Some of the Metrics analysed in the Generator Report Card 2018

On 7th February 2019, AGL Energy provided it’s 2019 half-year results, including a note about the availability of its aggregated portfolio.  Given we were working on the Availability Metric at the time, we posted this view of availability at its three main thermal assets – Loy Yang A, Bayswater and Liddell.

This same availability data was translated into an unavailability metric (with reference to each unit’s Registered Capacity in each dispatch interval) to provide a trended view of aggregate unavailability across 48x operational coal units across the NEM as at 31st December 2018.

Some of what we did in the Generator Report Card was discussed in this article published on 14th June.

Back in July 2018, we posted this article exploring the surprise expressed by a range of people when confronted with significant changes in their MLF from one year to the next – which, if the loss factor had dropped, directly represented a significant drop in top-line spot revenue figures.

In his “Lessons from the trenches” article from September 2018, Jonathon Dyson notes (about MLFs) that:

“physics of the power system beats financial models every day of the week”

In March 2019 we again see surprise expressed when MLFs drop as a result of more generation crowding into oversupplied areas remote from the customer.  In response to this, we posted this article.

It’s for this reason that we are investing considerable time in the exploration of MLF for all power stations over a 20-year period.

This is one of the metrics that are useful in understanding which generators are more dependable than others in their output.  We flagged that units might be Off-Target for a number of reasons, including:
Reason 1 = they might trip
Reason 2 = they might fail to start
Reason 3 = (if Semi-Scheduled) their output might not match AEMO’s forecast in AWEFS/ASEFS
Reason 4 = on occasions, also, generators might disregard dispatch targets.
Reason 5 = a unit might also be Off-Target because it was required to be, for the provision of FCAS Raise/Lower services in that dispatch interval.

We also mentioned the tabulated list of “worst performers” in this article on 7th March.

Category 1 = “Well Formed”, applying a strict interpretation of the AER Guidelines

Category 2 = Not in Category 1, but “Well Formed”, if applying a looser interpretation of the AER Guidelines

Category 3 = “Not Well Formed”, even with a looser interpretation of the

We posted some preliminary analysis that suggested a non-trivial number of rebids still “Not Well Formed” .

In the Generator Report Card, we are taking this analysis further.

1)  It can help us understand changes in patterns of operations for the grid/market as a whole – as discussed in this article on 7th March about “the death of baseload and rise of cycling”.

2)  At a more micro level, as well, it can help us understand degradation patterns and trends – hence being able to understand for every single generator:
(a)  If changes in operational patterns provide evidence of operational problems already with us; and/or
(b)  If changes in operational patterns hint at remnant life degradation, and hence the possible advent of operational problems at some point in the future.

For clarity, note that the specific DUIDs affected will vary from one month to the next, depending on the specifics of the constraint equations that have bound.

It’s for this reason that we have invested time in exploring constraints and their affect on generators (at both macro and micro levels) within the Generator Report Card.

On 28th February, we published this “Intermediate Guide to how Prices are Set” from on 28th February, to help explain some of the complexities inherent in the price setting process – and gues author Allan O’Neil followed up with this additional conceptual explanation on 22nd March 2019.

Because of these complexities we are exploring several different methods for providing a longitudinal answer to the question that is being increasingly asked “who’s set the price?”

As the this article helps to explain, this is not as simple to answer as it might initially seem.

On February 13th the AEMO published its “Quarterly Energy Dynamics – Q4 2018” (the PDF is here).  In particular this report noted that:

“There was also a relatively high number of sudden generator trips when completed to recent quarters, particularly in Queensland and New South Wales (Figure 11). At this stage, however, Q4 2018 results are not indicative of a longer-term trend, with the number of sudden unit trips in 2018 consistent with results in recent years”

In this particular report, the AEMO defines a “sudden trip” as:

” Definition of sudden unit trip for this report: when a unit is generating above its minimum generation level and falls to zero within two DIs.”

We’re taking this definition and having a look at what the performance has been, using this metric, over the longer-term.

——

Note that identifying trips is not as easy as it might initially seem when scanning over 20 years of data for hundreds of DUIDs.  There are a number of different algorithms that might be coded to produce a result – but (to date) none have proven themselves to be of high enough accuracy for us to be perfectly comfortable.
1)  Using this particular method, for instance, it will flag some “false positives” where a unit was being taken offline for planned maintenance and ramped to 0MW over 2 dispatch intervals.
2)  It is for this reason that we’re looking at several different metrics (including Number of Starts and Off-Target), to try to draw a more complete picture.

In this particular article reviewing what happened on 24th January the Instantaneous Reserve Plant Margin (IRPM) metric that’s been featured in NEMwatch for a number of years was featured – and proved a very useful non-price indicator of the balance between supply and demand.

It’s for the same reason we’re trending distributions of IRPM* over time as a macro indicator of the balance between supply and demand to provide some context to the more detailed analysis.

*  In the Generator Report Card, we calculated this slightly differently, as a Dispatchable Reserve Margin (DRM).

In the Generator Report Card, we take a look at all types of generators, to see how high temperatures affect their ability to achieve peak output levels.

Judging by the responses that we received for our challenge question, there were quite a few people not thinking holistically about the temperature limitations of all types of plant.  On 26th June we posted this follow-on article to shed more light on which unit “could not stand the heat”.

In the Generator Report Card, we explore this further.

The presentation (and the article) featured some data presented in the Report Card about Capacity Factors achieved for solar farms in the NEM, with both:
1)  A longer-term average shown for the few stations that have operated for more than a year; and
2)  Part-year data for 2018 for those stations that only began operations in that year.