The National Electricity Market, or NEM, is the electricity interconnection on the east coast of Australia. Ongoing growth of Variable Energy Resources (VERs) such as wind and PV has required a range of adaptations in both system operation and modelling, and is increasingly challenging the wholesale market design.
NEM demand ranges from 15-35 GW, and annual energy is approximately 200 TWh. Table 1 shows the capacity and energy mix.
However this doesn’t tell the whole story. VERs are unevenly spread across the 3000 mile long interconnection, and some of the interconnected sub-regions that form the NEM are already at world-leading VER penetration levels. 2019 has seen increasingly frequent multiple-hour periods of negative regional spot prices, and 4.8 GW of additional wind and utility PV is committed, with a further 41 GW proposed.
Much of the NEM coal fleet is expected to reach the end of its technical life within the next 10-15 years, with little private sector appetite for new coal development. Over the last several years gas has become an increasingly expensive fuel source.
The NEM wholesale energy market design is approximately 20 years old. It utilizes a 5-minute dispatch, and includes all generating resources above 30 MW in dispatch and congestion management. It has no capacity mechanism or formal day-ahead market.
While zero marginal cost VERs are increasingly economically displacing dispatchable synchronous generation within this market, some of this same synchronous generation is still required to provide grid strength to ensure the stability of networks with high levels of Inverter Based Resources (IBRs) such as wind, PV, HVDC or batteries.
As the NEM market design does not include centralized real-time commitment, or binding pre-commitment of generating units, the system operator has become increasingly reliant on interventions outside the market to ensure sufficient synchronous generation remains online.
A robust industry debate is now underway around wholesale market reform. This includes the appropriate role for real time price signals, how to recognise the full range of system services required in dispatch, including generation dispatchability itself, and the potential need for a new capacity reserve mechanism, to ensure financial viability of resources required to meet peak demand.
NEM market operation includes granular forecasting of VERs over a range of operational time frames. This forecasting function has historically sat with the system operator. However, recent changes allow and incentivise VER operators to self-forecast. To date, one operator has commenced self-forecasting, with others expected to follow.
In 2018 a Forecast Uncertainty Measure (FUM) was developed, providing a forward-looking assessment of the likely aggregate uncertainty in demand and VER output forecasts. This FUM now forms an important input into short term reserve assessment, and has become particularly important for decision making during extreme weather conditions and low reserve periods.
The NEM has used markets since 2001 to procure both primary and secondary frequency control reserves. Reserve procurement is incorporated into, and co-optimised with, the 5-minute energy market. Reserve requirements and their allocation to generation resources are updated every 5 minutes, considering current system conditions, including the largest loss of infeed risk, system demand, and system inertia.
VERs now successfully participate in these markets, providing both primary reserves, delivered automatically via local control, and secondary reserves, via centralised Automatic Generation Control (AGC). While the value of these reserve markets is low relative to the energy market, they form an important income source for some new resources, particularly grid-scale batteries.
However, the frequency performance of the NEM has declined significantly over the last several years. While there are multiple reasons, the technical requirements and economic incentives specified in these existing market arrangements have played a role. Rule changes affecting technical and commercial settings have recently been proposed to address this decline.
The limitations of traditional RMS power system modelling tools for assessing the stability of networks dominated by IBR have become increasing clear. Electromagnetic transient (EMT) modelling of significant sections of the NEM is now used to assess system performance and operating limits.
South Australia sits at one end of the NEM. It has a median demand of 1,380 MW, one AC interconnection to the rest of the NEM, and 1,930 MW of installed wind. In 2018/19, wind generation instantaneously peaked at 144% of regional demand, and exceeded 100% of regional demand approximately 5% of the time.
EMT models of this entire region have been used to assess the performance and interaction of multiple widely dispersed IBRs under lower grid strength conditions. This modelling indicates that 4-5 synchronous generators of 150-200 MVA each must remain online in South Australia at all times to ensure system stability.
All South Australian synchronous generation is thermal, and many larger units are relatively inflexible, with start-up times of several hours, and high minimum generation levels. This increases the operational challenges and economic impacts of managing unit commitment. None of the large generators can operate in synchronous condenser mode. The regional TSO is now procuring four synchronous condensers to reduce the cost of out-of-market generation commitment.
The island state of Tasmania is connected to the NEM via a 500 MW HVDC link. It has 308 MW of wind, a median demand of 1,220 MW, and is predominantly hydro-based. In 2015 it reached 79% instantaneous penetration of IBR (wind and HVDC imports), supported by operation of multiple hydro units in synchronous condenser mode.
With additional wind generation now under construction, wind and HVDC import capacity could exceed demand by early 2020. Minimum synchronous generation requirements to support adequate post-fault recovery and frequency response are currently being assessed, again using EMT modelling.
The blog post outlines just some of the responses to a changing generation mix in the NEM. As levels of VER rise further, the challenges of managing this large interconnection will continue to increase.
Australian Energy Market Operator