Solar and Snow: Unlikely allies on the energy landscape?
Prior to 2008, predicting photovoltaic (PV) system energy loss due to snow was a rare topic. The most significant markets for PV were in understandably sunny and warm climates, and no analytical models were available in the literature to predict the relationship between snow and energy loss. With the rapid expansion of North American PV markets in Canada, Colorado, and the New England and Atlantic states, DNV GL, via its predecessor firm BEW Engineering, jumped in to set up a first of its kind snow-on-PV test platform. This pioneering station, placed on a ranch near Lake Tahoe, California in 2009, used sets of paired clean and intermittently snow-covered PV modules, a datalogger, and a timed camera to record how tilt angle and snow patterns affected energy production. The Lake Tahoe testing yielded the first predictive model to correlate snow and PV array geometry. For the first time, monthly energy loss could be calculated for any climate and array tilt angle. That model, which is still in use, was published at the 2011 IEEE PV Specialists Conference as Photovoltaics and Snow: An Update from Two Winters of Measurements in the Sierra. The publication of this model improved the accuracy of our energy forecasts in challenging and emerging cold-weather markets, boosted our firm’s reputation in modeling complex PV systems, and prompted the sales of several additional comparable test platforms to private clients.
Fast forward to 2013, and DNV GL expanded our PV-snow research with two more test platforms, two enthusiastic institutional partners, and a two-year test plan. The National Renewable Energy Laboratory in Golden, CO and the Michigan Technological University in Houghton, MI began joint monitoring programs with DNV GL. Our Michigan Tech partners, with DNV GL engineer Tim Townsend, recently published a detailed article in the IEEE Journal of Photovoltaics titled Impact of Snow and Ground Interference on Photovoltaic Electric System Performance. This paper analyzes the first year of data from the Michigan test platform.
The current test platforms include four tilt angles ranging from 0 to 45 degrees as an improvement over the original test’s three tilt angles of 0 to 39 degrees. The new test, however, is also set up to measure an important commercial ground interference situation that the original testing was only partially able to address. In one extreme, sliding snow is never impeded by the ground or roofline below the module’s edge. At the opposite extreme, which often occurs on low-slope commercial rooftops, sliding snow piles up immediately, since the module’s lowest edge is placed adjacent to the ground. Between these extremes of ground interference, actual systems will experience losses that are the result of the combined and sometimes chaotic effects of array geometry and natural weather patterns. The just-published Journal article finds that while annual energy losses can range from 5 to 12% under normal Michigan conditions, these losses can be magnified by a factor of up 6 times worse if the array’s ability to shed the accumulated snow is regularly impeded by ground interference. The degree to which energy losses are magnified is shown to be worse for high tilt angles. Comparable tilt-angle effects have been observed at the Colorado site. The two-year test period wrapped up in October, and DNV GL will be publishing additional articles and an enhanced version of its unique analytical snow model in the coming months.
Read the full whitepaper “Impact of Snow and Ground Interference on Photovoltaic Electric Performance” here.