More than two hundred sixty grid-tied photovoltaic (PV) systems sized 30 kW to 1.1 MW installed in California during 2002 through 2004 received partial funding through the Self-Generation Incentive Program (SGIP). The SGIP is administered statewide by PG&E, SCE, SoCalGas, and the San Diego Regional Energy Office. The incentive is structured as a one-time capacity based payment made at the time of system completion. The first PV system incentive was paid in Summer 2002. Through the end of 2004, a total of 269 PV systems had received financial support through the program. The cumulative generation capacity of these systems exceeded 30 MW and corresponded to $101 million of incentives paid. While originally slated to run through 2004, recently the program was modified and extended through the end of 2007. PV systems participating in the program are being monitored to support evaluation of the program. These data have been used to assess impacts of the Program on peak demand and energy consumption. These data have also been incorporated into the Program’s cost-effectiveness assessment. Well over one-half of the PV systems have already been subject to metering yielding 15-minute interval generator output data. The cumulative size of the directly monitored PV systems currently exceeds 33 MW as of late 2005. In 2004, the statewide California Independent System Operator (ISO) electrical system peak occurred on September 8 during the 16th hour (from 3 to 4 PM PDT). During this hour the electrical demand for the California ISO reached 45,562 MW. On this day, there were 235 PV systems funded under the SGIP installed and operating; interval-metered data are available for 107 of these projects. The resulting estimate of peak demand impact coincident with the ISO peak load totals 9,938 kW. The estimated peak demand impact corresponds to 0.39 kW per 1.0 kWRebated of PV system size and is based on rebated capacity. Those unfamiliar with PV system size ratings and PV system operating characteristics may be surprised that the overall weighted-average peak demand impact was not substantially higher at this hour and time of year. To help put this result in perspective, it can be compared to a simple engineering estimate of peak power output based on published information regarding PV system performance. First, we begin with 1 kW [basis: rebated size] of horizontal PV system capacity. For purposes of determining rebates, PV system sizes are calculated as the product of cumulative estimated module DC power output under PTC conditions and inverter maximum DC to AC conversion efficiency. Factors such as manufacturing tolerance, soiling, module mismatch, temperature effects, and wiring losses may result in actual full-sun power output levels of about 0.76 kW/kWRebated. When the 3 to 4 PM angle of incidence effects for the month of September are included the expected output value drops significantly further. The peak-day operating characteristics of the 107 PV projects for which peak-day interval-metered data were available are summarized in the box plot of Figure 4. System sizes were used to normalize power output values prior to plotting summary statistics of PV output profiles for individual projects. The normalized values represent PV power output per unit of system size. Treatment in this manner enables direct comparison of the power output characteristics of PV systems of varying sizes. The vertically oriented boxes represent ranges within which 75% of project-specific values lie. The vertical lines represent the total range (i.e., maximum and minimum) of project-specific values. The energy production of the group of metered PV systems varied according to season. In Figure 7, normalized energy production by month is illustrated (on the right axis). These values represent the monthly average capacity factor for the on-line PV system capacity. As expected, normalized energy production levels reach their maximum values in the summer season and decrease towards the winter season as the intensity and duration of incident solar radiation falls off, coupled with increased incidence of storms and other weather disturbances off the Pacific Ocean, which affect the availability of solar radiation upon the PV modules.
Opera.DL: Deep Learning Modelling for Photovoltaic System Monitoring
