1. School of Building Services Science and Engineering, Xi'an University of Architecture and Technology, Xi'an 710055, China 2. School of Information and Control Engineering, Xi'an University of Architecture and Technology, Xi'an 710055, China
A hybrid prediction model which combining radial basis function (RBF) neural network optimized by tree-seed algorithm (TSA) and long short-term memory (LSTM) neural network was proposed, in order to improve the accuracy, robustness, and generalization ability of building energy consumption prediction. Firstly, the complete ensemble empirical mode decomposition with adaptive noise algorithm was used to decompose the building energy consumption data into a group of intrinsic mode function (IMF) components and a residual component, and the components were divided into high-frequency components and low-frequency components by sample entropy algorithm. Then, least absolute contraction and selection operator (LASSO) method was used for feature selection. Finally, the RBF model optimized by the TSA algorithm and the LSTM model were used to predict the low-frequency components and high-frequency components respectively, and the final prediction results were obtained by superposition and reconstruction. Model evaluation results showed that the accuracy of the hybrid prediction model was 98.72%. Compared with RBF, TSA-RBF, and LSTM models, the prediction effect of the hybrid model is better. Meanwhile, the model has strong robustness and generalization ability, and can be more effectively used for hourly building electricity consumption prediction.
Jun-qi YU,Si-yuan YANG,An-jun ZHAO,Zhi-kun GAO. Hybrid prediction model of building energy consumption based on neural network. Journal of ZheJiang University (Engineering Science), 2022, 56(6): 1220-1231.
Fig.1Flow chart of energy consumption data decomposition by CEEMDAN
Fig.2Structure of RBF neural network
Fig.3Flow chart of TSA algorithm
Fig.4Structure of LSTM neural network
Fig.5Modeling process of TSA-RBF-LSTM
变量名称
符号
单位
流动人员数量
β1
人/m 2
照明使用
β2
kW·h
电气使用
β3
kW·h
室外干球温度
β4
℃
室外湿球温度
β5
℃
室外露点温度
β6
℃
室外相对湿度
β7
%
风速
β8
m/s
风向
β9
(°)
太阳辐照度
β10
W/m 2
上一时刻太阳辐照度
β11
W/m 2
上一时刻能耗
β12
kW·h
能耗
WS
kW·h
Tab.1Input and output variables in sample data
Fig.6Sample data of building energy consumption
Fig.7Decomposition results of building energy consumption by CEEMDAN
IMF分量
所选特征
IMF 1
β1、 β2、 β3、 β4、 β7、 β8、 β10、 β11、 β12
IMF 2
β1、 β2、 β3、 β4、 β7、 β11、 β12
IMF 3
β1、 β2、 β3、 β4、 β5、 β7、 β11、 β12
IMF 4
β1、 β2、 β3、 β4、 β11、 β12
IMF 5
β1、 β2、 β3、 β11、 β12
IMF 6
β1、 β3、 β11、 β12
IMF 7
β1、 β3、 β11、 β12
IMF 8
β1、 β2、 β3、 β4、 β7、 β11、 β12
IMF 9
β1、 β2、 β3、 β4、 β7、 β11、 β12
Tab.2Feature selection results by LASSO
Fig.8Relationship between tuning parameter $\lambda $ and each component ${\;\beta _i}$ of coefficient vector
子算法
参数符号
说明
数值
TSA
Npop
种群规模
50
ST
搜索趋势
0. 1
d
优化维数
n+( n+1) m
Miter
最大迭代次数
500
RBF
n
输入层神经元数量
特征数
m
隐含层神经元数量
n
c
中心取值范围
[?5.0,5.0]
b
基宽取值范围
[0. 01,10.00]
w
权值取值范围
[?2.0,2.0]
LSTM
NIU
输入层神经元数量
特征数
NHU
隐含层神经元个数
16 NIU
RIL
初始学习率
0. 005
ME
最大训练次数
200
Tab.3Parameter setting of TSA-RBF-LSTM model
Fig.9Comparison of predictive results for four prediction models
