Building Energy Prediction Research Articles

A multi-source transfer learning model based on LSTM and domain adaptation for building energy prediction

Transfer learning can use the knowledge learned from the operating data of other buildings to facilitate the energy prediction of a target building. However, most of the current research focuses on the transfer of a single source building of the same type or from the same region. A single source domain produces domain shift due to the difficulty of aligning the distribution with the target domain. To address this problem, this paper proposes a novel multi-source transfer learning energy prediction model based on long short-term memory (LSTM) and multi-kernel maximum mean difference (MK-MMD) domain adaptation. This model was used for the short-term energy prediction of different types of buildings lacking historical data. In addition, dynamic time warping (DTW) was used to select the source domain. Multiple multi-source models and corresponding single-source models were compared on a collection of buildings in the Higashida area of Fukuoka Prefecture, Japan. On the experimental datasets, the results showed that DTW relatively accurately measured the similarity between building energy consumption datasets. Compared with the results of the single-source transfer learning models, the multi-source transfer learning models achieved better average prediction performance, and their mean absolute percentage error (MAPE) improved the prediction accuracy by 6.88–15.37%.

A Review of Data-Driven Building Energy Prediction

Building energy consumption prediction has a significant effect on energy control, design optimization, retrofit evaluation, energy price guidance, and prevention and control of COVID-19 in buildings, providing a guarantee for energy efficiency and carbon neutrality. This study reviews 116 research papers on data-driven building energy prediction from the perspective of data and machine learning algorithms and discusses feasible techniques for prediction across time scales, building levels, and energy consumption types in the context of the factors affecting data-driven building energy prediction. The review results revealed that the outdoor dry-bulb temperature is a vital factor affecting building energy consumption. In data-driven building energy consumption prediction, data preprocessing enables prediction across time scales, energy consumption feature extraction enables prediction across energy consumption types, and hyperparameter optimization enables prediction across time scales and building layers.

Exploring the effect of longwave radiation exchange on the energy balance of building façades in subtropical regions

The energy balance of the building façade affects the accuracy of building energy prediction. Although algorithms for energy balance have been applied in building energy simulation software, the longwave radiation exchange of building façades is often considered in a simple (implicit) way. This study proposes the concept of ambient equivalent temperature for quantifying the longwave radiation emitted by other buildings and trees in the surrounding environment. A more detailed radiation transfer model was developed to assess the energy balance of the longwave radiation of building façades. The input parameters of the model were simplified and the applicability of the model was improved by fitting the ambient equivalent temperature and meteorological parameters. Through quantitative and qualitative analysis, the effect of view factor parameters on the model output was assessed, and a universal value for the sky view factor (SVF) was determined to calculate the longwave radiation exchange of the building façade. The calculation accuracy of the longwave radiation was 71.4% higher in the simplified SVF model than in the original model. These findings provide a more accurate method for calculating the longwave radiation energy balance of building façades and have implications for building energy prediction and building energy efficiency design.

Comparison of empirical modal decomposition class techniques applied in noise cancellation for building heating consumption prediction based on time-frequency analysis

Empirical Modal Decomposition (EMD), and improved or modified techniques derived from EMD, collectively referred to as Empirical Modal Decomposition class (EMDC) techniques. EMDC techniques have a wide range of applications in building energy analysis, especially time–frequency analysis based noise cancellation in data-driven building energy prediction. However, there is a gap in the literature related to the choice of EMDC techniques in data-driven models. This paper provides a framework for a comprehensive comparison of EMD, Ensemble Empirical Mode Decomposition (EEMD), Variational Mode Decomposition (VMD) and Empirical Wavelet Transform (EWT) techniques for building heat consumption prediction modeling. A real building is used as an example to compare the noise cancellation potential of these techniques and the prediction accuracy under various data-driven models. The results demonstrated that noise cancellation using the EMDC techniques significantly improves Signal-Noise Ratio, regularity, and consistency with the original signal trend. The prediction models trained using the noise-cancelled data have the Root Mean Squared Error and the Mean Absolute Error reductions of 22.5 % and 31.3 % on average, respectively. Meanwhile, the predicted signals of the models inherit the noise cancellation benefits of the noise-cancelled training data.

