Performance of wholesale electricity markets with high wind penetration

Electricity markets around the world have added significant Variable Renewable Generation (VRG) in the form of wind and solar. As well as having zero greenhouse gas emissions, these technologies have limited predictability, variable output and low operating costs. Some form of policy is also usually used to encourage investment in VRG, such as the various renewable-specific policies that are currently in place across Australia or our discontinued carbon price. However, the economic performance and investibility of wholesale electricity markets with high levels of VRG is not clear, particularly when we consider realistic levels of renewable generation uncertainty.

In a paper recently published in Energy Economics, MEI researchers examined the performance of a wholesale electricity market with increasing wind penetration. The conditions that determine investibility for new generation are identified, and the effectiveness of renewable-specific and carbon-based policies are examined.

Comparison is first made to the canonical result that competitive marginal pricing should induce investors to build the optimal mix of generation via market signals alone. This is the fundamental theoretical basis on which ‘energy-only’ markets, like Australia’s National Electricity Market (NEM), rely to send investment signals to generators to enter or leave the market. It is not obvious that this result will always hold in renewable-rich systems for several reasons.

To do this, the impact of either a VRG incentive and a carbon price is first considered in the generation planning problem. The resulting generation fleets are then subjected to unit commitment that considers lumpiness of investment and VRG forecast uncertainty, followed by conventional economic dispatch of the committed units.

For wind penetrations up to approximately 50% by annual energy, it is found that unit commitment and renewable generation forecast uncertainty do not cause significant departure from this canonical economic result if the generation fleet is ‘optimal’. Optimality in this sense is determined in the planning problem that did not feature unit commitment, and which allows thermal generators to be built or retired as greater renewable generation or abatement was mandated. In contrast, the wholesale market with a fixed thermal fleet experiences progressively lower rates of returns to thermal generators as the level of renewable generation and/or abatement increased. This is simply because of over-capacity, with subsidised new wind and lack of capacity exit depressing wholesale market prices.

These results have implications for planning studies and market design. First, they show that standard planning tools will only produce an investible generating fleet if unit commitment effects are minor and there is always fleet optimality. Both of these conditions are not necessarily the case in practice. Related to this, unit commitment and forecast uncertainty appear to start having a significant impact once wind penetrations exceed approximately 50% by annual energy. Around this point, differences between total system prices and costs start to emerge and some thermal generators start to experience low returns.

Please contact Michael Brear to find out more about this work.