Editor’s Introduction: With another review of the Renewable Energy Target commencing we felt it was important to revisit the results of a modelling exercise assessing potential wind power grid integration technical issues undertaken by the Australian Energy Market Operator back in late 2013. This study attracted little attention but gave strikingly different answers to prior modelling exercises, suggesting greater grid integration costs for levels of wind consistent with achieving the Renewable Energy Target. While this study was fine for its purpose of helping AEMO to explore potential technical changes that might be required to manage high levels of wind penetration, it made simplifying assumptions that made it unsuitable for assessing the likely economic costs of achieving the Renewable Energy Target. To head off the potential for this study to be misinterpreted and misused in the forthcoming review of the Renewable Energy Target, we asked Jenny Riesz to provide this review of the report.
AEMO’s Wind Integration Studies report, released in late 2013, suggests that technical constraints and grid limitations could lead to the significant curtailment by 2020 of around 35 per cent of the wind energy generated in Victoria, and around 15 per cent of the wind energy generated in South Australia.
Have other studies failed to capture the impact of grid constraints that mean meeting the 41,000 GWh Renewable Energy Target will be much more expensive than we thought?
Compared to previous modelling studies of the RET, the Australian Energy Market Operator study has two key differences that lead to this result:
AEMO’s analysis takes a more careful look at the grid constraints that are likely to apply in 2020, with around 11.5 GW of wind operating.
AEMO’s study doesn’t take account of the fact that wind developers are likely to shift development away from congested network areas when these constraints begin to become important.
AEMO has been working on various issues related to the integration of wind for over a decade. This study builds upon this body of work, and provides a more rigorous assessment of grid constraints than has been attempted previously.
Most market modelling studies use a representation of network limitations, often applying the constraint equations released annually by AEMO with the National Transmission Network Development Plan. These equations define a wide range of grid limitations, including intra- and inter-regional thermal line limits, transient, voltage and oscillatory stability limits, and limits relating to Frequency Control Ancillary Services.
However, the equations aren’t complete; they represent a reduced set of the most important system normal constraints used by the National Electricity Market Dispatch Engine in the actual dispatch process.
AEMO’s NTNDP constraint equations arguably provide a decent approximation of the present network, operating in its current form. However, they don’t capture many of the important limitations of the grid that will become apparent as the generation mix changes. AEMO’s study made a first attempt to project how these grid constraints might change as more wind generation enters the market – this has undoubtedly contributed to the different curtailment outcomes compared with earlier modelling studies.
But this isn’t the full story. Importantly, the AEMO studydid not attempt to optimise the locations of wind installed in the NEM, based on updated knowledge of grid constraints.The study used a distribution of new wind development based on a number of factors, including the wind resource and the announced intentions of wind developers. Some consideration of known grid limitations was taken into account, but this didn’t include the new constraints implemented in this study. This modelling therefore explores the “worst case” of what might happen if wind developers blindly install their projects, with no consideration for mounting grid congestion.
This led to the wind distribution illustrated in Figure 1. In this scenario wind development is highly concentrated in the regions with the smallest electricity demand (South Australia, Tasmania and Victoria) with much lower proportions entering in New South Wales and Queensland, the two highest demand regions (based on projections for 2020). Thus, South Australia reaches a wind penetration level of almost 70 per cent, and Tasmania almost 50 per cent. No wonder there is significant wind curtailment! By concentrating the wind into the smallest demand regions, it is only to be expected that we see growing pressure on the power system.
Figure 1: Distribution of wind development assumed in AEMO study for 2020
In reality, wind developers are likely to respond to these network constraints, and move wind projects to less congested parts of the network in NSW and Queensland. In our system, what matters is ultimately the amount of generation you can export to market, so curtailment translates directly into lost revenue. In this scenario, for example, the heavy curtailment of wind farms in Victoria means they are only achieving capacity factors of 24 per cent, while much better capacity factors of between 34 per cent and 37 per cent are available from alternative sites in NSW and Queensland.
Furthermore, constraints on interconnectors mean more frequent price separation between regions, and wind farms in South Australia are already receiving lower average wholesale prices than elsewhere in the grid. Astute wind developers are well aware of this, and will carefully examine likely future revenues taking grid constraints into account.
Once network congestion starts to become a more serious reality, it will become commonplace for new wind projects to undertake even more detailed network constraint studies, and the findings will directly impact their choices of where they locate. In the event that an unlucky initial few proceed without undertaking sufficiently detailed network congestion analysis, they could find the results to be catastrophic to their bottom line. Everyone will learn from that hard lesson.
So what does this study really show?
For wind developers, it highlights the growing need to pay close attention to network congestion in any new projects under consideration. Wind developers will ignore this at their peril – the honeymoon period of excess network availability in South Australia and Victoria is coming to an end, and it’s possible that changing control regimes are imminent which may help or hinder new wind projects. Of course, some level of curtailment may be economically efficient, especially when accessing a particularly good wind resource. Precise analysis to determine acceptable levels will become essential.
For AEMO, it means that the market is rapidly changing, and the impacts could be highly material, depending on what market participants decide to do. It’s therefore important that this detailed analysis continues and the constraint equations and other aspects of market operation are adapted in a timely manner.
For policy makers, it certainly doesn’t suggest that the 41,000 GWh RET can’t be met, but it does highlight that it’s non-trivial to transition our complex electricity system to low carbon generation sources. The important work of AEMO and others to understand how we need to adapt our market and operational systems needs to be supported and carefully understood. Doing this work in a timely manner will ensure that we transition smoothly and without high excess costs.
