On the Advanced Solar + Storage discussion board, Wes Kennedy answers student questions about distribution circuit capacity and AC coupling for retrofits…

Student 1: Hey Wes, I have a quick comment on defining penetration levels. I think the definition you provided is a solid measurement, but as you know, utilities (and specifically those in Hawaii) have defined penetration by other factors. Initially, penetration was defined as a ratio of peak generation capacity (of all PV on a circuit) to the peak annual load (a baseline they were pulling from anytime in years past). So a 15% penetration meant that on a circuit which, at some point in the last 20 years, had a peak power load of 1MW, there was 150kW of PV (nameplate rating of inverters combined primarily, as PTC was almost never lower than AC ratings).

A couple of years ago they moved to a complementary metric of comparing peak generating capacity to daytime minimum load (DML). So a circuit with a DML of 150kw and 150kw of PV was considered 100% penetration for DML. Over the last two years, the position on how generating capacity relates to DML and its effects on the circuit under extreme conditions has evolved, and last year utilities determined they could handle up to 250% penetration, DML, without significant adverse effects. Of course, some circuits received certain upgrades to accommodate.

Under the PUC’s ruling last month (the one that did away with generation), they reiterated the utility’s proposal of moving to a new complementary metric that can basically be described as access capacity. Basically, this means they can individually define how much generation and under what conditions a circuit can operate, and they can decide what level of export the circuit can handle. It sounds like a much more intelligent way to approach the matter, but it also puts the judgement almost entirely in the utility’s hands without little for accountability measures.

What are you seeing elsewhere in terms of determining reasonable levels of capacity on distribution circuits? What metrics do you think make the most sense for allowing the appropriate levels of capacity while retaining accountability measures? Thanks!

Wes Kennedy: Hi Student 1. I just so happened to attend a conference at NREL a couple weeks ago on just this topic. The 250% DML penetration was derived from modeling done at NREL, specifically looking at issues around TOV (Transient Over Voltage).

Modern grid tie inverters are able to reduce to zero volts in a very few cycles, even in extreme cases like 1000% DML and load drop suddenly to zero. So the TOV scare was essentially proven to be a boogey man, and HECO, etc. (to their credit) changed policy because of it.

This conference was about the very points you bring up in your question. How do you determine safe penetration at the feeder level? The truth is, there is no rule of thumb, because more than a dozen factors affect the decision. Utilities are getting smarter and more comfortable running advanced grid simulations to determine that value. Even then, the advanced grid features built into modern inverters can alleviate much of the concern with high penetration. The latest concern is high line voltage on feeders. Playing with pf, small lags of .98 can alleviate much of the issue. HI is considering mandating a .98 fixed on all inverters after a certain date. Dynamic pf is the next step.


Student 2: For AC Coupling it would seem that the grid forming inverter would have to be able to form grid conditions within the parameters established for the grid tied inverter in order to bring the PV generation back online, in the event that the systems disconnects from the grid. Is it common for these grid forming inverters to be able to accomplish this, especially since most grid tie inverters (synchronous) have different settings by region?

Student 3: Two of the most established grid forming inverters for smaller microgrids are the SMA Sunny Island and the Schneider XW+. They both can work with AC Coupled “standard” inverters, but those inverters need to have an option to allow this (backup mode on Sunny Boys), and control of their operation is then done via frequency shifting. Wes could certainly weigh in more authoritatively on this.

Larger units like those from Princeton Power are grid-forming but really need to work more with their own technology. The 30kw GTIB from Princeton can work as a bidirectional battery inverter-converter, or as a direct coupled PV inverter-charger. I’m not sure how well third party grid-interactive inverters would work in AC coupled mode.

Student 2: Thanks, Student 3!

I think this is perhaps my biggest question right now, as it has implications both for current projects under design as well as providing potential retrofit options for our clients. For example, say we have a client with an existing Enphase system and now want to look into an AC coupled system that could serve critical loads in the event the grid went down. It seems the concept of AC coupling is to bring in a second, grid forming inverter that can bring the grid tie system back online when the utility grid is down. However, from what you are saying, it sounds like the inverters available to achieve AC coupling in a retrofit situation are pretty limited.

Wes, can you provide any further insight on this?

Wes Kennedy: Hi Gang! AC coupling makes a lot of sense for retrofits. Any of the inverter chargers will create a stable enough grid to get a grid tie inverter to sync to. Now the question arrises with controls. What happens when pv production exceeds loads (including charging the battery)? The SMA Sunny Island raises frequency of the islanded grid. SMA Sunny Boys can be put in a back up mode where by they respond to the change in frequency as a command to move off the mpp, throttle production in proportion to the load… very slick design. Other inverters see the raise in frequency as a bad grid and kick off line. these creates a very crude on/off controller. This could, potentially, happen dozens of times during a sunny daytime outage. Because of this, many pv inverters were reluctant to allow ac coupling with Sunny Islands for fear of premature failure of the inverter.

Other inverter chargers rely on other methods to regulate over charging. I believe Schnieder and Outback both rely on voltage trip relays that will drop an inverter off line when dc bus voltage gets too high. SolarPro Magazine did a great article a couple years back comparing all the ac coupling schemes for the various players. Read it here.

Ready to spend 9-weeks gaining an advanced technical understanding of residential, commercial and utility-scale solar-plus-storage systems? Want expert insight on policy, financial models, and emerging markets and opportunities? Enroll in Wes Kennedy‘s Advanced Solar + Storage course today and earn 30 NABCEP Advanced Credit Hours6 NABCEP Technical/JTA3 NABCEP NEC6 NABCEP Elective in the process. 

Download Wes’ FREE Battery Selector Tool, used in Advanced Solar + Storage:

 Person medium 2918b43

About Instuctor Wes Kennedy

Kennedy was most recently the senior application engineer for SMA America’s Hybrid Energy Solutions group. In this position, project scope ranged from residential PV with battery backup to MW-scale stand-alone microgrids primarily in North America and the Caribbean. He has worked in the solar industry since 1996, when virtually all projects were battery-based. He began his career with pioneering solar companies Jade Mountain and Real Goods, cofounded Colorado EPC firm Namaste Solar, and managed the engineering staffs at groSolar and Abound Solar. His skill sets include engineering, design, training, education, installation, O&M, software modeling, sales, marketing and management. Kennedy currently resides in Boulder County, Colorado.

Wes is teaching: