The Final AEMO Report on the big-SA Blackout deals up some hard truths, and contradicts its earlier claim that the “energy mix” didn’t matter. The key theme here is about the system inertia. The Blackout on Sept 28 last year was an accident waiting to happen, and it wasn’t storm damage to lines that caused it. The blackout would not have happened if wind power had not been so dominant.
The transition to a 35% wind powered system left the SA grid very vulnerable. On Sept 28 last year, the safety settings on wind turbines were overly sensitive and when voltages “bumped” the turbines shut off suddenly, but those shutoffs hit the system too fast, and that caused the interconnector to shut off too, sacrificing SA to protect the rest of the national grid. The settings themselves are not the main issue — because they can be changed to prevent a repeat. It is a fixable problem — what is harder to fix, is the lack of inertia, and the sheer complexity. These are the biggest challenges of any renewables grid. We can fix even those problems, but at what cost in order to change the weather 100 years from now?
The AEMO report mentions that the “intermittency” of wind power was not a problem, but that’s a strawman. It is true on the blistering scale that a grid crisis unfolds at, that intermittency is not an issue. (The graph in Figure 14 covers just five hundreths of a second). But on a longer scale the-intermittency-problem set the scene. Intermittency is a problem for the pricing, the market, and in the long run, it, and the subsidies to compensate for it, was what made cheap coal powered stations unviable.
The Final Report “Black System South Australia” 273p
External Review on the Final Report
The key difference was the lack of synchronous generators
The Heywood interconnector has broken before but this time SA didn’t have enough “inertia”. (Credit to the ABC for doing a decent job on this):
AEMO said unforeseen separation and complete loss of the Heywood Interconnector has occurred six times in the past 17 years.
But in every other instance, the system stayed alive.
“The key differentiator between the 28 September 2016 event and the other three events is that there was significantly lower inertia in SA in the most recent event, due to a lower number of on-line synchronous generators,” the report said.
“This resulted in a substantially faster rate of change of frequency compared to the other events, exceeding the ability of the under-frequency load-shedding scheme to arrest the frequency fall before it dropped below 47Hz.”
The independent review blames synchronous inertia:
– The system inertia on the SA side was not sufficient to maintain the frequency drop (once the Haywood interconnector tripped) and to make the under frequency load shedding (UFLS) effective. This is a key point. This is illustrative from Table 11.
The Heywood Interconnector has suffered a “complete loss” before. This time the system had more asynchronous (i.e. mostly renewable) generators.
The state with “must run” renewables will need to have “must run” thermal generation:
– The system inertia requirements must be carefully evaluated to ensure system stability in the event of extreme disturbances as the one that led to the black system. This is identified in the report. ‘Must run’ thermal generation may have to be identified. Synchronous condensers may be investigated as a potential solution if the thermal generation dispatch is expected to be low under specific load conditions. Generation mix is identified mentioned throughout the report as an important consideration from a power system security perspective.
More renewables means the whole grid system needs to be redesigned:
System Inertia: As systems lose inertia due to the retirement/displacement of thermal plants and the expansion of wind farms, frequency control becomes more important as frequency control is directly related to inertia. This report was very wise to note the relatively low level of inertia during this disturbance. The whole topic of frequency control and RoCoF [Rate of Change of Frequency] should be examined on a system wide basis and all devices that protect frequency should be reviewed for their appropriateness. The paper correctly states that the governor’s settings did not play an impact on the outcome of this disturbance. However, if it is impossible to consider under frequency schemes with higher activation frequencies due to frequency regulation, then this could be a sign that governor dead band needs to be reviewed system wide.
From the AEMO Executive Summary
Previously grid managers could rely on synchronous generators (coal, gas, nukes, hydro) to provide services like “voltage control, frequency control, inertia, and system strength.” Now (with more asynchronous renewables) we have to redevelop our entire grid, plan ahead, pay extra, and figure out how to get these services:
The generation mix now includes increased amounts of non-synchronous and inverter-connected plant. This generation has different characteristics to conventional plant, and uses active control systems, or complex software, to ride through disturbances. With less synchronous generation online, the system is experiencing more periods with low inertia and low available fault levels, so AEMO is working with industry on ways to use the capability of these new types of power generation to build resilience to extreme events. As the generation mix continues to change across the NEM, it is no longer appropriate to rely solely on synchronous generators to provide essential non-energy system services (such as voltage control, frequency control, inertia, and system strength). Instead, additional means of procuring these services must be considered, from non-synchronous generators (where it is technically feasible), or from network or non-network services (such as demand response and synchronous condensers).
How low are we aiming – apparently we need more inertia but only so we have time to drop enough people off the grid:
The following factors must be addressed to increase the prospects of forming a stable SA island and avoiding a Black System: Sufficient inertia to slow down the rate of change of frequency and enable automatic load shedding to stabilise the island system in the first few seconds. This will require increases in SA inertia under some conditions, as well as improvements to load shedding systems combined with reduced interconnector flows under certain conditions.
They could have said, we need more inertia so that South Australians have reliable electricity. But they didn’t.
Your future is load shedding. Get used to it.
h/t David B