Fig.10Correlation of actual values and predicted values for four models
预测模型
ERMSE
EMAPE/%
RBF
26.5549
5.37
TSA-RBF
15.7670
2.96
LSTM
18.6225
3.36
TSA-RBF-LSTM
6.8925
1.28
Tab.4Comparison of prediction accuracy for four models
预测模型
是否CEEMDAN分解
ERMSE
EMAPE/%
RBF
是
27.254 7
5.41
否
42.362 5
8.79
LSTM
是
19.392 1
3.32
否
28.745 9
5.68
Tab.5Comparison of prediction accuracy for two models
Fig.11Comparison of predictive results of RBF and LSTM for each order of IMF components
Fig.12Iterative comparison of different optimization algorithms
Fig.13Absolute error box-plot of four models
Fig.14Proof of generalization ability
[1]
中国建筑节能协会 中国建筑能耗研究报告2020[J]. 建筑节能(中英文), 2021, 49 (2): 1- 6 China Association of Building Energy Efficiency China building energy consumption annual report 2020[J]. Building Energy Efficiency, 2021, 49 (2): 1- 6
[2]
DONG Z, LIU J, LIU B, et al Hourly energy consumption prediction of an office building based on ensemble learning and energy consumption pattern classification[J]. Energy and Buildings, 2021, 241: 110929
doi: 10.1016/j.enbuild.2021.110929
[3]
AMASYALI K, EL-GOHARY N M A review of date-driven building energy consumption prediction studies[J]. Renewable and Sustainable Energy Reviews, 2018, 81: 1192- 1205
doi: 10.1016/j.rser.2017.04.095
[4]
ILBEIGI M, GHOMEISHI M, DEHGHANBANADAKI A Prediction and optimization of energy consumption in an office building using artificial neural network and a genetic algorithm[J]. Sustainable Cities and Society, 2020, 61: 102325
doi: 10.1016/j.scs.2020.102325
[5]
MOHANDES S R, ZHANG X, MAHDIYAR A A comprehensive review on the application of artificial neural networks in building energy analysis[J]. Neurocomputing, 2019, 340: 55- 75
doi: 10.1016/j.neucom.2019.02.040
[6]
BOROWSKI M, ZWOLIŃSKA K Prediction of cooling energy consumption in hotel building using machine learning techniques[J]. Energies, 2020, 13 (23): 1- 19
[7]
JOVANOVIĆ R Ž, SRETENOVIĆ A A, ŽIVKOVIĆ B D Ensemble of various neural networks for prediction of heating energy consumption[J]. Energy and Buildings, 2015, 94: 189- 199
doi: 10.1016/j.enbuild.2015.02.052
[8]
WANG Z, HONG T, PIETTE M A Data fusion in predicting internal heat gains for office buildings through a deep learning approach[J]. Applied Energy, 2019, 240: 386- 398
doi: 10.1016/j.apenergy.2019.02.066
[9]
邵光成, 章坤, 王志宇, 等 基于IABC-RBF神经网络的地下水埋深预测模型[J]. 浙江大学学报: 工学版, 2019, 53 (7): 1323- 1330 SHAO Guang-cheng, ZHANG Kun, WANG Zhi-yu, et al Groundwater depth prediction model based on IABC-RBF neural network[J]. Journal of Zhejiang University: Engineering Science, 2019, 53 (7): 1323- 1330
[10]
LI K, SU H, CHU J Forecasting building energy consumption using neural networks and hybrid neuro-fuzzy system: a comparative study[J]. Energy and Buildings, 2011, 43 (10): 2893- 2899
doi: 10.1016/j.enbuild.2011.07.010
[11]
ZHOU G, MOAYEDI H, BAHIRAEI M, et al Employing artificial bee colony and particle swarm techniques for optimizing a neural network in prediction of heating and cooling loads of residential buildings[J]. Journal of Cleaner Production, 2020, 254: 120082
doi: 10.1016/j.jclepro.2020.120082
[12]
