Search

Accurate short-term forecasting of building cooling load (CLF) in an integrated energy system (IES) is essential for effective building energy management. However, the existing CLF models for IES often treat each building as an independent entity and neglect the spatiotemporal correlation among buildings. To address this research gap and achieve accurate CLF, this paper proposes a new hybrid deep learning model that considers spatiotemporal coupling. First, the coupled spatial–temporal features among different buildings were analyzed, and the meteorological factors were screened based on the Spearman's rank order correlation coefficient (SROCC). Second, synchrosqueezing wavelet denoising (SWT) was adopted to denoise the historical cooling load (CL) data, remove high-frequency noise, and improve data quality. Third, the TTGAT-GTC model was constructed for the CLF of an IES. A temporal trend-aware graph attention network (TTGAT) captured the spatial correlation of CL between buildings. A gated temporal convolution layer (GTC) was constructed to extract the trend in the dynamic temporal variation in historical load. Residual and skip connections were applied to avoid gradient disappearance and increase the computational efficiency of the model. To validate the effectiveness of the proposed SWT-TTGAT-GTC model, this paper compared the proposed model with four benchmark models using MAPE, RMSE, MAE, and R2. The experimental results showed that the performance of the proposed CL forecasting model is superior and that the proposed model appropriately introduces the spatio-temporal coupling information between buildings. • The spatiotemporal information between related buildings is appropriately considered in CL. • TTGAT captures the spatial correlation of cooling load between buildings. • GTC extracts the trend in dynamic temporal variation between buildings. • Residual and skip connections are used to avoid gradient disappearance. [ABSTRACT FROM AUTHOR]

This paper reports a new knowledge-driven engineered feature learning approach in response to the Global AI Challenge for Building E&M Facilities held by the Electrical and Mechanical Service Department (EMSD) of the Hong Kong SAR. The results were awarded with a Grand Prize by the competition organizer. A dynamically engineered multi-modal feature learning (DEMMFL) method is proposed for predicting the cooling load of two office buildings. The DEMMFL model is estimated with the Lasso-ridge regression and compared with other well-known methods such as the Lasso. The novel approach applies control system knowledge to engineer useful features and explore load patterns for multi-mode modeling. Deep learning methods including LSTM, GRU, and AutoGluon are implemented for automated machine learning and tested in parallel to compare the performance of the proposed model with existing methods. The proposed model is demonstrated to predict long-term cooling load most accurately using engineered features from weather information only. • A new dynamic feature engineering method for office building cooling load prediction. • Slow and fast dynamic features engineered from available weather data only. • Control-knowledge based feature to learn on-hours dynamic responses. • Multi-modal approach with eight day-type features to learn occupancy loads. • Deep recurrent learning and AutoML methods compared to the proposed method. [ABSTRACT FROM AUTHOR]

PV windows are considered to be a promising building-integrated PV technology due to their excellent electrical, daylighting and thermal performance. However, the shading of the PV cells increases the heating load in winter and the high temperature of PV cells exacerbates the cooling load in summer. Moreover, a significant portion of the heat absorbed by the PV modules is usually wasted. To address these issues, this paper proposed a multi-functional PV/T window. This system could recover the heat on PV cells for producing warm air in winter and hot water in other seasons. As a result, the proposed system can realize the combined PV/T utilization of solar energy and reduce the seasonal thermal needs of the building. First, the mathematical model of the proposed window was developed and validated. Next, the developed model was integrated into building energy software to evaluate the overall energy performance of the proposed window in four different cities of Yangtze River region: Shanghai, Nanjing, Wuhan, and Chongqing in China. These cities experience the subtropical humid climate (Cfa in the Köppen-Geiger classification) and belong to the zones with medium global horizontal irradiation. The sensitivity analysis revealed that the optimal slat angles are between 0o and 40o, while the optimal orientation angles range from 30o to 60o. Selecting low-iron glass as the inner glass resulted in the lowest overall energy consumption. By adopting the MFPV/T window, the annual operation cost of the studied case could be reduced by 34% (6.4 $/m2), 36% (7.0 $/m2), 28% (5.7 $/m2), and 16% (2.9 $/m2) in these four cites, respectively. Additionally, the annual CO 2 emissions could be curtailed by 53.5, 58.5, 47, and 22.3 kg/m2, respectively. • A novel PV/T window was proposed for addressing the common issues of PV windows. • The mathematical model of proposed window was developed and validated. • The parameter sensitivity was analyzed. • The overall energy performance was predicted in Yangtze River region. [ABSTRACT FROM AUTHOR]

