REDUCTION AND IDENTIFICATION OF THERMAL DISCRETE TRANSFER-FUNCTIONS

For the reduction and identification of thermal discrete transfer functions of infinite order, existing techniques are briefly examined and a rule of thumb is shown. A linear identification model based on the standard 2-norm error is introduced in order to satisfy explicitly the original steady-state performance and also to minimize the transient performance errors of the reduced system. For this purpose, the solution involving an optimization problem is obtained for a thermal system of wall-response factors or room-weighting factors.