Data fusion in predicting internal heat gains for office buildings through a deep learning approach

Wang, Zhe; Hong, Tianzhen; Piette, Mary Ann

doi:10.1016/j.apenergy.2019.02.066

Cited by 85 publications

(38 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The gray-box models lack ways to identify and reflect the schedule and intensity change of occupant, lighting, and plug loads. As indicated in [17], the internal heat gains account for an increasingly higher proportion in modern buildings with a high-efficiency level of envelope.…”

Section: Previous Workmentioning

confidence: 99%

See 1 more Smart Citation

Building thermal load prediction through shallow machine learning and deep learning

2020

Self Cite

View full text Add to dashboard Cite

Building thermal load prediction informs the optimization of cooling plant and thermal energy storage. Physics-based prediction models of building thermal load are constrained by the model and input complexity. In this study, we developed 12 data-driven models (7 shallow learning, 2 deep learning, and 3 heuristic methods) to predict building thermal load and compared shallow machine learning and deep learning. The 12 prediction models were compared with the measured cooling demand. It was found XGBoost (Extreme Gradient Boost) and LSTM (Long Short Term Memory) provided the most accurate load prediction in the shallow and deep learning category, and both outperformed the best baseline model, which uses the previous day's data for prediction. Then, we discussed how the prediction horizon and input uncertainty would influence the load prediction accuracy. Major conclusions are twofold: first, LSTM performs well in shortterm prediction (1h ahead) but not in long term prediction (24h ahead), because the sequential information becomes less relevant and accordingly not so useful when the prediction horizon is long. Second, the presence of weather forecast uncertainty deteriorates XGBoost's accuracy and favors LSTM, because the sequential information makes the model more robust to input uncertainty. Training the model with the uncertain rather than accurate weather data could enhance the model's robustness. Our findings have two implications for practice. First, LSTM is recommended for short-term load prediction given that weather forecast uncertainty is unavoidable. Second, XGBoost is recommended for long term prediction, and the model should be trained with the presence of input uncertainty.

show abstract

Section: Previous Workmentioning

confidence: 99%

“…The long short term memory (LSTM) is a special form of RNN, which is designed to handle long sequential data. LSTM has been used successfully to forecast internal heat gains [17] or building energy usage [26]. However, to the best of the authors' knowledge, LSTM has not been used for building thermal load prediction.…”

Section: Previous Workmentioning

confidence: 99%

Building thermal load prediction through shallow machine learning and deep learning

2020

Self Cite

View full text Add to dashboard Cite

show abstract

“…Machine learning could be helpful to address both problems. First, machine learning could be used to predict weather, occupancy [93] and building load [94], and then take the predictive information into optimization. Second, machine learning could enable the controller to learn from the building operation data, identifying states, updating parameters, and adapting itself to any changes in the target building.…”

Section: Machine Learning For Building Controlmentioning

confidence: 99%

State-of-the-art on research and applications of machine learning in the building life cycle

Hong

Wang

Luo

et al. 2020

Energy and Buildings

Self Cite

248

View full text Add to dashboard Cite

Fueled by big data, powerful and affordable computing resources, and advanced algorithms, machine learning has been explored and applied to buildings research for the past decades and has demonstrated its potential to enhance building performance. This study systematically surveyed how machine learning has been applied at different stages of building life cycle. By conducting a literature search on the Web of Knowledge platform, we found 9579 papers in this field and selected 153 papers for an in-depth review. The number of published papers is increasing year by year, with a focus on building design, operation, and control. However, no study was found using machine learning in building commissioning. There are successful pilot studies on fault detection and diagnosis of HVAC equipment and systems, load prediction, energy baseline estimate, load shape clustering, occupancy prediction, and learning occupant behaviors and energy use patterns. None of the existing studies were adopted broadly by the building industry, due to common challenges including (1) lack of large scale labeled data to train and validate the model, (2) lack of model transferability, which limits a model trained with one data-rich building to be used in another building with limited data, (3) lack of strong justification of costs and benefits of deploying machine learning, and (4) the performance might not be reliable and robust for the stated goals, as the method might work for some buildings but could not be generalized to others. Findings from the study can inform future machine learning research to improve occupant comfort, energy efficiency, demand flexibility, and resilience of buildings, as well as to inspire young researchers in the field to explore multidisciplinary approaches that integrate building science, computing science, data science, and social science.

show abstract

“…humidity, CO2-to sense occupancy information, this study, therefore, takes into consideration of those parameters. The dataset feature pool can be roughly expressed as: adequately high prediction accuracy could be achieved with as few inputs as possible [48]. In this study, two steps are conducted.…”

Section: Features From Wi-fi Connectionsmentioning

confidence: 99%

Cross-source sensing data fusion for building occupancy prediction with adaptive lasso feature filtering

Wang

Hong

et al. 2019

Building and Environment

Self Cite

View full text Add to dashboard Cite

Fusing various sensing data sources is able to improve the accuracy and reliability of building occupancy detection. Efficiently fusing environmental sensors and wireless network signals is seldom studied for its computational and technical challenges. This study aims to propose an integrated model that is able to extract critical data features for environmental and Wi-Fi probe dual sensing sources to promote computational efficiency. The adaptive lasso model was introduced for the feature extraction and reduction process. To validate the proposed model, an onsite experiment was conducted and two occupancy resolutions, real-time and four-level occupancy resolutions, were compared. The results suggested that eight features among all twelve features are most relevant. The mean absolute error of the selected data features is about 2.18 for real-time occupancy and F1_accuracy is about 84.36% for four-level occupancy.

show abstract

Data fusion in predicting internal heat gains for office buildings through a deep learning approach

Cited by 85 publications

References 40 publications

Building thermal load prediction through shallow machine learning and deep learning

Building thermal load prediction through shallow machine learning and deep learning

State-of-the-art on research and applications of machine learning in the building life cycle

Cross-source sensing data fusion for building occupancy prediction with adaptive lasso feature filtering

Contact Info

Product

Resources

About