Developing a whole building cooling energy forecasting model for on-line operation optimization using proactive system identification

Optimal automatic operation of buildings and their subsystems in responding to signals from a smart grid is essential to reduce energy demand, and to improve the power resilience. In order to achieve such automatic operation, high fidelity and computationally efficiency whole building energy forecasting models are needed. Currently, data-driven (black box) models and hybrid (grey box) models are commonly used in model based building control. However, typical black box models often require long training period and are bounded to building operation conditions during the training period. On the other hand, creating a grey box model often requires (a) long calculation time due to parameter optimization process; and (b) expert knowledge during the model development process. This paper attempts to quantitatively evaluate the impacts of two significant system characteristics: system nonlinearity and response time, on the accuracy of the model developed by a system identification process. A general methodology for building energy forecasting model development is then developed. How to adapt the system identification process based on these two characteristics is also studied. A set of comparison criteria are then proposed to evaluate the energy forecasting models generated from the adapted system identification process against other methods reported in the literature, including Resistance and Capacitance method, Support Vector Regression method, Artificial Neural Networks method, and N4SID subspace algorithm. Two commercial buildings: a small and a medium commercial building, with varying chiller nonlinearity, are simulated using EnergyPlus in lieu of real buildings for model development and evaluation. The results from this study show that the adapted system identification process is capable of significantly improve the performance of the energy forecasting model, which is more accurate and more extendable under both of the noise-free and noisy conditions than those models generated by other methods.