A hybrid agent-based machine learning method for human-centred energy consumption prediction

Occupant behaviour has significant impacts on the performance of machine learning algorithms when predicting building energy consumption. Due to a variety of reasons (e.g., underperforming building energy management systems or restrictions due to privacy policies), the availability of occupational data has long been an obstacle that hinders the performance of machine learning algorithms in predicting building energy consumption. Therefore, this study proposed an agent-based machine learning model whereby agent-based modelling was employed to generate simulated occupational data as input features for machine learning algorithms for building energy consumption prediction. Boruta feature selection was also introduced in this study to select all relevant features. The results indicated that the performances of machine learning algorithms in predicting building energy consumption were significantly improved when using simulated occupational data, with even greater improvements after conducting Boruta feature selection.

Deep-Chain Echo State Network With Explainable Temporal Dependence for Complex Building Energy Prediction

Building energy prediction plays critical roles in the study of green building and smart city. The most challenging issue is to predict energy demand profiles over multiple time steps, which may have inconsistent timescales. Due to complex temporal dependence, existing prediction approaches cannot satisfy certain requirements in multistep (MS) or multitimescale (MTS) applications. In this article, a deep-chain echo state network (DCESN) is proposed to enhance the mapping capability for the MS demand prediction. The DCESN composed of many submodules of echo state network (ESN) belongs to the single-input multi-output (SIMO) model, and neuron states generated by sequential steps are utilized to prevent accumulative error in the recursive chain. Due to the novel learning mechanism of DCESN, numerical coefficients of temporal dependence are presented to explain the short-term and long-term impacts on future energy consumption. Experimental results in four cases indicate that the proposed DCESN has promising performance on MS and MTS prediction, and temporal dependence can be explained in a visible way. Comparative results of DCESN, sliding-window ESN, and long-short term memory (LSTM) demonstrate that the proposed learning mechanism could prevent error accumulation effectively.

Identifying the optimal heterogeneous ensemble learning model for building energy prediction using the exhaustive search method

Heterogeneous ensemble learning has received increasing attention in building energy prediction research because it provides more stable and accurate predictions than the conventional single model-based methods. Although existing studies have proved the feasibility of adopting heterogeneous ensemble learning in building energy prediction, they have not fully explored the method of constructing an optimal heterogeneous ensemble learning model based on limited base model resources. This paper attempts to identify the optimal heterogeneous ensemble learning model for building energy prediction by using the exhaustive search method. Six machine learning methods are used as the alternative base model for constructing heterogeneous ensemble learning models. All 57 heterogeneous ensemble learning models that can be built from the six base models are compared and the optimal model is identified. The hourly building energy prediction experiment is performed on an institutional building located at the University of Florida to verify the proposed exhaustive search method. Three modules representing different semesters of the test building are developed and tested to examine the predictive performance of the proposed heterogeneous ensemble learning models. The research findings demonstrate the feasibility of the proposed exhaustive search method in identifying the optimal heterogeneous ensemble learning model. The optimal heterogeneous ensemble learning model found in this study can reduce the MAPE by 4.9% − 6.6% compared with its most accurate base model in the model testing stage. The experimental results and research findings provide critical guidance for applying heterogeneous ensemble learning to data-driven building energy prediction.

A novel method of creating machine learning-based time series meta-models for building energy analysis

The meta-models have been widely used to replace computationally expensive engineering-based building energy models for model calibration, sensitivity analysis, and performance optimization of buildings. However, most meta-models are used for monthly or annual building energy analysis applying the same procedure as building energy or load prediction. A few studies thoroughly explore the characteristics of meta-models when creating hourly machine learning-based meta-models for building energy assessment, which is very different from conventional building load or energy prediction. Therefore, this paper proposes a new method of creating time series models for building energy analysis based on machine learning techniques. The main feature of this new method is to include two steps (model preselection and data folding) by taking advantage of the sequential nature during the process of constructing meta-models. A case study of office buildings is used to demonstrate the application of this new approach using the EnergyPlus program and R computational environment. The results indicate that the time-series meta-models constructed using this new method have very high accuracy with a moderate computational cost. For the best models, the R2 values are above 0.99 and the CV(RMSE) (coefficient variation of root mean square error) values are below 0.06 for both heating and cooling energy prediction. The ensemble machine learning models, including Cubist, gradient boosting machine, and stacking, are recommended to create the meta-models for hourly building energy analysis. The distributed lag models combined with these ensemble models can be used to utilize the time-series features of hourly building energy assessment. The method proposed here can be extended to the time-series energy analysis in different time scales (sub-hourly, hourly, or daily) for other energy systems (such as PV, solar thermal, and wind).

