Skip to main content
SpringerLink
Log in

An improved transfer learning strategy for short-term cross-building energy prediction using data incremental

  • Research Article
  • Advances in Modeling and Simulation Tools
  • Published:
Building Simulation Aims and scope Submit manuscript

Abstract

The available modelling data shortage issue makes it difficult to guarantee the performance of data-driven building energy prediction (BEP) models for both the newly built buildings and existing information-poor buildings. Both knowledge transfer learning (KTL) and data incremental learning (DIL) can address the data shortage issue of such buildings. For new building scenarios with continuous data accumulation, the performance of BEP models has not been fully investigated considering the data accumulation dynamics. DIL, which can learn dynamic features from accumulated data adapting to the developing trend of new building time-series data and extend BEP model’s knowledge, has been rarely studied. Previous studies have shown that the performance of KTL models trained with fixed data can be further improved in scenarios with dynamically changing data. Hence, this study proposes an improved transfer learning cross-BEP strategy continuously updated using the coarse data incremental (CDI) manner. The hybrid KTL-DIL strategy (LSTM-DANN-CDI) uses domain adversarial neural network (DANN) for KLT and long short-term memory (LSTM) as the Baseline BEP model. Performance evaluation is conducted to systematically qualify the effectiveness and applicability of KTL and improved KTL-DIL. Real-world data from six-type 36 buildings of six types are adopted to evaluate the performance of KTL and KTL-DIL in data-driven BEP tasks considering factors like the model increment time interval, the available target and source building data volumes. Compared with LSTM, results indicate that KTL (LSTM-DANN) and the proposed KTL-DIL (LSTM-DANN-CDI) can significantly improve the BEP performance for new buildings with limited data. Compared with the pure KTL strategy LSTM-DANN, the improved KTL-DIL strategy LSTM-DANN-CDI has better prediction performance with an average performance improvement ratio of 60%.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Similar content being viewed by others

Abbreviations

b :

bias matrix

c t :

storage cell

f t :

forget gate

G λ :

gradient reversal layer

h t :

hidden layer

i t :

input gate

I :

unit matrix

k :

number of iterations

l i :

domain labels for true values

\(\overline {{l_i}} \) :

domain labels for predicted values

L d :

domain classification loss value

L(D t ) :

loss function of fine-tune

L ft :

weight parameter of the change layer

L MMD :

MMD loss

L y :

regression prediction loss value

m :

total number of iterations

N :

amount of training data

o t :

output gate

P :

data distributions

r :

length of the minimum total batch in the training data

R :

function symbol

G f :

feature extractor

G y :

regression predictor

G d :

domain classifier

tanh:

activation function tanh

x :

input data

y i :

true values of energy consumption

\(\overline {{y_i}} \) :

predicted values of energy consumption

u :

weight matrix

w :

weight matrix

α :

hyperparameter

θ f :

network connection weights of feature extractor

θ y :

network connection weights of regression predictor

θ d :

network connection weights of domain classifier

i :

number of data

S:

source domain

t :

moment of input data

T:

target domain

BAS:

building automation system

BED:

building energy data

BEP:

building energy prediction

BES:

building energy system

CDI:

coarse incremental learning

CV_RMSE:

coefficient of variation of root-mean squared error

DANN:

domain adversarial neural network

DIL:

data incremental learning

DTL:

deep transfer learning

LSTM:

long short-term memory

MAPE:

mean absolute percentage error

KTL:

knowledge transfer learning

RMSE:

root mean square error

PIR:

performance improvement ratio

References

  • Ahmad MW, Mourshed M, Rezgui Y (2017). Trees vs Neurons: Comparison between random forest and ANN for high-resolution prediction of building energy consumption. Energy and Buildings, 147: 77–89.

    Google Scholar 

  • Ahmad T, Zhang H, Yan B (2020). A review on renewable energy and electricity requirement forecasting models for smart grid and buildings. Sustainable Cities and Society, 55: 102052.

