Dispersed PV: The Antidote to Short-Term Power Variability

Variability of photovoltaic systems is an important and often misunderstood topic.  SunPower has been a pioneer and leader in efforts to better characterize variability and its impacts, and has helped develop a substantial body of work on the topic.  It has been consistently demonstrated that the variability of a single, relatively small PV system is much greater than that of many distributed PV systems – and that large, utility-scale PV systems demonstrate substantially lower variability than small systems in the same location.  The following article by guest blogger Richard Perez provides an excellent introduction to the topic.  
 

Owners of PV systems know first-hand that passing clouds cause rapid changes in the amount of energy produced from second to second. The impact of this short term power “variability” on system owners is usually seen rather than felt because during times of low production, electricity from the grid picks up the slack. This is good for PV system owners, but makes it difficult for grid operators, who are charged with providing reliable energy for their customers. 

 

One way for utilities to manage variability is to ramp up other power plants, but that increases the costs and complexity of managing the electric power distribution system. Fortunately, recent studies have found that renewable energy sources that are distributed across a wide geographic range greatly reduce the effects of weather-related variability to the overall electric system. This means that more “dispersed” residential and commercial PV is good for utilities. 

 

Electric grid operators became aware of the difficulty variability posed when information about short-term variability in a 3.5-megawatt plant in Springerville, Arizona was widely circulated in 2009. The analysis revealed frequent ramping up and down of the plant’s production as a result of passing clouds. This prompted utilities and agencies across the U.S. to ask themselves: “How would power fluctuations be handled if PV reached a sizeable fraction of power production?” The implication of the Springerville analysis was that short-term fluctuations in power production within a particular region are an obstacle to large-scale PV deployment. 

 

As a result of the Springerville analysis, the questions about short-term variability were taken up by the U.S. Department of Energy and the California Solar Initiative. Unsurprisingly, research confirmed that conditions can be highly variable at any given location. But on the positive side, research also revealed that spreading PV systems out over a larger area mitigates the problems of short-term variability. The truth of this finding is supported by the probability theory’s law of large numbers, and has been proven through recent studies. 

 

The following image illustrates this principle by showing how distance can “smooth” variability. The data in the top part of the figure shows 10-second solar radiation (irradiance) at a single location in Napa, California, on November 21, 2010. The data in the bottom half of the figure presents the same irradiance data, but measured at 25 locations in a 1.5 square mile grid rather than at a single location. The data in the bottom half of the figure is much smoother, without large short-term fluctuations.
 

A similar study in New York compared the variability of a single PV system versus a large number of systems deployed over a 25 square mile area. The study found that power output variability of PV systems distributed across a region is similar to the demand-side variability impacts that utilities have experienced for many years. Namely, that a single customer might be quite “noisy,” with local fluctuations caused by the starts and stops of systems and equipment, while the city-wide load experiences almost no short-term fluctuations. In the same way, the power fluctuations at a single PV system location can be substantial, but fluctuations decrease as the footprint of distributed systems increases.