model based on CEEMDAN and LSTM[J]. Physica A:Statistical mechanics and its applications, 2019, 519: 127-139.
[7] ZHOU F, HUANG Z, ZHANG C. Carbon price forecasting based on CEEMDAN and LSTM[J]. Applied Energy, 2022, 311: 118601.
[8] TORRES M E, COLOMINAS M A, SCHLOTTHAUER G, et al. A complete ensemble empirical mode decomposition with adaptive noise[C]//2011 IEEE international conference on acoustics, speech and signal processing (ICASSP). IEEE, 2011: 4144-4147.
[9] 石琳,刘文慧,曹富军,等.基于CEEMDAN-SVM-LSTM的高炉煤气利用率组合预测[J].中国测试,2023,49(1):86-91.
SHI L, LIU W H, CAO F J, et al. Combination prediction of blast furnace gas utilization based on CEEMDAN-SVM-LSTM [J].China Measurement& Test, 2023,49 (1): 86-91.
[10] 张林, 刘继春. 基于EEMD-SE和PSO-KELM的短期负荷区间预测方法[J]. 中国电力, 2021, 54(3): 132-140.
ZHANG L, LIU J C. Short term load interval prediction method based on EEMD-SE and PSO-KELM [J].Electric Power, 2021, 54 (3): 132-140.
[11] 徐岩, 靳伟佳, 朱晓荣. 基于遗传粒子群算法的光伏并网逆变器参数辨识[J]. 太阳能学报, 2021, 42(7): 103-109.
XU Y, JIN W J, ZHU X R. Parameter identification of photovoltaic grid connected inverters based on genetic particle swarm optimization algorithm [J].Acta Energiae Solaris Sinica, 2021, 42 (7): 103-109.
[12] 党建武, 从筱卿. 基于CNN和GRU的混合股指预测模型研究[J]. 计算机工程与应用, 2021, 57(16): 167-174.
DANG J W, CONG X Q.Research on a hybrid stock index prediction model based on CNN and GRU [J].Computer Engineering and Applications, 2021, 57 (16): 167-174.
[13] RAFI S H, DOOBA S R, HOSSAIN E. A short-term load forecasting method using integrated CNN and LSTM network[J]. IEEE Access, 2021, 9: 32436-32448.
[14] 廖清阳, 王军, 胡凯强,等. 基于深度并行CNN-BiLSTM的能源互联网电负荷和热负荷联合预测模型[J]. 中国测试, 2022, 48(4): 146-153.
LIAO Q Y, WANG J, HU K Q, et al. Joint forecasting model of electric load and thermal load of energy internet based on deep parallel CNN-BiLSTM [J].China Measurement& Test, 2022, 48 (4): 146-153.