Search

An accurate and fast building load prediction model is critically important for guiding building energy system design, optimizing operational parameters, and balancing a power grid between energy supply and demand. A physics-based simulation tool is traditionally used to provide the building load demand; however, it is constrained by its complex model development process and requirement for engineering judgments. Machine learning algorithms (i.e., data-driven models) based on big data can bridge this gap. In this study, we used the massive energy data generated by a physics-based tool (EnergyPlus) to develop three data-driven models (i.e., LightGBM, random forest (RF), and long-short term memory (LSTM)) and compared their prediction performances. The physics-based models were developed using office prototype building models as baselines, and ranges were provided for selected key input parameters. Three different input feature dimensions (i.e., six-, nine-, and fifteen-input feature selections) were investigated, aiming to meet different demands for practical applications. We found that LightGBM significantly outperforms the RF and LSTM algorithms, not only with respect to prediction accuracy but also in regard to computation cost. The best prediction results show that the coefficient of variation of the root mean squared error (CVRMSE), squared correction coefficient (R2), and computation time are 5.25%, 0.9959, and 7.0 s for LightGBM, respectively, evidently better than the values for the algorithms based on RF (18.54%, 0.9482, and 44.6 s) and LSTM (22.06%, 0.9267, and 758.8 s). The findings demonstrate that a data-driven model is able to avoid the process of establishing a complicated physics-based model for predicting a building’s thermal load, with similar accuracy to that of a physics-based simulation tool.

The penetration rates of intermittent renewable energies such as wind and solar energy have been increasing in power grids, often leading to a massive peak-to-valley difference in the net load demand, known as a “duck curve”. The power demand and supply should remain balanced in real-time, however, traditional power plants generally cannot output a large range of variable loads to balance the demand and supply, resulting in the overgeneration of solar and wind energy in the grid. Meanwhile, the power generation hours of the plant are forced to be curtailed, leading to a decrease in energy efficiency. Building demand response (DR) is considered as a promising technology for the collaborative control of energy supply and demand. Conventionally, building control approaches usually consider the minimization of total energy consumption as the optimization objective function; relatively few control methods have considered the balance of energy supply and demand under high renewable energy penetration. Thus, this paper proposes an innovative DR control approach that considers the energy flexibility of buildings. First, based on an energy flexibility quantification framework, the energy flexibility capacity of a typical office building is quantified; second, according to energy flexibility and a predictive net load demand curve of the grid, two DR control strategies are designed: rule-based and prediction-based DR control strategies. These two proposed control strategies are validated based on scenarios of heating, ventilation, and air conditioning (HVAC) systems with and without an energy storage tank. The results show that 24–55% of the building’s total load can be shifted from the peak load time to the valley load time, and that the duration is over 2 h, owing to the utilization of energy flexibility and the implementation of the proposed DR controls. The findings of this work are beneficial for smoothing the net load demand curve of a grid and improving the ability of a grid to adopt renewable energies.