AIL Research monitored four novel photovoltaic heat pumps (PVHP) for the Electric Power Research Institute (EPRI).  The test sites were in Mesa, AZ, Houston, TX, Marfa, TX,  and Huntersville, NC.

The PVHP used backup power from the utility grid when it required more power than the photovoltaic (PV) array could deliver.  The PVHP had no provisions for energy storage, so the PV array floated at its open-circuit voltage when the heat pump was off.

Three test sites were residences and one was a NASA laboratory.  The work at the laboratory produced considerable heat, so this site required cooling the entire year.  At all sites, the PV array was sized to provide 50% to 60% of the full-load power of the heat pump.  (The nominal output of the arrays ranged from 2.0 kW to 3.2 kW, and the heat pumps' cooling capacities were 3, 4 and 5 tons.)  Despite this modest undersizing of the PV arrays, their utilization at the residences sites was poor when averaged over the entire year: 23% to 47% of the full potential output of the arrays was delivered to the heat pumps.  During summer months, the utilization of the arrays was much higher, reaching 85% for one month at one site.  At the NASA laboratory, 74% of the potential output of the array was used, due to the high cooling loads that persisted for the entire year.

The PVHP was not very effective at reducing electrical demand.  For the residences, the cooling loads tended to peak in the late afternoon when solar insolation was relatively low.  The cooling loads at the laboratory had a very broad peak that began midday but extended into the late afternoon.  For the ten hottest days at each site, the PVHP reduced the average peak electrical demand from the grid by 70 to 570 W at the residences and 1100 W at the laboratory.

EPRI's final report TR-109256 entitled "Field Testing of Photovoltaic Heat Pumps" presents detailed results for these field tests.