Application of Multilayer Extreme Learning Machine for Efficient Building Energy Prediction

Building energy efficiency is vital, due to the substantial amount of energy consumed in buildings and the associated adverse effects. A high-accuracy energy prediction model is considered as one of the most effective ways to understand building energy efficiency. In several studies, various machine learning models have been proposed for the prediction of building energy efficiency. However, the existing models are based on classical machine learning approaches and small datasets. Using a small dataset and inefficient models may lead to poor generalization. In addition, it is not common to see studies examining the suitability of machine learning methods for forecasting the energy consumption of buildings during the early design phase so that more energy-efficient buildings can be constructed. Hence, for these purposes, we propose a multilayer extreme learning machine (MLELM) for the prediction of annual building energy consumption. Our MLELM fuses stacks of autoencoders (AEs) with an extreme learning machine (ELM). We designed the autoencoder based on the ELM concept, and it is used for feature extraction. Moreover, the autoencoders were trained in a layer-wise manner, employed to extract efficient features from the input data, and the extreme learning machine model was trained using the least squares technique for a fast learning speed. In addition, the ELM was used for decision making. In this research, we used a large dataset of residential buildings to capture various building sizes. We compared the proposed MLELM with other machine learning models commonly used for predicting building energy consumption. From the results, we validated that the proposed MLELM outperformed other comparison methods commonly used in building energy consumption prediction. From several experiments in this study, the proposed MLELM was identified as the most efficient predictive model for energy use before construction, which can be used to make informed decisions about, manage, and optimize building design before construction.

Building Energy Models at Different Time Scales Based on Multi-Output Machine Learning

Machine learning techniques are widely applied in the field of building energy analysis to provide accurate energy models. The majority of previous studies, however, apply single-output machine learning algorithms to predict building energy use. Single-output models are unable to concurrently predict different time scales or various types of energy use. Therefore, this paper investigates the performance of multi-output energy models at three time scales (daily, monthly, and annual) using the Bayesian adaptive spline surface (BASS) and deep neural network (DNN) algorithms. The results indicate that the multi-output models based on the BASS approach combined with the principal component analysis can simultaneously predict accurate energy use at three time scales. The energy predictions also have the same or similar correlation structure as the energy data from the engineering-based EnergyPlus models. Moreover, the results from the multi-time scale BASS models have consistent accumulative features, which means energy use at a larger time scale equals the summation of energy use at a smaller time scale. The multi-output models at various time scales for building energy prediction developed in this research can be used in uncertainty analysis, sensitivity analysis, and calibration of building energy models.

Building interpretable predictive models with context-aware evolutionary learning

Building prediction models with the right balance between performance and interpretability is currently a great challenge in machine learning. A large number of recent studies have focused on either building intrinsically interpretable models or developing general explainers for blackbox models. Although these methods have been widely adopted, their interpretability or explanations are not always useful because of the lack of contexts considered in training machine learning models and producing explanations. This paper aims to tackle this significant challenge by developing a context-aware evolutionary learning algorithm (CELA) for building interpretable prediction models. A new context extraction method based on unsupervised self-structuring learning algorithms is developed to treat data in contexts. The proposed algorithm overcomes the limitations of existing evolutionary learning methods in handling a large number of features and large datasets by training specialised interpretable models based on the automatically extracted contexts. The new algorithm has been tested on complex regression datasets and a real-world building energy prediction task. The results suggest CELA can outperform well-known interpretable machine learning (IML) algorithms, the state-of-the-art evolutionary algorithm, and can produce predictions much closer to the results of blackbox algorithms such as XGBoost and artificial neural networks than the compared IML methods. Further analyses also demonstrate that the CELA’s prediction models are smaller and easier to interpret than those obtained by the evolutionary learning algorithm without context awareness.