HUANG Y, LI C Accurate heating, ventilation and air conditioning system load prediction for residential buildings using improved ant colony optimization and wavelet neural network[J]. Journal of Building Engineering, 2021, 35: 101972
doi: 10.1016/j.jobe.2020.101972
[13]
GAO X, QI C, XUE G, et al Forecasting the heat load of residential buildings with heat metering based on CEEMDAN-SVR[J]. Energies, 2020, 13 (22): 6079
doi: 10.3390/en13226079
[14]
DING Y, ZHANG Q, YUAN T, et al Model input selection for building heating load prediction: a case study for an office building in Tianjin[J]. Energy and Buildings, 2018, 159: 254- 270
doi: 10.1016/j.enbuild.2017.11.002
[15]
LIU M D, DING L, BAI Y L Application of hybrid model based on empirical mode decomposition, novel recurrent neural networks and the ARIMA to wind speed prediction[J]. Energy Conversion and Management, 2021, 233: 113917
doi: 10.1016/j.enconman.2021.113917
[16]
KIRAN M S TSA: tree-seed algorithm for continuous optimization[J]. Expert Systems with Applications, 2015, 42 (19): 6686- 6698
doi: 10.1016/j.eswa.2015.04.055
[17]
XI Z, XU A, KOU Y, et al Target maneuver trajectory prediction based on RBF neural network optimized by hybrid algorithm[J]. Journal of Systems Engineering and Electronics, 2021, 32 (2): 498- 516
doi: 10.23919/JSEE.2021.000042
[18]
CHEN G, TANG B, ZENG X, et al Short-term wind speed forecasting based on long short-term memory and improved BP neural network[J]. International Journal of Electrical Power and Energy Systems, 2022, 134: 107365
doi: 10.1016/j.ijepes.2021.107365
[19]
TORRES M E, COLOMINAS M A, SCHLOTTHAUER G, et al. A complete ensemble empirical mode decomposition with adaptive noise [C]// 2011IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). Prague: IEEE, 2011: 4144-4147.
[20]
TIBSHIRANI R Regression shrinkage and selection via the lasso[J]. Journal of the Royal Statistical Society: Series B (Statistical Methodological), 1996, 58 (1): 267- 288
doi: 10.1111/j.2517-6161.1996.tb02080.x
[21]
HUANG N E, SHEN Z, LONG S R, et al The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis[J]. Proceedings of the Royal Society Series A: Mathematical, Physical and Engineering Sciences, 1998, 454 (1971): 903- 995
doi: 10.1098/rspa.1998.0193
[22]
SUN H, ZHAI W, WANG Y, et al Privileged information-driven random network based non-iterative integration model for building energy consumption prediction[J]. Applied Soft Computing, 2021, 108: 107438
doi: 10.1016/j.asoc.2021.107438
[23]
RICHMAN J S, MOORMAN J R Physiological time-series analysis using approximate entropy and sample entropy[J]. American Journal of Physiology: Heart and Circulatory Physiology, 2000, 278 (6): H2039- H2049
doi: 10.1152/ajpheart.2000.278.6.H2039
[24]
施莹, 林建辉, 庄哲, 等 基于振动信号时频分解-样本熵的受电弓裂纹故障诊断[J]. 振动与冲击, 2019, 38 (8): 180- 187 SHI Ying, LIN Jian-hui, ZHUANG Zhe, et al Fault diagnosis for pantograph cracks based on time-frequency decomposition and sample entropy of vibration signals[J]. Journal of Vibration and Shock, 2019, 38 (8): 180- 187
[25]
WU Q, LIN H Daily urban air quality index forecasting based on variational mode decomposition, sample entropy and LSTM neural network[J]. Sustainable Cities and Society, 2019, 50: 101657