• A rolling optimization strategy is proposed for microgrid operation optimization. • Flexible load and hybrid energy storage are applied jointly to power regulation. • A priority-based power fluctuations smoothing method is first proposed. • Energy trigger mechanism is designed to keep the optimal state of supercapacitor. The rise of microgrids provides an effective solution to the problem of local consumption of renewable energy sources. However, the power fluctuations are the crucial issue for the widespread adoption of the grid-connected microgrid with renewable energy sources. We aim to economically and locally solve the problem of grid-connected power fluctuations of microgrid. In this paper, a novel rolling optimization strategy considering grid-connected power fluctuations smoothing for microgrids is provided. Firstly, the mathematical model of the microgrid is described, which contains the grid-connected power limits and supercapacitor-battery hybrid energy storage system. Then, a priority-based smoothing method of power fluctuations is proposed for the first time. As a flexible load, the heating/cooling load is used as virtual energy storage to participate in power regulation. Finally, the rolling optimization strategy integrated with the energy trigger mechanism is designed to achieve the operation optimization. The objectives of this paper are to help the microgrids improve grid-connection friendliness and minimize the daily operating costs that include fluctuation penalties. Simulation results on different scenarios for a grid-connected microgrid show that the novel rolling optimization strategy has better performance in any scenario. Compared with the traditional strategy, the total operating costs of the proposed strategy can save 5.67% for a given scenario. We can conclude that the proposed rolling optimization strategy can effectively reduce cost payment and meet the requirement of grid-connected power fluctuations smoothing. [ABSTRACT FROM AUTHOR]

• Backup capability of solar and storage differs by region in the U.S. • Small batteries can meet a limited set of critical loads in most counties. • Electric cooling and heating loads are typically hardest loads to backup. • Household building attributes like efficiency can impact results up to 20% • Resiliency value needs to be high to motivate adoption of solar-plus-storage. Customer concerns over electric system resilience could drive early adoption of behind-the-meter solar-plus-storage (BTM PVESS), especially as wildfire, hurricane, and other climate-driven risks to electric grids become more pronounced. However, the resilience benefits of BTM PVESS are poorly understood, especially for residential customers, owing to lack of data and methodological challenges, making it difficult to forecast adoption trends. In this paper, we develop a methodology to model the performance of BTM PVESS in providing backup power across a wide range of customer types, geography / climate conditions, and long duration power interruption scenarios, considering both whole-building backup and backup of specific critical loads. We combine novel, disaggregated end-use load profiles across the continental United States with temporally and geospatially aligned solar generation estimates. We then implement a PVESS dispatch algorithm to calculate the amount of load served during interruptions. We find that PVESS with 10 kWh of storage can meet a limited set of critical loads in most United States counties during any month of the year, though this capability drops to meeting only 86% of critical load, averaged across all counties and months, when heating and cooling are considered critical. Backup performance is lowest in winter months where electric heat is common (southeast and northwest U.S.) and in summer months in places with large cooling loads (southwest and southeast U.S.). Winter backup performance varies by roughly 20% depending on infiltration rates, while summer performance varies by close to 15% depending on the efficiency of the central air-conditioning system. Differences in temperature set-points in Harris County correspond to a 40% range in winter backup performance and a 20% range in summer performance. Economic calculations show that a customer's resilience value of PVESS must be high to motivate adoption of these systems. [ABSTRACT FROM AUTHOR]

• The H 2 -O 2 PEMFC adopts a DEAC operation mode to achieve 100 % utilization of reactants. • PEMEC-PEMFC achieves self-sufficiency of reactants with water self-supply. • The number of PEMEC varies according to solar radiation and light time. • The hourly loads of households can be satisfied by adopting the pre-operation mode. Targeting the shortcomings of oxygen waste and reactant supplement of electrolyzer, this paper proposes a novel combined cooling, heating and power (CCHP) system based on proton exchange membrane electrolysis cell (PEMEC) and H 2 /O 2 proton exchange membrane fuel cell (PEMFC) with water self-supply. In this system, H 2 /O 2 PEMFC adopts a dead-ended anode and cathode operation mode to achieve 100 % hydrogen and oxygen utilization and water recovery. For the application scenario of a household in Shanghai, China, the performance of the CCHP system in summer and winter is investigated. When the system gives priority to meeting the electric and cooling load, there will be insufficient heat at the initial stage of operation. The pre-operation model can realize the hourly matching of residential electric, heating and cooling loads. After one day of operation in summer and winter, there are 19.1 kWh and 45.5 kWh electricity surplus, 7.27 × 106 kJ and 1.97 × 106 kJ heat surplus, 1200 mol and 300 mol hydrogen surplus, respectively. Moreover, the estimated payback period for this system is 7.6 years, which will be significantly shortened with the decrease of the cost of photovoltaic panels and PEMEC. [ABSTRACT FROM AUTHOR]