    Google Scholar 

  • Chen Y, Xu P, Chu Y, et al. (2017). Short-term electrical load forecasting using the Support Vector Regression (SVR) model to calculate the demand response baseline for office buildings. Applied Energy, 195: 659–670.

    Google Scholar 

  • Fan C, Xiao F, Zhao Y (2017). A short-term building cooling load prediction method using deep learning algorithms. Applied Energy, 195: 222–233.

    Google Scholar 

  • Fan C, Sun Y, Xiao F, et al. (2020). Statistical investigations of transfer learning-based methodology for short-term building energy predictions. Applied Energy, 262: 114499.

    Google Scholar 

  • Fang X, Gong G, Li G, et al. (2021a). A general multi-source ensemble transfer learning framework integrate of LSTM-DANN and similarity metric for building energy prediction. Energy and Buildings, 252: 111435.

    Google Scholar 

  • Fang X, Gong G, Li G, et al. (2021b). A hybrid deep transfer learning strategy for short term cross-building energy prediction. Energy, 215: 119208.

    Google Scholar 

  • Fang X, Gong G, Li G, et al. (2022). Deep reinforcement learning optimal control strategy for temperature setpoint real-time reset in multi-zone building HVAC system. Applied Thermal Engineering, 212: 118552.

    Google Scholar 

  • Feng R, Grana D, Balling N (2021). Imputation of missing well log data by random forest and its uncertainty analysis. Computers & Geosciences, 152: 104763.

    Google Scholar 

  • Foucquier A, Robert S, Suard F, et al. (2013). State of the art in building modelling and energy performances prediction: A review. Renewable and Sustainable Energy Reviews, 23: 272–288.

    Google Scholar 

  • Fu G (2018). Deep belief network based ensemble approach for cooling load forecasting of air-conditioning system. Energy, 148: 269–282.

    Google Scholar 

  • Ganin Y, Ustinova E, Ajakan H, et al. (2017). Domain-adversarial training of neural networks. In: Csurka G (Ed), Domain Adaptation in Computer Vision Applications. Cham, Switzerland: Springer.

    Google Scholar 

  • Gao Y, Ruan Y, Fang C, et al. (2020). Deep learning and transfer learning models of energy consumption forecasting for a building with poor information data. Energy and Buildings, 223: 110156.

    Google Scholar 

  • García-Laencina PJ, Sancho-Gómez J-L, Figueiras-Vidal AR (2010). Pattern classification with missing data: A review. Neural Computing and Applications, 19: 263–282.

    Google Scholar 

  • González-Vidal A, Mendoza-Bernal J, Niu S, et al. (2023). A transfer learning framework for predictive energy-related scenarios in smart buildings. IEEE Transactions on Industry Applications, 59: 26–37.

    Google Scholar 

  • Jin Y, Acquah MA, Seo M, Han S (2022). Short-term electric load prediction using transfer learning with interval estimate adjustment. Energy and Buildings, 258: 111846.

    Google Scholar 

  • Li A, Xiao F, Fan C, et al. (2021). Development of an ANN-based building energy model for information-poor buildings using transfer learning. Building Simulation, 14: 89–101.

    Google Scholar 

  • Li G, Zhao X, Fan C, et al. (2021). Assessment of long short-term memory and its modifications for enhanced short-term building energy predictions. Journal of Building Engineering, 43: 103182.

    Google Scholar 

  • Li A, Fan C, Xiao F, et al. (2022a). Distance measures in building informatics: An in-depth assessment through typical tasks in building energy management. Energy and Buildings, 258: 111817.

    Google Scholar 

  • Li G, Li F, Ahmad T, et al. (2022b). Performance evaluation of sequence-to-sequence-Attention model for short-term multi-step ahead building energy predictions. Energy, 259: 124915.

    Google Scholar 

  • Li G, Li F, Xu C, et al. (2022c). A spatial-temporal layer-wise relevance propagation method for improving interpretability and prediction accuracy of LSTM building energy prediction. Energy and Buildings, 271: 112317.

