System parameter identification theory and uncertainty analysis methods for multi-zone building heat transfer and infiltration

Methods for on-site measurement of building thermal performance system parameters such as coefficient of heat loss, solar heat gain, effective thermal capacity, infiltration rate, and effective mixing volume are very important, yet a nontrivial task. Although these are steady-state parameters, on-site measurements are exposed to changing meteorological conditions and are affected by the thermal capacity of the building. In addition, these parameters should generally be estimated by using a multi-zone model such as inter-zone flow rates. In this regard, a state space equation model, referred to as a “thermal network model,” has been devised to generalize such multi-zone heat transfer system and tracer gas diffusion system measurements. This model is composed of three parameter types, and we have developed a system parameter identification theory and uncertainty analysis method using least squares, as well as actual measurement systems. In the present paper, we improve the least-squares regression equation, the uncertainty analysis method, and the reliability evaluation method. We investigate appropriate excitation waveforms and frequencies for heating and tracer gas release, as well as a low-pass filter for pre-processing measurement data. We verify these theories and methods using computer-simulated measurement.

► State space equation model for multi-zone building heat and tracer gas transfer.
► A composite regression equation formulated using two constraints of quadratic forms.
► Low-frequency sinusoidal excitation is suitable for a rough system identification model.
► A low-pass filter using the moving term average of measurement data is necessary.
► A discrepancy ratio is defined to identify system identification premise failures.