Originally posted by: Marko_Obert

Hi,

I have issues with the constraint "ctROCOF":

// Decision Variables
dvar float+ Pg[UnitID][subset];                  
dvar boolean ug[UnitID][subset];          
dvar float TotalInertia[subset];  
dvar float TotalInertiaTrip[UnitID][subset];

    forall(h in subset, u in UnitID)
           ctROCOF:
                 Pg[u][h] * c >= TotalInertiaTrip[u][h];

Result in: CPLEX Error  5002: Q in ''ctROCOF#0#-1'' is not positive semi-definite.   

I think this error is indirectly caused by the constraints to calculate "TotalInertiaTrip":

    forall(h in subset)
         calcInertia:
             TotalInertia[h] == sum(u in UnitID)(Inertia[u] * ug[u][h]);
    
    forall(h in subset, u in UnitID)
         calcDeltaInertia:
             TotalInertiaTrip[u][h] == (TotalInertia[h] - Inertia[u])*ug[u][h]

Observation: If up[u][h] is zero Pg[u][h] and TotalInertiaTrip[u][h] is zero, too.

What can I do to solve this issue?

Kind regards

Hi,

have you had a look at

https://www.ibm.com/developerworks/community/forums/html/topic?id=aa9aa3db-4fbc-4209-a767-5b5e54902cbd&ps=25

?

regards

https://www.linkedin.com/pulse/ist-mathematische-optimierung-und-wie-kann-sie-helfen-alex-fleischer/

Originally posted by: Marko_Obert

Hi Mr. Fleischer,

thank you for the link. I tried option 2. Now, error 5002 disappeared, but the constraint still does not work as expected:

    forall(h in subset)
         calcInertia:
             TotalInertia[h] == sum(u in UnitID)(Inertia[u] * ug[u][h]);
    
    // Calculate total inertia, after a generator trips
    forall(h in subset, u in UnitID)
         (ug[u][h]==1) => (TotalInertiaTrip[u][h] == TotalInertia[h] - Inertia[u]);
             
    forall(h in subset, u in UnitID)
         (ug[u][h]==0) => (TotalInertiaTrip[u][h] == 0);

   // Calculate the rate of change of frequency, when a generaotr trips
    forall(h in subset, u in UnitID)
           ctROCOF:
              TotalInertiaTrip[u][h] >= Pg[u][h] * c;

Is my implementation of option 2 correct?

well this looks good