Climatic change impacts on the energy requirements for the built environment sector

The population growth, rising living standards, and climate change impacts will considerably affect the energy needs for the built environment in the coming decades. Because of the climatic changes, the shift in humidity levels and dry/wet bulb temperatures would change the required energy to maintain the same spatial comfort conditions. Furthermore, the heating load may be lowered due to the higher winter temperatures; however, the cooling load will be significantly augmented for some regions. Thus, it is imperative to quantify the impacts of building heating and cooling loads concerning climate change. This will help emphasize increasing energy capacity, improving efficiency, and developing energy policies to reduce and adapt to the climate change implications. The present study focuses on analyzing the recent works addressing and exploring the effects of climate change on the built environment’s energy consumption. This paper examines and discusses the accepted approaches and projections of building energy change. The survey shows that the estimations of heating/cooling loads are significantly affected by the type of building, city, and other regional factors. Finally, the challenges associated with predicting building energy requirements are presented and discussed.

A New Explication of Minimum Variable Sets (MVS) for Building Energy Prediction Based on Building Performance Database

Building energy simulation plays a significant role in buildings, with applications such as building performance evaluation, retrofit decisions and the optimization of building operations. However, the wide range of model inputs has limited much research into empirically customized case studies due to the insufficient availability of data inputs or the lack of systematic feature selection of key inputs. To address this gap, this study proposes the concept of minimum variable sets (MVSs) for building energy-prediction models to improve the general applicability of building energy prediction using forward simulation. An MVS, in this paper, refers to a variable set that contains the most indispensable energy-related variables/features for annual building energy prediction. This study developed MVSs for office buildings by applying feature engineering algorithms to a Building Performance Database (BPD), which was established by integrating the design of experiments (DoE) method with high-dimensional data-space metrics, as well as parallel simulation. Supervised feature dimension reduction methods and multiple statistical criteria were adopted to choose different numbers of indispensable variables from the primary 16 building variables. The hierarchical MVSs that consist of the selected variables are characterized by the most influential features for building energy prediction, with certain requirements for prediction accuracy. To further improve the feasibility of MVSs, this study utilized two separate office buildings located in Shanghai and California as validation cases and provided comparable prediction accuracies across different sizes of MVS. The results showed that the MVS that has 12 variables has higher prediction accuracy than that which has 9 variables, followed by that which has 7 variables. Finally, the quantitatively hierarchical correlations between different sizes of MVS with different prediction accuracies for annual building energy could provide potential support for reasonable decision-making regarding building energy model variables, especially when comprehensive consideration is needed of the limited cost and data availability, and the acceptable accuracy of building energy.

Semi-supervised learning based framework for urban level building electricity consumption prediction

The spatial feature of building energy consumption in a city is essential for urban level energy planning and policy making. With the increasing availability of urban level building energy benchmarking datasets, machine learning has shown a powerful capability of making data-driven predictions on urban level building energy consumption. However, the building energy benchmarking datasets usually only cover large buildings, which are not sufficient representations of all buildings in a city. Besides building energy benchmarking datasets, many other urban level open datasets are also valuable to building energy prediction, but they do not contain building energy use data, in other words, they are unlabeled. This study proposes a novel framework based on semi-supervised learning to make effective use of the unlabeled datasets to develop more generic urban level data-driven building energy prediction models, and energy mapping with higher space resolution. The framework consists of preliminary labeling, selection of pseudo labeled samples and predictive modelling. Several machine learning algorithms are proposed and compared for generating pseudo labels of building electricity consumption for unlabeled datasets of small and medium-sized buildings. A selection process consisting of convergence testing and screening is designed to select pseudo labeled samples with high credibility to enlarge the labeled dataset. A novel two-level performance evaluation method is proposed to evaluate the performance of the framework at both urban level and district level to enhance the spatial resolution of the predictions. The framework is implemented to model and map the individual electricity consumptions of all buildings in two years in the districts of New York City using multiple open datasets. The results show significant improvement in terms of prediction accuracy at both levels. In addition, the applicability of the model to various buildings in the city is remarkably enhanced.

