Hosting capacity analyses are the first step for utilities seeking to create a holistic approach to incorporating distributed energy resources.
A new Electric Power Research Institute (EPRI) report released this week offers a very useful review of the latest methods and metrics of hosting capacity analyses in some key U.S. markets, including California, New York and Minnesota. Each state has pushed utilities to undertake a version of an HCA, using different data inputs and methodologies, but aimed at a set of common goals.
In the most restrictive sense, utilities need to know the limits of individual circuits and subsections, to give developers and customers the data to understand why they can, or can’t, bring new DERs online without adversely impacting power quality or reliability and requiring infrastructure upgrades
But HCAs also serve a positive role of adding data and streamlining processes to assist developers with interconnections, as well as to allow utilities to better plan DERs into their current, and future, grid.
"In the future, this enhanced level of analysis will enable utilities to determine the ability of the distribution system to utilize services from DER non-wires alternatives, the impacts of DER on grid reconfiguration, operational strategies, and smart inverter technologies,” EPRI wrote.
But to serve these needs, HCAs will need to demonstrate their ability to serve across multiple, sometimes contradictory, measures. First, it requires granularity of data on location, feeder design and operation, and DER technology. Second, HCAs must be scalable to entire distribution systems, repeatable to consider individual feeder modifications, transparent and proven to gain confidence, available to the industry at large, and consistent across various jurisdictions and technology implementations. That can conflict with getting the most granular data possible.
As EPRI’s report points out, there are many complex and divergent ways to collect, measure, weigh and calculate the data that goes into HCAs. These tools have been organized into different categories, such as “stochastic,” “iterative,” “streamlined” and “hybrid” — but these terms mean different things in different contexts, EPRI notes. And because they’re all still in development, various efforts to compare one method to another have yielded unclear results.
The stochastic, streamlined and iterative methods
Stochastic HCAs are the earliest attempt to grasp the problem. EPRI defines “a stochastic analysis" as when a "DER is added at multiple locations across the feeder representing a future planning scenario when there could be a distribution of DER on the feeder.” But because it doesn’t allow for plugging in real-world DER values, EPRI deems it “an effective method to be used as a research tool, but it is not recommended for broad application beyond that. The approach is used for "select R&D projects where direct modeling of DERs is needed on only select feeders.”
California has driven development of two more methods: streamlined and iterative. These are the two methods developed by the state’s investor-owned utilities to provide an Integration Capacity Analysis (ICA) — California’s term for an HCA — for the state’s Distribution Resource Plan proceeding.
PG&E’s approach was a so-called “streamlined” method, aimed at reducing cost and time. The advantages are its ability to manage time-based analyses and run “what-if” scenarios such as DER forecasts, reconfiguration, smart inverter settings, and DER mitigation strategies.
But PG&E’s streamlined method also sacrificed accuracy in capturing some of the dynamic effects on more complex circuits, EPRI noted. It also only considers single-site DERs, and “does not currently consider the aggregate impacts of distributed DERs (e.g., rooftop PV) needed when planning for future DER scenarios.” And for these reasons, it’s unclear whether its data can be used to streamline interconnection processes.
Southern California Edison and San Diego Gas & Electric chose a distinct approach, dubbed the “iterative” method, meant to yield much more specific findings. This method uses distribution planning tools such as CYME and Synergi to perform the voltage and thermal impact assessments rather than utilizing a calculation-based approach. In simple terms, the iterative method “essentially uses power flow simulations on modeled circuits, then increases the DER at each node until a violation occurs."
This is a very computation-intensive approach compared to the streamlined method, EPRI notes. It takes 27 hours to perform calculations on one circuit, compared to about 10 minutes per circuit for the streamlined method. That makes it impractical for entire systems.
But unlike the streamlined method, the iterative method allows for “an agnostic hosting capacity result,” EPRI wrote, “a unique aspect of the ICA approach that enables the rigorous hosting capacity assessments to be performed upfront while allowing the actual DER-specific results to then be derived offline.”
This method should also, in theory, yield results that can be used for interconnection purposes, EPRI noted. Unfortunately, the heavy load it places on distribution modeling software means that real-world implementations have had to take shortcuts that undermine that level of accuracy, whether in the more complex protection-based hosting analysis, or in using “search and find” routines to determine the hosting capacity result rather than incrementally stepping through DER penetration levels.