• The size and shape of a building have considerable influence on both embodied and operational efficiency. • Lower window-to-wall ratios lead to lower heating and cooling across all building types and climates. • Mechanical ventilation with heat recovery significantly increases operational efficiency for most building types and climate zones. • The choice of structural frame type has significant impact on embodied carbon and cost, but negligible influence on heating and cooling loads. • Compactness and low window-to-wall ratios simultaneously benefit embodied and operational efficiency and are decisive to enable meeting the Passivhaus standard. Buildings account for over one-third of global emissions and energy use. Meeting climate pledges will require achieving high operational energy efficiency with low embodied impacts in new construction. Yet, a systematic identification of the relative influence of building design parameters on both operational and embodied efficiencies has rarely been attempted. In this paper we explore for the first time the sensitivity of a wide range of design and operation parameters in terms of embodied carbon, construction cost, as well as heating and cooling loads for multi-storey buildings. We devised a model to estimate the relative importance of a large set of input variables, describing a building's shape, size, layout, structure, ventilation, windows, insulation, air, and use for residential and office multi-storey buildings, across different climates. We found that increasing building compactness, using steel or timber instead of concrete frames, lowering window-to-wall ratio, choosing the most suitable glazing, and employing mechanical ventilation with heat recovery are the most important measures to decrease embodied emissions and operational energy. The most significant trade-offs with construction cost were found for the choice of frame material and in the decision whether to install mechanical ventilation. We estimate that 28–44% of yearly heating and cooling energy and 6 Gt cumulative embodied CO 2 e until 2050 could be saved in multi-storey buildings, without employing new technologies. [ABSTRACT FROM AUTHOR]

• A novel Dendrite Net based on the adaptive mean square gradient is proposed. • Energy consumption prediction and analysis model of buildings is obtained based on energy consumption data of buildings. • The prediction accuracy of the proposed model is higher than most neural network methods. • The optimization direction of building envelopes is found for reducing energy consumption. • The maximum carbon dioxide emissions can be reduced by 158.17 kg. Artificial neural networks have been widely applied in construction industries. Due to dendrites of biological neurons participating in the pre-calculation of neural networks, the structure of the traditional artificial neural network needs to be adjusted subjectively. Thus, a novel dendrite net based on the adaptive mean square gradient is proposed in this paper. The energy consumption of buildings is predicted and analyzed by the proposed method to cut the carbon dioxide emissions. The adaptive mean square gradient method is used to update the weight matrix of the dendrite net method, which can avoid errors caused by selecting hidden layer nodes to improve the prediction accuracy of the proposed method. Finally, the proposed method is used to energy saving and emission reducing of the construction industry. Compared with other methods, the experimental results show the availability of the proposed method. Through predicting the heating and cooling loads based on the proposed method, the construction plan is adjusted to decrease the energy consumption of buildings. Moreover, the appliances energy consumption is predicted and analyzed by the proposed method to improve energy efficiency and cut carbon dioxide emissions. [ABSTRACT FROM AUTHOR]