Neural Network-Based Building Energy Models for Adapting to Post-Occupancy Conditions: A Case Study for Florida

Abstract Building energy models (BEMs) are usually developed by subject matter experts during the design phase to help with decision making for achieving a more energy-efficient design at a minimum cost. The energy performance of a building is subject to significant changes as its operational parameters vary (e.g., occupancy, schedule of operation, etc.) due to different reasons such as change in building spaces application, demands, pandemic situation, among other reasons. In other words, a BEM that is created based on “as-designed” condition to predict building energy consumption (EC) can potentially become much less accurate during the lifetime of the building given the potential changes to the “in-operation” conditions. While BEMs can be adjusted to address operational changes, the end-user (i.e., building owner, manager, etc.) usually does not possess the knowledge to work with physics-based models (e.g., eQUEST) and therefore the initial BEM may no longer be of use to them. In the present paper, an approach is proposed and assessed through which a physics-based model is developed using eQUEST and simulated for several different operating conditions. The resulting data are then used for training an artificial neural network (ANN) which can serve as a simple and data-driven model for prediction of building energy consumption in response to changes in operating conditions. A case study is performed for a building on the campus of Florida Institute of Technology, to explore the changes that occurred in the building schedule of operation during COVID-19 pandemic and its impact on the performance of BEM. The inputs to the ANN are considered average daily values for outside dry bulb temperature, total daily global horizontal irradiation, hours of operation for the heating, ventilation, and air conditioning (HVAC) system for the main building, and hours of operation for the HVAC system for the conference room, while the output is considered as the monthly energy consumption of the building. The trained ANN is then tested against the actual measured data for energy consumption (post-construction) under different scenarios and good agreement between the results is found. The approach presented in this work aims to serve as a methodology for using data-driven surrogate models that can be used beyond the construction phase of the building and in response to sudden changes in building operating conditions.

A Comprehensive Study on Integrating Clustering with Regression for Short-Term Forecasting of Building Energy Consumption: Case Study of a Green Building

Integrating clustering with regression has gained great popularity due to its excellent performance for building energy prediction tasks. However, there is a lack of studies on finding suitable regression models for integrating clustering and the combination of clustering and regression models that can achieve the best performance. Moreover, there is also a lack of studies on the optimal cluster number in the task of short-term forecasting of building energy consumption. In this paper, a comprehensive study is conducted on the integration of clustering and regression, which includes three types of clustering algorithms (K-means, K-medians, and Hierarchical clustering) and four types of representative regression models (Least Absolute Shrinkage and Selection Operator (LASSO), Support Vector Regression (SVR), Artificial Neural Network (ANN), and extreme gradient boosting (XGBoost)). A novel performance evaluation index (PI) dedicated to comparing the performance of two prediction models is proposed, which can comprehensively consider different performance indexes. A larger PI means a larger performance improvement. The results indicate that by integrating clustering, the largest PI for SVR, LASSO, XGBoost, and ANN is 2.41, 1.97, 1.57, and 1.12, respectively. On the other hand, the performance of regression models integrated with clustering algorithms from high to low is XGBoost, SVR, ANN, and LASSO. The results also show that the optimal cluster number determined by clustering evaluation metrics may not be the optimal number for the ensemble model (integration of clustering and regression model).

A short-term building energy consumption prediction and diagnosis using deep learning algorithms

Short-term energy consumption prediction of buildings is crucial for developing model-based predictive control, fault detection, and diagnosis methods. This study takes a university library in Xi’an as the research object. First, a time-by-time energy consumption prediction model is established under the supervised learning approach, which uses a long short-term memory (LSTM) network and a Multi-Input Multi-Output (MIMO) strategy. The experimental results validate the model’s validity, which is close enough to physical reality for engineering purposes. Second, the potential of the people flows factor in energy consumption prediction models is explored. The results show that people flow has great potential in predicting building energy consumption and can effectively improve the prediction model performance. Third, a diagnostic method, which can recognize abnormal energy consumption data is used to diagnose the unreasonable use of the building during each hour of operation. The method is based on differences between actual and predicted energy consumption data derived from a short-term energy consumption prediction model. Based on actual building operation data, this work is enlightening and can serve as a reference for building energy efficiency management and operation.

