Setting Parallel R–L Load in IEEE Node 611 Model Produces Unrealistic Values

Question

Hi everyone,

I’m modelling the IEEE 13-Bus Test Feeder in Typhoon HIL and replacing the series R–L load at Node 611 with a parallel R–L branch to represent the spot load.

I want to set the resistance and inductance based on the desired active (P) and reactive (Q) power at nominal voltage, using the standard relationships:

P = V² / R
Q = V² / X , where X = 2πfL

So for my case:

• Line-to-line voltage: V_LL = 4.16 kV

• Line-to-neutral voltage: V_ph = 0.9777 × (4160 / √3)

• Frequency: f = 60 Hz

• Desired P = 170 kW, Q = 80 kVAr

Then:

L = (0.9777 × 4160 / √3)² / (80,000 × 2π × 60)
R = (0.9777 × 4160 / √3)² / 170,000

However, when I run the model and try to dynamically adjust P by varying the resistor in real time (using a variable resistor block in Typhoon and SCADA input), the resulting power readings become extremely large—much higher than expected.

Am I missing something in how the parallel R–L load is modeled or how the dynamic resistance update is interpreted by Typhoon? Could it be a scaling issue with the per-phase voltage, or the fact that power depends on V²/R while the system voltage changes slightly under load?

Any advice on how to correctly parameterize and dynamically adjust P while keeping Q fixed would be appreciated

I have made a standalone model simulating the node611 from ieee13 node feeder and im still getting the same huge numbers. The schematic and scada panel are linked below. 

schematic model download

scada panel download

Answer 1

This might have something to do with the snubber inductance in your variable resistor (doc). In addition to this, variable inductor acts as a current source and it might also need snubbers. If not properly parameterized, both of these can destabilize the model. I would try to first confirm that the resistor indeed acts "almost like an ideal resistor", i.e., that the snubber inductance didn't destabilize the circuit (especially at low load - large resistor value). 

Three-phase Variable Load component should do the snubber parameterization for you, but it's quite expensive in terms of resources. If you're tight with resources, debugging focused on the variable elements would be the way to go.