doi: 10.1016/j.scs.2019.101657
[26]
SUN Y, HAGHIGHAT F, FUNG B C M A review of the-state-of-the-art in data-driven approaches for building energy prediction[J]. Energy and Buildings, 2020, 221: 110022
doi: 10.1016/j.enbuild.2020.110022
[27]
GAO Z, YU J, ZHAO A, et al A hybrid method of cooling load forecasting for large commercial building based on extreme learning machine[J]. Energy, 2022, 238: 122073
doi: 10.1016/j.energy.2021.122073
[28]
韩波, 李衡, 王志波, 等 气溶胶光学厚度估测中的LASSO特征选择方法[J]. 武汉大学学报:信息科学版, 2018, 43 (4): 536- 541 HAN Bo, LI Heng, WANG Zhi-bo, et al A feature selection approach via LASSO for aerosol optical thickness estimation[J]. Geomatics and Information Science of Wuhan University, 2018, 43 (4): 536- 541
[29]
李鱼强, 潘天红, 李浩然, 等 近红外光谱LASSO特征选择方法及其聚类分析应用研究[J]. 光谱学与光谱分析, 2019, 39 (12): 3809- 3815 LI Yu-qiang, PAN Tian-hong, LI Hao-ran, et al NIR spectral feature selection using LASSO method and its application in the classification analysis[J]. Spectroscopy and Spectral Analysis, 2019, 39 (12): 3809- 3815
[30]
EFRON B, HASTIE T, JOHNSTONE I, et al. Least angle regression [EB/OL]. [2021-10-27]. https://hastie.su.domains/Papers/LARS/LeastAngle_2002.pdf.
[31]
甘敏, 彭晓燕, 彭辉 RBF神经网络参数估计的两种混合优化算法[J]. 控制与决策, 2009, 24 (8): 1172- 1176 GAN Min, PENG Xiao-yan, PENG Hui Two hybrid parameter optimization algorithms for RBF neural networks[J]. Control and Decision, 2009, 24 (8): 1172- 1176
doi: 10.3321/j.issn:1001-0920.2009.08.010
[32]
YANG Z, MOURSHED M, LIU K, et al A novel competitive swarm optimized RBF neural network model for short-term solar power generation forecasting[J]. Neurocomputing, 2020, 397: 415- 421
doi: 10.1016/j.neucom.2019.09.110
[33]
KOSE E Optimal control of AVR system with tree seed algorithm-based PID controller[J]. IEEE Access, 2020, 8: 89457- 89467
doi: 10.1109/ACCESS.2020.2993628
[34]
DUAN J, WANG P, MA W, et al A novel hybrid model based on nonlinear weighted combination for short-term wind power forecasting[J]. International Journal of Electrical Power and Energy Systems, 2022, 134: 107452
doi: 10.1016/j.ijepes.2021.107452
[35]
KINGMA D P, BA J L. Adam: a method for stochastic optimization [EB/OL]. [2021-10-27]. https://arxiv.org/pdf/1412.6980.pdf.
[36]
BALAJI E, BRINDHA D, ELUMALAI V K, et al Automatic and non-invasive Parkinson's disease diagnosis and severity rating using LSTM network[J]. Applied Soft Computing, 2021, 108: 107463
doi: 10.1016/j.asoc.2021.107463
[37]
CHANG Z, ZHANG Y, CHEN W Electricity price prediction based on hybrid model of adam optimized LSTM neural network and wavelet transform[J]. Energy, 2019, 187: 115804
doi: 10.1016/j.energy.2019.07.134
[38]
尹爱军, 赵磊, 吴宏钢 相关法动平衡校正中的3σ准则误差处理方法[J]. 重庆大学学报, 2013, 36 (10): 22- 26 YIN Ai-jun, ZHAO Lei, WU Hong-gang Error process based on 3σ rule used in balancing by correlation theory[J]. Journal of Chongqing University, 2013, 36 (10): 22- 26
[39]
胡沛然, 陈少辉 权重归一化拉格朗日插值及其空间降尺度应用[J]. 遥感信息, 2019, 34 (6): 63- 71 HU Pei-ran, CHEN Shao-hui Weight normalization based Lagrange interpolation and its application in downscaling[J]. Remote Sensing Information, 2019, 34 (6): 63- 71
doi: 10.3969/j.issn.1000-3177.2019.06.011
