Publicación: Analysis of smart thermostat thermal models for residential building
Cargando...
Fecha
2022-06-02
Editor/a
Director/a
Tutor/a
Coordinador/a
Prologuista
Revisor/a
Ilustrador/a
Derechos de acceso
info:eu-repo/semantics/openAccess
Título de la revista
ISSN de la revista
Título del volumen
Editor
Elsevier
Resumen
This work studies the thermal behavior of residential buildings by using the data provided by smart thermostats and weather forecast data. We consider an equivalent circuit model depending on four parameters related to the heater power, the solar energy, heat capacity, and the thermal resistance of the building. We employ a random ordinary differential equation to overcome the natural model uncertainty. We develop a differential equation constrained least-squares-based parameter identification approach for uncertainty quantification. We provide results related to the analytic properties of the parameter-to-solution map involving a derivative characterization. Furthermore, using a discretization scheme adapted to the given data, we derive discrete formulas for the deterministic identification model and compute the statistical moments of the random model. Following a machine learning approach, we propose an algorithm that consists of three phases. The first is a training phase where we identify the parameter uncertainties on a training dataset.
In the second phase, we establish a normal distribution of the parameters using these uncertainties. In the final phase, we simulate the temperature on a test dataset by solving the random model. We test this algorithm on real data from two residential buildings.
The detailed numerical experiments show the feasibility and the efficacy of the developed framework.
Descripción
Categorías UNESCO
Palabras clave
Constrained quadratic optimization, Uncertainty quantification, Machine learning, Thermal building models, Ordinary differential equations, Parameter estimation
Citación
Arias, J., Khan, A. A., Rodriguez-Uría, J., Sama, M. (2022). Analysis of smart thermostat thermal models for residential building. Applied Mathematical Modelling, 110, 241-261. DOI 10.1016/j.apm.2022.05.041.
Centro
Facultades y escuelas::Facultad de Ciencias
Departamento
Matemática Aplicada I