Dr Jenny Rieszis a research associate at the Centre for Energy and Environmental Markets at the University of New South Wales.
Editor’s Introduction: With another review of the Renewable Energy Target commencing we felt it was important to revisit the results of a modelling exercise assessing potential wind power grid integration technical issues undertaken by the Australian Energy Market Operator back in late 2013. This study attracted little attention but gave strikingly different answers to prior modelling exercises, suggesting greater grid integration costs for levels of wind consistent with achieving the Renewable Energy Target. While this study was fine for its purpose of helping AEMO to explore potential technical changes that might be required to manage high levels of wind penetration, it made simplifying assumptions that made it unsuitable for assessing the likely economic costs of achieving the Renewable Energy Target. To head off the potential for this study to be misinterpreted and misused in the forthcoming review of the Renewable Energy Target, we asked Jenny Riesz to provide this review of the report.
AEMO’s Wind Integration Studies report, released in late 2013, suggests that technical constraints and grid limitations could lead to the significant curtailment by 2020 of around 35 per cent of the wind energy generated in Victoria, and around 15 per cent of the wind energy generated in South Australia.
Have other studies failed to capture the impact of grid constraints that mean meeting the 41,000 GWh Renewable Energy Target will be much more expensive than we thought?
Compared to previous modelling studies of the RET, the Australian Energy Market Operator study has two key differences that lead to this result:
AEMO has been working on various issues related to the integration of wind for over a decade. This study builds upon this body of work, and provides a more rigorous assessment of grid constraints than has been attempted previously.
Most market modelling studies use a representation of network limitations, often applying the constraint equations released annually by AEMO with the National Transmission Network Development Plan. These equations define a wide range of grid limitations, including intra- and inter-regional thermal line limits, transient, voltage and oscillatory stability limits, and limits relating to Frequency Control Ancillary Services.
However, the equations aren’t complete; they represent a reduced set of the most important system normal constraints used by the National Electricity Market Dispatch Engine in the actual dispatch process.
AEMO’s NTNDP constraint equations arguably provide a decent approximation of the present network, operating in its current form. However, they don’t capture many of the important limitations of the grid that will become apparent as the generation mix changes. AEMO’s study made a first attempt to project how these grid constraints might change as more wind generation enters the market – this has undoubtedly contributed to the different curtailment outcomes compared with earlier modelling studies.
But this isn’t the full story. Importantly, the AEMO study did not attempt to optimise the locations of wind installed in the NEM, based on updated knowledge of grid constraints. The study used a distribution of new wind development based on a number of factors, including the wind resource and the announced intentions of wind developers. Some consideration of known grid limitations was taken into account, but this didn’t include the new constraints implemented in this study. This modelling therefore explores the “worst case” of what might happen if wind developers blindly install their projects, with no consideration for mounting grid congestion.
This led to the wind distribution illustrated in Figure 1. In this scenario wind development is highly concentrated in the regions with the smallest electricity demand (South Australia, Tasmania and Victoria) with much lower proportions entering in New South Wales and Queensland, the two highest demand regions (based on projections for 2020). Thus, South Australia reaches a wind penetration level of almost 70 per cent, and Tasmania almost 50 per cent. No wonder there is significant wind curtailment! By concentrating the wind into the smallest demand regions, it is only to be expected that we see growing pressure on the power system.
Figure 1: Distribution of wind development assumed in AEMO study for 2020
In reality, wind developers are likely to respond to these network constraints, and move wind projects to less congested parts of the network in NSW and Queensland. In our system, what matters is ultimately the amount of generation you can export to market, so curtailment translates directly into lost revenue. In this scenario, for example, the heavy curtailment of wind farms in Victoria means they are only achieving capacity factors of 24 per cent, while much better capacity factors of between 34 per cent and 37 per cent are available from alternative sites in NSW and Queensland.
Furthermore, constraints on interconnectors mean more frequent price separation between regions, and wind farms in South Australia are already receiving lower average wholesale prices than elsewhere in the grid. Astute wind developers are well aware of this, and will carefully examine likely future revenues taking grid constraints into account.
Once network congestion starts to become a more serious reality, it will become commonplace for new wind projects to undertake even more detailed network constraint studies, and the findings will directly impact their choices of where they locate. In the event that an unlucky initial few proceed without undertaking sufficiently detailed network congestion analysis, they could find the results to be catastrophic to their bottom line. Everyone will learn from that hard lesson.
So what does this study really show?
For wind developers, it highlights the growing need to pay close attention to network congestion in any new projects under consideration. Wind developers will ignore this at their peril – the honeymoon period of excess network availability in South Australia and Victoria is coming to an end, and it’s possible that changing control regimes are imminent which may help or hinder new wind projects. Of course, some level of curtailment may be economically efficient, especially when accessing a particularly good wind resource. Precise analysis to determine acceptable levels will become essential.
For AEMO, it means that the market is rapidly changing, and the impacts could be highly material, depending on what market participants decide to do. It’s therefore important that this detailed analysis continues and the constraint equations and other aspects of market operation are adapted in a timely manner.
For policy makers, it certainly doesn’t suggest that the 41,000 GWh RET can’t be met, but it does highlight that it’s non-trivial to transition our complex electricity system to low carbon generation sources. The important work of AEMO and others to understand how we need to adapt our market and operational systems needs to be supported and carefully understood. Doing this work in a timely manner will ensure that we transition smoothly and without high excess costs.
Dr Jenny Riesz is a research associate at the Centre for Energy and Environmental Markets at the University of New South Wales.