And each utility’s implementation of the iterative method have yielded varying results, EPRI noted. For example, all three California IOUs applied it to their CA DRP DEMO A/B projects, but the CYME and Synergi platforms yielded quite different short-circuit impedance and voltage profiles.
These findings underscore a key point EPRI makes — that utilities will always have to make tradeoffs in the HCA process. “Assumptions are necessary due to the uncertainties in underlying data and therefore a single '100% accurate' answer is not achievable. Assumptions around voltage regulation, future load profiles, DER profiles and characteristics, phasing, etc. all impact hosting capacity.”
The DRIVE hybrid method
EPRI’s Distribution Resource Integration and Value Estimation (DRIVE) hosting capacity method is aimed specifically at balancing the need to keep computation burdens within practical limits with capturing the most critical grid responses required to ensure that hosting capacity follows the modeling.
It combined parts of PG&E’s streamlined method, such as using a select number of power flow cases to characterize the feeder response, then performing calculations for “DER scenario impacts and hosting capacities.” But it’s “also similar to the iterative ICA method” in that the tool uses protection analysis to make it more efficient.
DRIVE also combines both iterative calculations for single-site DERs and stochastic analyses for multiple-site DERs, and it can manage what-if analyses.
As EPRI points out, it’s by far the most widely used method to date, with work underway by more than 50 utilities, the U.S. Department of Energy, the California Public Utilities Commission and the New York State Energy Research & Development Authority, and planning platforms including CYME, Synergi, Milsoft, OpenDSS, Gridlab-D, DEW, PVL and PowerFactory now supporting its implementations.
Still, EPRI’s hybrid method does have its disadvantages. “The initial implementation, while effective, was found to provide conservative results in some cases,” it wrote. And while “the results are still useful in informing interconnection processes,” it’s still “different than that traditionally used for detailed interconnection studies.” The current version doesn’t support DER portfolios, since it uses specific DER characteristics for its calculations. It also analyzes one feeder at a time and doesn’t yet calculate substation impacts.
Findings and next steps
To date, these different HCA methods have had a limited opportunity to be tested on their own, let alone in comparison with one another. EPRI’s report laid out three such comparison efforts so far, two within California, and one internal, that largely prove that it’s too early to try to compare such nascent technologies.
For example, all three California investor-owned utilities compared their chosen methods to each other during their DRP Demo A projects, but “at the time of comparison, both methods were still in various stages of development,” with the iterative method further along than the streamlined one.
But the results were inconclusive, showing both that the same methods could yield different results, and that different methods could yield largely similar results — all depending on assumptions built into different calculations.
SCE and SDG&E’s implementations of the iterative method yielded different results, because they used separate software vendors that performed calculations differently. Likewise, when SCE and SDG&E tried out their own versions of PG&E’s streamlined process, each found some variations in implementation that also impacted results. At the same time, different methods ended up yielding similar results, as with SDG&E’s work with EPRI to compare DRIVE to the iterative ICA method.
One of EPRI’s key findings was that “mandating how hosting capacity should be calculated sets the industry up for costly risks.” The report cites as an example California, where regulators chose the iterative approach for its greater accuracy. But its computational intensity has forced utilities to rely on alternative methods for some calculations, which “may lead to unnecessarily limited hosting capacity in some cases.”
"While the industry has tried to draw a distinct line between various approaches, mainly 'iterative' and 'streamlined,' in the future this will be irrelevant as there will likely be little means of distinction,” EPRI wrote.
Methods are important, but the results are what matter most. Providing transparency is more important than the underlying algorithms used. “EPRI recommends developing and publishing results from test feeders, thus allowing the industry to compare and validate results consistently as methods continue to develop.”
That’s going to become critical, as the distribution system becomes ever more complex. “Grid modernization initiatives, using DERs as non-wires solutions, transactive energy — all of these changes will increase the complexity of distribution analysis,” the report noted. “Hosting capacity analytics are a key component in the assessment of distribution systems, and as such the industry should continue to focus on improving the methods outlined here. Advancements to the capabilities of hosting capacity will be critical to ensuring the results of this analysis capture the needs of tomorrow.”