    Google Scholar 

  • Li G, Wu Y, Liu J, et al. (2022d). Performance evaluation of short-term cross-building energy predictions using deep transfer learning strategies. Energy and Buildings, 275: 112461.

    Google Scholar 

  • Li J, Zhang C, Zhao Y, et al. (2022e). Federated learning-based short-term building energy consumption prediction method for solving the data silos problem. Building Simulation, 15: 1145–1159.

    Google Scholar 

  • Li G, Chen L, Liu J, et al. (2023a). Comparative study on deep transfer learning strategies for cross-system and cross-operation-condition building energy systems fault diagnosis. Energy, 263: 125943.

    Google Scholar 

  • Li W, Tang R, Wang S (2023b). A fully distributed robust optimal control approach for air-conditioning systems considering uncertainties of communication link in IoT-enabled building automation systems. Energy and Built Environment

  • Lissa P, Schukat M, Keane M, et al. (2021). Transfer learning applied to DRL-Based heat pump control to leverage microgrid energy efficiency. Smart Energy, 3: 100044.

    Google Scholar 

  • Liu J, Zhang Q, Li X, et al. (2021). Transfer learning-based strategies for fault diagnosis in building energy systems. Energy and Buildings, 250: 111256.

    Google Scholar 

  • Low R, Tekler ZD, Cheah L (2020). Predicting commercial vehicle parking duration using generative adversarial multiple imputation networks. Transportation Research Record: Journal of the Transportation Research Board, 2674: 820–831.

    Google Scholar 

  • Ma J, Cheng JCP, Jiang F, et al. (2020). A bi-directional missing data imputation scheme based on LSTM and transfer learning for building energy data. Energy and Buildings, 216: 109941.

    Google Scholar 

  • Masana M, Liu X, Twardowski B, et al. (2023). Class-incremental learning: Survey and performance evaluation on image classification. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45: 5513–5533.

    Google Scholar 

  • Mat Daut MA, Hassan MY, Abdullah H, et al. (2017). Building electrical energy consumption forecasting analysis using conventional and artificial intelligence methods: A review. Renewable and Sustainable Energy Reviews, 70: 1108–1118.

    Google Scholar 

  • Miller C, Kathirgamanathan A, Picchetti B, et al. (2020). The Building Data Genome Project 2, energy meter data from the ASHRAE Great Energy Predictor III competition. Scientific Data, 7: 368.

    Google Scholar 

  • Miller C, Picchetti B, Fu C, et al. (2022). Limitations of machine learning for building energy prediction: ASHRAE Great Energy Predictor III Kaggle competition error analysis. Science and Technology for the Built Environment, 28: 610–627.

    Google Scholar 

  • Ng RW, Begam KM, Rajkumar RK, et al. (2021). An improved self-organizing incremental neural network model for short-term time-series load prediction. Applied Energy, 292: 116912.

    Google Scholar 

  • O’Neill Z, Eisenhower B (2013). Leveraging the analysis of parametric uncertainty for building energy model calibration. Building Simulation, 6: 365–377.

    Google Scholar 

  • Parisi GI, Kemker R, Part JL, et al. (2019). Continual lifelong learning with neural networks: A review. Neural Networks, 113: 54–71.

    Google Scholar 

  • Park H, Park DY, Noh B, et al. (2022). Stacking deep transfer learning for short-term cross building energy prediction with different seasonality and occupant schedule. Building and Environment, 218: 109060.

    Google Scholar 

  • Peirelinck T, Kazmi H, Mbuwir BV, et al. (2022). Transfer learning in demand response: A review of algorithms for data-efficient modelling and control. Energy and AI, 7: 100126.

    Google Scholar 

  • Pinto G, Messina R, Li H, et al. (2022a). Sharing is caring: An extensive analysis of parameter-based transfer learning for the prediction of building thermal dynamics. Energy and Buildings, 276: 112530.

    Google Scholar 

  • Pinto G, Wang Z, Roy A, et al. (2022b). Transfer learning for smart buildings: A critical review of algorithms, applications, and future perspectives. Advances in Applied Energy, 5: 100084.

