Weekend read: In the testing loop

From pv magazine Issue 10 – 2022

At 4:18 p.m. on Sept. 29, 2016, South Australia’s power went out. About 850,000 residents in the state lost electricity, some for several days. The chain of events which led to this staggering glitch was set off by tornadoes simultaneously damaging two major transmission lines. Disturbances on the network led wind farms in the mid-north to switch off – an unforeseen reaction that ultimately contributed to the blackout, investigations from Australia’s Energy Market Operator (AEMO) later confirmed.

The wind farms’ unexpected behaviour arose from a protection mechanism within their control settings, part of the complex software responsible for running the inverter-based plants. Alongside further stoking burning debate around renewables and power system security, the dramatic blackout led Australia’s market operator to impose a far more cautious regime in connecting renewable energy systems to the network.

Six years on, grid connection remains fraught. “The whole connection risk has become a topmost risk for developers and investors,” Rajesh Arora, the technical director of power and industrial at global infrastructure consultancy AECOM, tells pv magazine.

Alongside causing a mass exodus of contractors from the Australian market, delays in the process from initial connection approval to registration and commissioning have cost companies millions, perhaps billions, of dollars, dragged legal disputes all the way to the Supreme Court, and seen investor confidence in the sector plummet.

For Australia to meet its emissions reduction targets, having passed the Climate Change Bill 2022 in September, Arora says the grid connection process must be streamlined. He was part of a working group set up by AEMO and industry group, the Clean Energy Council, formed for precisely that purpose.

A key question raised was how project simulation models can be made to better align with actual onsite results, since every mismatch entails that project contractors take more time to tweak their models and rebook with market operators.

Mark Parker, co-founder and technical director of electrical engineering contractor EPEC Group, explains how delays wreck revenues. A project he worked on took less than a day to rectify a mismatch, but two months to confirm the solution with network operators. “It was a 100 MW solar farm, and at that time based on their power purchase agreements, they would be paid roughly $1,000 (USD 672) per megawatt per day. So they were essentially losing up to $100,000 of revenue per day,” says Parker. “It highlights just how important it is to avoid these delays. In probably every project we’ve dealt with, we’ve found issues with commissioning.”

Aggregated testing

In thermal generation projects, companies taking on the majority of the electrical system – think the GEs and Siemens of the world – might wear 70% to 80% of the project’s total costs. Therefore, it is clearly in their interest to test the full aggregated system before deploying it.

But the commercial model is markedly different in the realm of renewables. For a solar project, an inverter and Power Plant Controller (PPC) might be less than 10% of the total project cost, Parker says. While manufacturers test their own equipment, it hardly makes sense to test components within the aggregated system. Aside from the complexity, that type of testing just doesn’t fall into anyone’s scope. “Businesses don’t really operate in this field of tying it together,” Parker says. “So there’s still a gap.”

To plug that, the EPEC Group co-founder returned to his old stomping ground, the University of Queensland, to find senior lecturer Richard Yan, who is working on a hardware in the loop platform which had yet to find its practical application.

Parker proposed the idea of testing aggregated solar generation control systems, consisting primarily of the power plant controller and the power quality meter. “It’s the PPC and the meter communicating with the inverters, that’s where we’ve found most of the issues that delay the commissioning processes,” Parker says.

With Yan onboard, the industry and research team worked closely to develop a business case for a platform which can combine different brands of components into an aggregated system in the lab, testing them long before site works begin.

The platform aims to bridge the fickle space between software models developed for the first stage of the connection process, and the actual field commissioning tests of the later stages. “Previously we only had the software and the plants. Now we have the platform in the middle – it’s not pure software, it’s not pure hardware,” Yan tells pv magazine. “That’s the reason we call it ‘precommissioning platform’.”

This kind of aggregated system testing is already standard in the aviation and auto industries, says Yan, and his team firmly believe it will prove beneficial for renewable power systems too. EPEC Group, which has also been supplying equipment to Yan, is quick to point out the platform is not a research and development tool. “It’s about proving the generator control system is correct before it’s commissioned in the field,” Parker says. “We plug in the inputs, we get the outputs and make sure that matches our design.” Bart Sedgwick, EPEC’s general manager adds, “the main thing here is we’re trying to bring certainty to the connection process.”

The platform received $498,000 in funding from the government-backed Australian Renewable Energy Agency back in January. As of September, the preliminary testing platform has been completed – successfully bringing together solar project simulation models to interface with real PPC and power quality meter hardware and take real-time measurements.

The next stage of the project will see the university team standardise its aggregated control system testing procedures, before validating the results against real field data from the Woolooga Solar Farm in Queensland, which is currently in commissioning, hopefully demonstrating the pre-commissioning platform’s accuracy.

Should all go to plan, the university and EPEC Group will seek to commercialise the platform, targeting contracting companies and developers. It’s a bit too soon to say how much the platform would cost, but Yan estimates it would be in the order of AUD 1 million to AUD 2 million.

The bigger vision is to convince the Australian Energy Market Operator – a stakeholder partner in the project – to make the platform’s use standard practice in Australia.

Component performance

While the platform concept has been warmly welcomed by renewable developers, contractors and market operators, the reception among manufacturers has been mixed.

The reason for that is closed loop lab tests make visible component’s actual behaviour, Yan explains. In tests he’s already conducted between two tier one brands of power quality meters, there have been stark differences. “On paper, the meter doesn’t say it will have this type of behaviour but that’s the real behaviour.”

Even through the platform’s primary purpose is to test aggregated systems, it could also afford developers greater oversight in choosing hardware technologies because they could test components in aggregated systems before committing.

Cooling high stakes

The platform also allows in-depth testing of the power system’s reaction to faults – like, for instance, the types of fault that occurred on South Australia’s network in the lead up to the blackout. “This precommissioning platform gives opportunity to validate those types of issues in a lab environment where we can do a range of tests that we can’t do in reality,” Parker says. Without intensive lab testing, the crucial work of fault testing remains haphazard: “You never find out what’s the true behaviour,” says Yan.