Demand Side Management (DSM) programs provide utility companies with amethod to shift consumer electricity usages away from peak electricity hours. DSMprograms use alternative appliance usage schemes that maintain their usefulness whileproviding ancillary services for utilities. This thesis aims to develop a linear controlmethodology that can provide signi cant ancillary services for utilities without re-ducing customer comfort.A prototype enclosure was built and equipped with a heater and thermalmeasuring equipment. Data was collected during a 17 hour temperature regulationexperiment using a bang-bang controller similar to those commonly used for residen-tial heating control. An experimental thermal system identi cation methodology wasdeveloped for online system identi cation. First and second order mathematical mod-els were developed for thermal system identi cation. The mathematical models werecalibrated using data collected experimentally and used to estimate the net thermalresistance and capacitance using system identi cation techniques.The enclosure system model was also used to determine if peak power couldbe reduced by slowly varying loads utilizing a di erent type of controller. Two di er-ent linear control techniques (using K-Factor and PI approaches) and the associatedpower electronics circuitry were implemented and tuned in PSpice platform. Both controller systems successfully leveled the load and reduced the peak power demand.Finally the prototype enclosure was modi ed to include a linear controller us-ing an available DC power supply and a buck converter power stage. The PI controlscheme was used with a 60  phase margin for smoother and faster settling charac-teristics. The phase margin was acquired using appropriate linear approximation ofsystem transfer functions. The temperature response of the experimental system wascompared to theoretical responses.