    Google Scholar 

  • Sayed AN, Himeur Y, Bensaali F (2022). Deep and transfer learning for building occupancy detection: A review and comparative analysis. Engineering Applications of Artificial Intelligence, 115: 105254.

    Google Scholar 

  • Sendra-Arranz R, Gutiérrez A (2020). A long short-term memory artificial neural network to predict daily HVAC consumption in buildings. Energy and Buildings, 216: 109952.

    Google Scholar 

  • Singh MM, Singaravel S, Geyer P (2021). Machine learning for early stage building energy prediction: Increment and enrichment. Applied Energy, 304: 117787.

    Google Scholar 

  • Skillington K, Crawford RH, Warren-Myers G, et al. (2022). A review of existing policy for reducing embodied energy and greenhouse gas emissions of buildings. Energy Policy, 168: 112920.

    Google Scholar 

  • Somu N, M R GR, Ramamritham K (2020). A hybrid model for building energy consumption forecasting using long short term memory networks. Applied Energy, 261: 114131.

    Google Scholar 

  • Stekhoven DJ, Bühlmann P (2012). MissForest—non-parametric missing value imputation for mixed-type data. Bioinformatics, 28: 112–118.

    Google Scholar 

  • Sun Y, Haghighat F, Fung BCM (2020). A review of the-state-of-the-art in data-driven approaches for building energy prediction. Energy and Buildings, 221: 110022.

    Google Scholar 

  • Sun Y, Haghighat F, Fung BCM (2022). Trade-off between accuracy and fairness of data-driven building and indoor environment models: A comparative study of pre-processing methods. Energy, 239: 122273.

    Google Scholar 

  • Sun Y, Li J, Xu Y, et al. (2023). Deep learning versus conventional methods for missing data imputation: A review and comparative study. Expert Systems with Applications, 227: 120201.

    Google Scholar 

  • van Buuren S, Oudshoorn CGM (2000). Multivariate Imputation by Chained Equations: Mice V1.0 User’s manual. Netherlands Organization for Applied Scientific Research (TNO).

  • Vincent P, Larochelle H, Bengio Y, et al. (2008). Extracting and composing robust features with denoising autoencoders. In: Proceedings of the 25th International Conference on Machine learning, Helsinki, Finland.

  • Wang Z, Srinivasan RS (2017). A review of artificial intelligence based building energy use prediction: Contrasting the capabilities of single and ensemble prediction models. Renewable and Sustainable Energy Reviews, 75: 796–808.

    Google Scholar 

  • Wang Z, Wang Y, Zeng R, et al. (2018). Random Forest based hourly building energy prediction. Energy and Buildings, 171: 11–25.

    Google Scholar 

  • Wang W, Hong T, Xu X, et al. (2019). Forecasting district-scale energy dynamics through integrating building network and long short-term memory learning algorithm. Applied Energy, 248: 217–230.

    Google Scholar 

  • Wang Z, Liu J, Zhang Y, et al. (2021). Practical issues in implementing machine-learning models for building energy efficiency: Moving beyond obstacles. Renewable and Sustainable Energy Reviews, 143: 110929.

    Google Scholar 

  • Wang C, Yuan J, Huang K, et al. (2022a). Research on thermal load prediction of district heating station based on transfer learning. Energy, 239: 122309.

    Google Scholar 

  • Wang Z, Xia L, Yuan H, et al. (2022b). Principles, research status, and prospects of feature engineering for data-driven building energy prediction: A comprehensive review. Journal of Building Engineering, 58: 105028.

    Google Scholar 

  • Xiao T, Xu P, He R, et al. (2022a). Status quo and opportunities for building energy prediction in limited data Context—Overview from a competition. Applied Energy, 305: 117829.

    Google Scholar 

  • Xiao Z, Gang W, Yuan J, et al. (2022b). Impacts of data preprocessing and selection on energy consumption prediction model of HVAC systems based on deep learning. Energy and Buildings, 258: 111832.

