Turning Intermittency Into an Advantage

Changing perspectives on just about any subject can alter the understanding and many times the outcome of long held principles. For example, Human-kind's understanding of the way that diseases are transmitted is undergoing changes today as the Covid-19 pandemic is brought under control.  Another example is the way that the large growth of distributed electrical generation that has entered the PJM marketplace.  It can be viewed and integrated into the market either as a negative or as a beneficial plus.  The degree to which it is viewed positively or negatively depends a great deal on how the resource is packaged and presented.

Solar PV is such a resource.  It has entered the electrical generation market in large part as a distributed resource in Pennsylvania, with limits of 3 megawatts or less for nonresidential applications and at 50 kilowatts or less when behind the meter of a residence. On a one-on-one basis these systems don't impact the radar screen at the PJM (our regional power grid regulator) control center. However, when aggregated across the region, several hundred solar PV systems can easily account for many megawatts of clean, carbon neutral, renewable generation. While those attributes are undoubtedly positive, the growth in this resource has also raised concern about intermittency at the grid operator level.  

Intermittency of solar PV resources, or the fact that system output fluctuates with the availability of the sun, has labeled this resource as intermittent and as such of low value from a generation capacity perspective. However, the intermittency of Solar PV has some characteristics that may allow it to be valued at a higher level by the regional grid operator.  Thus far, these characteristics have not been incorporated into grid management and valuation models.  

As an aggregate group, on well over a hundred days of the year, solar PV systems in Pennsylvania offset the need to burn large amounts of fossil fuel annually. Each independent system also provides an instantaneous measure of the amount of sunlight available for conversion to electricity based on their location within a region. Cumulatively aggregating the solar energy available would create a predictive tool for tracking the movement of cloud cover across a region. This information would enable the tool to measure the aggregate electrical production output across the region and predict any changes to the output as cloud formations track across the region, lowering output across areas that are reported and recorded as the changing electrical output from solar PV systems large and small. This can all be collected reported and logged on the internet, instantaneously.  One plus to this approach would be that the readings are of actual power production, rather than satellite images or other data that would need to be converted to an approximate power production value.  

As the collection grows of these daily changes in solar insolation reaching the earth's surface, a library of data will accumulate that can be used to predict electrical production from solar PV resources across the region. Using weather forecasts, the tool can select a "proxy day" by choosing from the library a historic day that closely match the forecast. Using the proxy day as the base forecast, the electrical output from the aggregate group could be scheduled into PJM's day ahead forecast. As new systems join the aggregate group the new data points will reduce the granularity of information and the accuracy of the predictive tool will increase as will the capacity of the resource to be scheduled.

Ultimately the aggregated set of data in a region will demonstrate that the distributed solar power systems have gained reliability as a generation resource that can be scheduled into the PJM generation queue. In real time, the system will be outputting aggregate generation by Locational Marginal Pricing (LMP) node in the PJM model. As such these resources will become eligible for capacity payments based on the reliability factor they establish over time as the aggregation group operates and accumulates performance data.

Treating aggregated distributed solar PV resources as a generation source while logging production data to form a historic record for use in day ahead scheduling is a very important first step in overcoming the solar PV intermittency challenge. In the next of this series of articles, we'll look at filling the voids and balancing the output to improve the solar PV reliability profile, further increasing its value in the marketplace.

Prepared by Edward Johnstonbaugh, Penn State Extension