On Monday 6th July 2026 we noted ‘TransGrid directed to (temporarily) disconnect 3 batteries for System Strength’, and then followed that up with ‘Three questions, in relation to the 3 x NSW batteries directed off for System Strength reasons’.

… there were a couple questions in there:

(a)  one of them revolving around how often it is that AEMO needs to direct a TNSP to disconnect any unit (for system strength, or other reasons)?

(b)  there were separate questions about disconnecting batteries for system strength reasons.

Yesterday (Wednesday 15th July 2026) we saw a somewhat similar case, via MN144556 at 12:55:34 (NEM time) with respect to the Lilyvale Solar Farm:

‘——————————————————————-
MARKET NOTICE
——————————————————————-

From :              AEMO
To   :              NEMITWEB1
Creation Date :     15/07/2026     12:55:34

——————————————————————-

Notice ID               :         144556
Notice Type ID          :         MARKET INTERVENTION
Notice Type Description :         Reserve Contract / Direction / Instruction
Issue Date              :         15/07/2026
External Reference      :         Direction – Qld region to Powerlink to disconnect Lilyvale Solar Farm

——————————————————————-

Reason :

AEMO ELECTRICITY MARKET NOTICE

Direction – Qld region to Powerlink to disconnect Lilyvale Solar Farm
Refer to Market Notice 144557

In accordance with section 116 of the National Electricity Law AEMO is issuing a direction to Powerlink.  For the purposes of the National Electricity Rules this is a direction under clause 4.8.9(a1)(1).

Direction issued to:  Powerlink at 1253 hrs 15/07/2026
Required actions: Disconnect Lilyvale Solar Farm from the network at 1253 hrs 15/07/2026
Services provided: N/A
Circumstances necessitating the Direction: System Strength
Circumstances necessitating the Direction additional detail: intervention is required to maintain the power system in a secure operating state
Expected duration:  until further notice.

Powerlink does not incur any civil monetary liability for a relevant action taken in accordance with this direction unless the action is taken in bad faith.

Manager NEM Real Time Operations

——————————————————————-
END OF REPORT
——————————————————————-’

MN144557 noted the AEMO Intervention event, and then MN144558 published at 14:15:13:

‘——————————————————————-
MARKET NOTICE
——————————————————————-

From :              AEMO
To   :              NEMITWEB1
Creation Date :     15/07/2026     14:15:13

——————————————————————-

Notice ID               :         144558
Notice Type ID          :         MARKET INTERVENTION
Notice Type Description :         Reserve Contract / Direction / Instruction
Issue Date              :         15/07/2026
External Reference      :         Cancellation: Direction issued to: Powerlink to disconnect Lilyvale Solar Farm

——————————————————————-

Reason :

AEMO ELECTRICITY MARKET NOTICE

Cancellation: Direction issued to: Powerlink to disconnect Lilyvale Solar Farm 

Refer to Market Notice 144556

Direction is cancelled from: 1410 hrs 15/07/2026

Manager NEM Real Time Operations

——————————————————————-
END OF REPORT
——————————————————————-’

MN144559 noted the end of the AEMO Intervention event.

Note that the first Market Notice landed ~2 minutes after the timing of the direction to Powerlink, which makes me wonder how long AEMO had to make this decision.

1)  Note that Semi-Scheduled units are fairly frequently curtailed to zero (including via ‘zero inverter’ constraints) for system strength reasons

2)  What my curiosity is about here is the reason for the direction

… i.e. what changed suddenly requiring AEMO to pull the plug (almost literally speaking)?

A look at LILYSF1

Thankfully with ‘next day public’ data set now available and ez2view close at hand we can take a look … firstly with respect to the LILYSF1 unit via the ‘Unit Dashboard’ widget:

This is a snapshot taken at 08:00 on Thursday 16th July 2026, looking back over 2 days (so covering Wednesday 15th July 2026 and much of Tuesday 14th July 2026).

There’s much that could be highlighted, but particularly wanted to focus on five dispatch intervals, containing two stories:

Story 1)  about the constraint, and disconnection

The focus of this article is that:

1)  beginning the 12:25 dispatch interval, the ‘Q_STR_A19B_LVSF’ constraint equation bound.

(a)  This ‘constrained down’ the unit output:

i.  from an InitialMW of 65MW to a Target of 35MW

ii.  so curtailment of 30MW

(b)  Which the unit followed relatively closely for a Dispatch Error of –4MW.

2)  At the 12:30 dispatch interval, the unit was constrained down further, to 9MW

3)  Then to 0MW from 12:35

4)  That constraint equation remained bound until the 14:10 dispatch interval

We’ll look at the ‘Q_STR_A19B_LVSF’ constraint equation below, but wanted to briefly flag story 2…

Story 2)  about self-forecasting

However, whilst it jumps out to me, also worth noting what happened with the self-forecast on this unit:

1)  at 12:05 on Wednesday (i.e. two columns to the left of what you cans see in the image):

(a)   The unit was constrained

i.  via the ‘Q>BCLC_BCCP_CLWU’ constraint equation

ii.  in the ‘Q-RGLC_8875’ constraint set (as it had been since ~08:15)

iii.  This is one form of a group of constraints that separate out northern QLD units from central QLD units (there are a number of them).

