STOCHASIC CALIBRATION OF ENERGY MODEL FOR EXISTING OFFICE BUILDINGS

摘要

Energy consumption by existing buildings accounts for a significant portion of national energy use. To assess . energy performance of existing buildings as well as to make rational decision for building energy retrofit, it is necessary to have a scientific and objective energy model. Such energy model must be very simple, accurate enough and easy-to-use for daily practice. The approach presented in this paper utilizes ISO 13790 simple hourly calculation method, rather than resorting to full-blown dynamic simulation tools (e.g. EnergyPlus) which is very time & labor demanding. One of the interesting features of the approach is that it employs a stochastic parameter estimation technique so called Bayesian calibration. The energy model enhanced by the calibration technique can predict stochastic nature of building energy consumption. The paper reports the results of such application to existing office buildings in terms of accuracy, usability, and limitations.