With increasing amounts of data centre load now pushing demand forecasts into a new and demanding phase, the challenge is no longer simply building renewable generation and batteries. It is integrating them into an electricity system that must remain stable, reliable and affordable as its underlying technology changes and which will provide for a step change in system capacity post coal. As investment accelerates, and as it continues to accelerate over the rest of this decade, grid connection is a critical consideration for developers and investors, and a prime red flag for value of development projects. Yet one of the most important challenges is less visible than transmission constraints or connection queues: ensuring that the equipment connecting to the grid behaves as expected.

Matt Rennie, Executive Director, Rennie Advisory spoke with Mohammed Arif, a Lead Power Systems Engineer in Rennie’s Power Systems team, about why inverter modelling matters, how Australia’s connection process works and what developers can do to reduce the risk of costly delays.

A different kind of engineering challenge

The world of power system engineering has changed enormously over the last 20 years. While conventional generation has long benefited from decades of research and standardisation, with control systems well understood, and established models allowing engineers to predict how they will respond to disturbances, inverter-based resources are different. As the technology scales up, OEM models are still evolving, and as more new players enter the market, there is a commercial logic in manufacturers not always disclosing every parameter or element of their control logic.

Mohammed Arif has more than 17 years of power systems consulting experience, including 11 years with GE Energy Consulting working on conventional synchronous generation and control systems. His experience across multiple international markets gives him a useful perspective on how quickly inverter technology is evolving. As he explains, engineers cannot always simply inspect an inverter model to understand its behaviour. They may need to simulate it, identify problems and work with the original equipment manufacturer (OEM) to resolve them. As he describes it, the process is iterative. Engineers test the model, provide feedback to the OEM and repeat the exercise until the required performance is demonstrated. The control logic and the changes needed to meet technical requirements ultimately sit with the manufacturer. This creates an important dependency for developers: progress can depend as much on the OEM’s engineering capability and responsiveness as on the consultant’s expertise.

What the rules require

The National Electricity Rules (NER) establish the framework for connecting generation and storage to the National Electricity Market. Clause 5.3.4 governs connection applications, while clause 5.3.4A addresses negotiated access standards and clause 5.3.4B deals with system-strength mitigation requirements. Schedule 5.2 sets out technical performance requirements, including the standards that generating systems must meet. Clause S5.2.4 establishes obligations for providing power system data and models.

These are not simply paperwork requirements. They determine what a plant must demonstrate dynamically in a simulation environment before it can operate within the real power system. For this reason, AEMO and the connecting network service provider (NSP) require models that accurately represent the plant’s physical configuration, control systems and expected performance, requiring engineers to use simulation tools such as PSCAD and PSSE to assess the plant against its agreed performance standards and examine how it might affect the wider network. For network connection engineers, this plots a careful path of setting the model parameters, and updating and testing as the project progresses. Ultimately, while the work is complex, the aim is quite simple: to ensure that the assumptions used in power system studies remain consistent with the equipment ultimately installed.

From feasibility to proof

As Arif explains, in practice, the industry commonly describes the technical process using R0, R1 and R2 milestones. Before R0, feasibility studies examine the practicality of the proposed connection location, potential capacity, fault levels and relevant network contingencies. For hybrid projects, engineers also assess the interaction between different technologies, such as wind, solar and batteries. Once an OEM is selected, the R0 stage involves developing the plant models and undertaking the studies required by the NSP. The R1 stage involves further model refinement and demonstrating technical compliance through simulation. R2 takes place after installation, when testing verifies that the actual plant behaves consistently with its model. Under the NER’s commissioning and model-data requirements, this process supports the demonstration of compliance with the agreed performance standards. Commissioning, where the rubber hits the road in terms of the physical performance of the plant, involves staged increases in output, with testing and approval requirements at intermediate operating levels. While some solar projects, for example, progress through hold points at low output, 50 per cent and ultimately full capacity, if measured performance differs from the simulations, engineers must investigate and resolve the discrepancy before the relevant operating restrictions can be lifted. AEMO’s modelling requirements therefore provide for the final R2 model package and validation report to be submitted before the plant’s commercial operation. This is an important distinction. R0 and R1 develop and refine the model. R2 validates it against the physical plant’s performance.

Why doing more work upfront can save time and money

Most power system engineers work globally, and Rennie is no different, but Arif sees benefits in Australia’s process being more demanding than in some other markets. While developers must invest considerable engineering effort before their projects reach commercial operation, and model development can involve repeated iterations between consultants, OEMs, AEMO and NSPs, this upfront effort reduces uncertainty and cost at later stages. If a model has not been properly developed and tested, problems may only become apparent when the plant is installed and commissioned. At that point, resolving them can require changes to control settings, additional testing, further model submissions and repeated reviews. Not surprisingly, the financial consequences can be significant. A project may remain restricted to a lower commissioning hold point while engineers investigate discrepancies, meaning revenue can be deferred, contractors and engineering teams remain engaged, and investors face uncertainty over the timing of commercial operation. Arif sees R0 as the most challenging stage. Repeated rounds of review and resubmission can cause substantial delays and costs. But that front-loaded effort is the rationale for Australia’s approach. Refining the model through R0 and R1 identifies problems earlier, when they are generally less expensive and disruptive to resolve. R2 then validates the model against the installed plant’s performance. This style and depth of work will never eliminate commissioning delays, but it can reduce the risk of fundamental issues emerging at the most commercially sensitive point in a project.

Arif suggests two decisions as deserving particular attention. First, OEM selection should involve more than comparing equipment prices. Arif highlights the importance of a proven track record of OEM equipment in weak-grid conditions, where low short-circuit levels can make satisfactory inverter performance more challenging. A lower equipment price may offer little advantage if the technology subsequently creates significant connection difficulties.

Second, the grid consultant’s experience with the selected OEM matters. A consultant familiar with its models, engineering team and recurring technical issues is better positioned to anticipate problems and resolve them efficiently, which reduces costs and delays if done correctly. When studies take longer than expected, investors need a clear explanation of the problem, its likely consequences and the steps required to resolve it. Arif expects these issues may decline over time as inverter models become more standardised, as conventional generator models have. Until then, technical expertise, model quality and effective collaboration with OEMs remain essential.

Australia’s connection process requires considerable effort before a project reaches commercial operation. But that effort serves a wider purpose: giving system operators a more accurate understanding of the assets on which the electricity system depends.

For developers, grid connection engineering should not be treated as a compliance task at the end of development. It should inform technology selection, project design and investment decisions from the outset.

The time you spend proving the model is time you won’t spend explaining a hold point to your investors.

If your project is heading into R0, or towards commissioning and R2, talk to Arif and Rennie’s Power Systems team.