[Display omitted] • Power-to-gas facility is integrated into the trigeneration energy system. • Optimal operation of this integrated system is studied under multiple uncertainties. • A novel distributionally robust operation optimization model is proposed accordingly. • Economic and environmental benefits are improved by operating the integrated system. • Operation efficiency in various aspects are guaranteed by the proposed model. Power-to-gas is an emerging energy conversion technology. When integrating power-to-gas into the combined cooling, heating and power system, renewable generations can be further accommodated to synthesize natural gas, and additional revenues can be obtained by reutilizing and selling the synthesized gas. Therefore, it is necessary to address the optimal operation issue of the integrated system (Combined cooling, heating and power-Power-to-gas) for bringing the potential benefits, and thus promoting energy transition. This paper proposes a Wasserstein and multivariate linear affine based distributionally robust optimization model for the above issue considering multiple uncertainties. Specifically, the uncertain distribution of wind power and electric, thermal, cooling loads is modeled as an ambiguity set by applying the Wasserstein metric. Then, based on the ambiguity set, the proposed model with two-stage structure is established. In the first-stage, system operation cost (involving the energy exchange and carbon emission costs, etc.) is minimized under the forecast information. In the second stage, for resisting the interference of multiple uncertainties, the multivariate linear affine policy models are constructed for operation rescheduling under the worst-case distribution within the ambiguity set, which is capable of adjusting flexible resources according to various random factors simultaneously. Simulations are implemented and verify that: 1) both the economic and environmental benefits of system operation are improved by integrating power-to-gas; 2) the proposed model keeps both the conservativeness and computational complexity at low levels, and its solutions enable the effective system operation in terms of cost saving, emission reduction, uncertainty resistance and renewable energy accommodation. [ABSTRACT FROM AUTHOR]

• Open-source Modelica models for chiller plants with water-side economizer were developed. • The developed model can evaluate the control performance of multiple control sequences. • Flexible capability of the developed chiller plants with WSE models was demonstrated. • Opportunities in improving the control performance of water-side economizer were identified. There are several cooling mode control sequences for chiller plants with water-side economizers adopted in industry and academia, and it is widely known that this supervisory control significantly affects energy consumption; however, there is a lack of a modeling resource to allow multiple control sequences to be evaluated systematically under different settings, such as system configurations, load types, and climate locations. To fill this gap, this paper develops open-source Modelica models to simulate the control and energy performance of multiple cooling mode control sequences for chiller plants with water-side economizers. These models allow users to develop and test their advanced control sequences for chiller plants with water-side economizers for their target climates and system configurations. To demonstrate how these models can be utilized, a chiller plant with an integrated water-side economizer is simulated using two advanced cooling mode control sequences, two cooling load types, and six climates, for a total of 24 simulation cases. This study revealed that the energy saving potential varied from 8.6% to 36.8% for constant load profiles in all of the considered climates, and from 6.3% to 25.8% for variable load profiles in most of the climates. Results also showed that the developed system models are able to capture transient control details and reveal counterintuitive energy performance. [ABSTRACT FROM AUTHOR]

• Increasingly complex machine learning models lead to lower interpretability. • An attention-based neural network architecture for load prediction is proposed. • The architecture is tested using real operational data from a commercial building. • Proposed architecture shows enhanced prediction accuracy and interpretability. Machine learning has gained increasing popularity in building energy management due to its powerful capability and flexibility in model development as well as the rich data available in modern buildings. While machine learning is becoming more powerful, the models developed, especially artificial neural networks like Recurrent Neural Networks (RNN), are becoming more complex, resulting in "darker models" with lower model interpretability. The sophisticated inference mechanism behind machine learning prevents ordinary building professionals from understanding the models, thereby lowering trust in the predictions made. To address this, attention mechanisms have been widely implemented to improve the interpretability of deep learning; these mechanisms enable a deep learning-based model to track how different inputs influence outputs at each step of inference. This paper proposes a novel neural network architecture with an attention mechanism for developing RNN-based building energy prediction, and investigates the effectiveness of this attention mechanism in improving the interpretability of RNN models developed for 24-hour ahead building cooling load prediction. To better understand, explain and evaluate these neural network-based building energy prediction models, the obtained attention vectors (or metric) are used to visualize the influence of different parts of model inputs on the prediction result. This helps the users to understand why predictions are made by the model, as well as how input sequences proportionally influence the output sequences. Further analysis of attention vectors can provide interesting temporal information for understanding building thermal dynamics, like the thermal inertia of the building. The proposed attention-based architecture can be implemented in developing optimal operation control strategies and improving demand and supply management. The model developed based on this architecture is assessed using real building operational data, and shows improved accuracy and interpretability over baseline models (without adopting attention mechanisms). The research results help to bridge the gap between building professionals and advanced machine learning techniques. The insights obtained can be used as guidance for the development, fine-tuning, explanation and debugging of data-driven building energy prediction models. [ABSTRACT FROM AUTHOR]