Performance evaluation of short-term cross-building energy predictions using deep transfer learning strategies

Performing accurate building energy prediction (BEP) is one of the most important foundations for achieving energy resource allocation and developing energy efficiency measures. Buildings are diverse and operate under complex conditions leading to distribution differences of energy-related data among different buildings. Owing to such differences, the already reported data-driven BEP models exhibit poor cross-building prediction performance since they only use insufficient operation data of a single building. Although several deep transfer learning (DTL) strategies have been applied with improved cross-building prediction performance, there is still a lack of comparison of various DTL strategies, which would help to determine the optimal DTL strategy for different scenarios. Hence, three DTL strategies were compared: network-based Fine-tune, adversarial-based domain adversarial neural network (DANN), and mapping-based domain adaptive neural network (DaNN). The usefulness of the DTL strategy was validated using the open source dataset Building Data Genome Project 2 in the scenario of insufficient available data for real buildings. The influences of several factors were analysed, such as the available data volumes of the source and target domains within the training set. The applicability of different DTL strategies was discussed considering both accuracy and computational cost. Results show that the three DTL strategies outperform the traditional long short-term memory (LSTM) with an average BEP performance improvement ratio (PIR) of 0.75. For the extremely limited amount of available training data scenario, Fine-tune was recommended for the next-few-weeks prediction owing to its good balance between time cost and prediction performance. For the data shortage scenario of BEP tasks for nearly a year, DANN was recommended owing to its outperforming prediction accuracy. This provides insights for practical applications of DTL during the development of BEP models for cross-building BEP tasks without sufficient operation data.

Data-driven prediction of building energy consumption using an adaptive multi-model fusion approach

This paper develops an adaptive multi-model fusion approach to predict building energy consumption, aiming to give useful suggestions for better energy control. The building energy benchmarking dataset of Chicago in 2017 is selected as the case study, where 9 features are selected as the input variables aiming to estimate the weather normalized site energy use intensity of buildings. The training dataset is clustered using the K-means algorithm and sub-models are trained based on the clustered data using the XGBoost algorithm. The sub-models are then fused by assigning a weight considering both the model reliability and the matching degree and adopting a screening algorithm to weed out the unmatching sub-models, where the influence of the threshold in the screening algorithm is studied. The root mean square error of the estimation results from a fused model is found to be 13.42 which achieves a 7.6% amelioration compared with a single model. Moreover, the adaptive multi-model fusion approach is also proved to outperform both the two-stage clustering-based regression method and the linear fusion method. Benefiting from proper treatment of samples in the fuzzy zones between clusters and the screening algorithm in the fusion process, the method proposed in our paper eventually serves as more advanced guidance in the analysis and control of building energy performance.

Machine learning to predict building energy performance in different climates

Digitalization is sweeping the world of buildings. Notably, the use of machine and deep learning techniques to develop buildings’ digital twins is becoming crucial to foster the energy transition of the construction sector and a sustainable urban growth. Digital twins can ensure a user-friendly, fast and reliable prediction of building energy loads and demands, thereby enabling a comprehensive optimization of planning, design and operation. Accordingly, this study investigates machine learning techniques to predict heating loads of a building in Rome (Italy, Mediterranean conditions, “Csa” climate in the Köppen and Geiger classification) and in Berlin (Germany, European backcountry, “Cfb”). Firstly, the real building, located in Benevento, is used to develop the artificial neural networks (ANNs), then implemented in MATLAB® to achieve meta-models of building energy behavior. NARX (nonlinear autoregressive model with exogenous inputs) networks are used and trained based on simulated data, provided by the well-known building simulation tool EnergyPlus using the software DesignBuilder® as interface. The meta-model inputs are related to weather conditions, while the required outputs concern the thermal energy load for space heating. The analysis is performed with reference to annual forecasts of energy demands. In all cases, the ANNs architecture is optimized to achieve the best fitness with EnergyPlus outputs. The results show that machine learning can be a precious and reliable tool to support energy design and operation of different buildings in different climates. Nonetheless, the meta-modeling procedure needs to be properly conducted by experts to set suitable frameworks and hyperparameter values of the ANNs, as well as to achieve a right and comprehensive interpretation of the results.