Physics-constrained Deep Learning of Multi-zone Building Thermal Dynamics
This work addresses the problem of accurate and interpretable thermal modeling for building energy management, representing an incremental advance by integrating physics-based knowledge into neural networks.
The paper tackled modeling building thermal dynamics by developing a physics-constrained deep learning method that incorporates structural priors and constraints to ensure physical realism, achieving significant accuracy improvements over prior state-of-the-art results using only 10 days of training data on a real-world office building with 20 thermal zones.
We present a physics-constrained control-oriented deep learning method for modeling building thermal dynamics. The proposed method is based on the systematic encoding of physics-based prior knowledge into a structured recurrent neural architecture. Specifically, our method incorporates structural priors from traditional physics-based building modeling into the neural network thermal dynamics model structure. Further, we leverage penalty methods to provide inequality constraints, thereby bounding predictions within physically realistic and safe operating ranges. Observing that stable eigenvalues accurately characterize the dissipativeness of the system, we additionally use a constrained matrix parameterization based on the Perron-Frobenius theorem to bound the dominant eigenvalues of the building thermal model parameter matrices. We demonstrate the proposed data-driven modeling approach's effectiveness and physical interpretability on a dataset obtained from a real-world office building with 20 thermal zones. Using only 10 days' measurements for training, we demonstrate generalization over 20 consecutive days, significantly improving the accuracy compared to prior state-of-the-art results reported in the literature.