Fix Synchronous Machines, when the original set-point in SSH is out of limits

[Haigutus]

To enable limits on PF when merging the limits wither must be changed to fit the set-point in original model or the set-point must be changed to be within limits

[Haigutus]

EQ

SSH

SV

[Haigutus]

Parameters to use: PQ_area = polygon(SynchronousMachine.InitialReactiveCapabilityCurve -> ReactiveCapabilityCurve <- CurveData) Q_limits = lines(SynchronousMachine.maxQ, SynchronousMachine.minQ) P_limits = lines(SynchronousMachine -> GeneratingUnit.maxOperatingP, SynchronousMachine -> GeneratingUnit.minOperatingP) PQ_setpoint = point(SynchronousMachine.p, SynchronousMachine.q) PQ_solution = point(SynchronousMachine <- Terminal <- SvPowerFlow.p, SynchronousMachine <- Terminal <- SvPowerFlow.q)

Logic:

1. Find all machines out of their PQ_area by SSH setpoint
2. Find all machines not in point 1. and out of their P and Q limits
3. Filter out all switched off machines PQ_setpoint == (0, 0)?
4. Remove all CurveData from machines found in point 1. and 2.
5. Expand Q and P limits to contain SSH setpoint

[Haigutus]

Possible issues

1. Is it guaranteed to have 1-1 Synchronous Machine and Generating Unit? As P limits are under Generating Unit and Q limits under SynchronousMachine
2. Different P limits under Generating Unit, witch ones are used?
3. What about SynchronousMachine.ratedS
4. What about SynchronousMachine <- Terminal <- OperationalLimitSet <- CurrentLimit.value
5. Inverted signs PQ curve vs SSH and SV p and q
6. What to do with PQ_area that is a line?
7. What is deadband to consider setpoint out of bounds?
8. How can one be sure that generator is switched off?
9. Units in PQ curve do not have defined multiplier Mega?

[Haigutus]

Responses to possible issues and new solution

The proposal for next merge is to update the script to disable regulating controls for Units/Machines that are out of PQ_curve or PQ_limits, this should set the machine to fixed PQ node and then limits should not stop the power flow in tools.

Responses that I picked up from today’s discussion:

1. Is it guaranteed to have 1-1 Synchronous Machine and Generating Unit? This is true for CGM creation process, but needs addressing in CGMES
2. Different P limits under Generating Unit, which ones are used? GeneratingUnit.maxOperatingP and GeneratingUnit.minOperatingP
3. What about SynchronousMachine.ratedS (marked green on the plots)? Not relevant for CGM creation process
4. What about SynchronousMachine <- Terminal <- OperationalLimitSet <- CurrentLimit.value? Not relevant for CGM creation process
5. Inverted signs PQ curve vs SSH and SV p and q? Not relevant for CGM creation process
6. What to do with PQ_area that is a line? Needs further discussion
7. What is deadband to consider setpoint out of bounds? Needs further discussion
8. How can one be sure that generator is switched off? Needs further discussion
9. Units in PQ curve do not have defined multiplier Mega? Not relevant for CGM creation process

[Haigutus]

Intended result is that controlEnabeled is set to False in SSH in cases where the machines are outside of PQ_area or PQ_limits.

SSH

[Haigutus]

If Machine RegualtingConEq.controlEnabled == false, should any RegualtingControls be applied even they are active? – I would expect not?

• If you have multiple devices within one Regulating control you can have some devices with RegulatingCondEq.controlEnabled = true and some with false. This would then specify which devices are active within the RegulatingControl
• If you have 1 device 1 control then if controlEnabled is false and .enabled is true the result should be the same if controlEnabled is true and .enabled is false. There could be a sanity check to verify if there are cases where you have .controlEnabled false and .enabled true , but at the end the load flow will need to make use of the 2 attributes in combination and give same result

[Haigutus]

Solution to this issue:

1. Switch off control on machine
2. Switch off all connected regulating controls*

*might not be best solution as one regulating control can target multiple machines, but for testing it is good by ensuring that all tools behave the same. And no such cases have been observed where there are multiple machines targeted by single control.

[Haigutus]

Result