    Google Scholar 

  • Yoon J, Jordon J, van der Schaar M (2018). GAIN: Missing data imputation using generative adversarial nets. arXiv: 1806.02920.

  • Zhang L, Wen J (2021). Active learning strategy for high fidelity short-term data-driven building energy forecasting. Energy and Buildings, 244: 111026.

    Google Scholar 

  • Zhang X, Sun Y, Gao D, et al. (2022). Similarity-based grouping method for evaluation and optimization of dataset structure in machine-learning based short-term building cooling load prediction without measurable occupancy information. Applied Energy, 327: 120144.

    Google Scholar 

  • Zhao Y, Li T, Zhang X, et al. (2019). Artificial intelligence-based fault detection and diagnosis methods for building energy systems: Advantages, challenges and the future. Renewable and Sustainable Energy Reviews, 109: 85–101.

    Google Scholar 

  • Zhuang F, Qi Z, Duan K, et al. (2021). A comprehensive survey on transfer learning. Proceedings of the IEEE, 109: 43–76.

    Google Scholar 

Download references

Acknowledgements

This work was jointly supported by the Opening Fund of Key Laboratory of Low-grade Energy Utilization Technologies and Systems of Ministry of Education of China (Chongqing University) (LLEUTS-202305), the Opening Fund of State Key Laboratory of Green Building in Western China (LSKF202316), the open Foundation of Anhui Province Key Laboratory of Intelligent Building and Building Energy-saving (IBES2022KF11), “The 14th Five-Year Plan” Hubei Provincial advantaged characteristic disciplines (groups) project of Wuhan University of Science and Technology (2023D0504, 2023D0501), the National Natural Science Foundation of China (51906181), the 2021 Construction Technology Plan Project of Hubei Province (2021-83), and the Science and Technology Project of Guizhou Province: Integrated Support of Guizhou [2023] General 393.

Author information

Authors and Affiliations

  1. School of Urban Construction, Wuhan University of Science and Technology, Wuhan, 430065, China

    Guannan Li, Yubei Wu, Tao Li, Jiajia Gao, Chengliang Xu & Zixi Wang

  2. Anhui Province Key Laboratory of Intelligent Building and Building Energy-saving, Anhui Jianzhu University, Hefei, 230601, China

    Guannan Li

  3. State Key Laboratory of Green Building in Western China, Xi’an University of Architecture and Technology, Xi’an, 710055, China

    Guannan Li & Tao Li

  4. College of Urban Construction, Nanjing Tech University, No. 200, North Zhongshan Road, Nanjing, 210009, China

    Chengchu Yan

  5. Key Laboratory of Low-grade Energy Utilization Technologies and Systems (Chongqing University), Ministry of Education of China, Chongqing University, Chongqing, 400044, China

    Guannan Li

  6. College of Civil Engineering, Hunan University, Changsha, 410082, China

    Xi Fang

  7. Hubei Provincial Engineering Research Center of Urban Regeneration, Wuhan University of Science and Technology, Wuhan, 430065, China

    Guannan Li, Tao Li, Jiajia Gao & Chengliang Xu

Authors
  1. Guannan Li
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Yubei Wu
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Chengchu Yan
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Xi Fang
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Tao Li
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Jiajia Gao
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Chengliang Xu
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Zixi Wang
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

All authors contributed to the study conception and design. Original draft, review, editing and funding acquisition were performed by Guannan Li. Review, editing and project administration were performed by Chengchu Yan. Data preparation and software were performed by Yubei Wu and Zixi Wang. Review and editing were performed by Xi Fang, Tao Li, Jiajia Gao and Chengliang Xu. The first draft of the manuscript was written by Yubei Wu and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Chengchu Yan.

Ethics declarations

We declare that we do not have any commercial or associative interest that represents a conflict of interest in connection with the work submitted

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Li, G., Wu, Y., Yan, C. et al. An improved transfer learning strategy for short-term cross-building energy prediction using data incremental. Build. Simul. 17, 165–183 (2024). https://doi.org/10.1007/s12273-023-1053-x

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s12273-023-1053-x

Keywords

Search

Navigation