(b)  And was receiving its UIGF from ASEFS (as it had been for the period of the constraint).

2)  for a single dispatch interval at 12:10 on Wednesday (i.e. the time before what’s shown in the three columns in the table):

(a)  The unit was unconstrained (briefly!)

(b)  The UIGF continued to come from ASEFS (i.e. some self-forecasters avoid issuing a self-forecast during SDC periods (there are a few reasons why they might do this), there is a lag between SDC coming off and the self-forecast self-suppressing).

3)  at 12:15 (i.e. the first column) we see that the ‘Q>BCLC_BCCP_CLWU’ constraint equation bound again.

(a)  But the delayed timer on the self-suppression of the self-forecast appears to have kicked in (ever so briefly) because the UIGF commenced again from the self-forecast

(b)  But not for very long…

4)  Because at 12:20 (with the north-central constraint still bound) LILYSF1 reverted back to UIGF from ASEFS

But let’s return to the main story…

The ‘Q_STR_A19B_LVSF’ constraint equation

Utilising the ‘constraint dashboard’ widget in ez2view and focused on the ‘Q_STR_A19B_LVSF’ constraint equation we see the following:

In this snapshot we can see the brief period (just over an hour from this point) with the Marginal Value of the constraint rise as large as –$8,352,000/MWh from 12:35 when the unit was fully constrained off.

On the ‘standing data’ tab that we see in the image we can access the description, which is:

‘Limit 80% to Lilyvale Solar Farm if Stan>=2+Cal>=1+Cal+Glad>=2+ (Stan+Cal+Glad) >=7, Kar>=2,NQLD>350&370(AVG),Ross_FN>150&170(AVG),(Townsville GT>=1,if kar<2),Zero otherwise.’

… which implies that the reason for the sudden drop in the RHS of the constraint (hence the reason the unit was constrained off) was a change in the level of system strength coming from other units in northern QLD.  No time to go into the RHS details to demonstrate.

This constraint equation is one member of …

… the ‘Q-LV_275_BUS_TX_2’ constraint set

Utilising the ‘constraint set details’ widget in ez2view we see the following:

There’s three things highlighted here.

1)  The description says:

‘Out = Lilyvale (H15) 275kV Bus no 1 or no2 and/or Lilyvale 275/132kV transformer 1 or 2 with 132kV network split between Baralaba and Moranbah , open fdrs 7114/1 at Baralaba end and 7382 and 7124/1 at Moranbah end’

2)  This outage is running out for another week, to Thursday 23rd July 2026;

3)  There are a number of Semi-Scheduled units as part of this.

However, in all of the above, (at least in my reading) there’s nothing particularly pointing to the reason why AEMO would need to totally disconnect Lilyvale Solar Farm from the grid.

… perhaps the inverters (even when the solar farm was at 0MW) were still interacting with the network in some way?

PS1 Similar again, on Thursday 16th July 2026

Today (Thursday 16th July 2026) we saw a ‘rinse and repeat’ with MN144563 at 12:09:02 (NEM time) with respect to the Lilyvale Solar Farm:

——————————————————————-
MARKET NOTICE
——————————————————————-

From : AEMO
To : NEMITWEB1
Creation Date : 16/07/2026 12:09:02

——————————————————————-

Notice ID : 144563
Notice Type ID : MARKET INTERVENTION
Notice Type Description : Reserve Contract / Direction / Instruction
Issue Date : 16/07/2026
External Reference : Direction – Qld region to Powerlink to disconnect Lilyvale Solar Farm

——————————————————————-

Reason :

AEMO ELECTRICITY MARKET NOTICE

Direction – Qld region to Powerlink to disconnect Lilyvale Solar Farm
Refer to Market Notice 144564

In accordance with section 116 of the National Electricity Law AEMO is issuing a direction to Powerlink. For the purposes of the National Electricity Rules this is a direction under clause 4.8.9(a1)(1).

Direction issued to: Powerlink at 1210 hrs 16/07/2026
Required actions: Disconnect Lilyvale Solar Farm from the network at 1210 hrs 16/07/2026
Services provided: N/A
Circumstances necessitating the Direction: System Strength
Circumstances necessitating the Direction additional detail: intervention is required to maintain the power system in a secure operating state
Expected duration: until further notice.

Powerlink does not incur any civil monetary liability for a relevant action taken in accordance with this direction unless the action is taken in bad faith.

Manager NEM Real Time Operations

——————————————————————-
END OF REPORT
——————————————————————-

PS2 not on Friday 17th July 2026

Correcting earlier update … this did not happen on Friday 17th July 2026.