A voltage change begins at the source bus, passes through the transformer and feeder impedance, may be intercepted by voltage regulation, and finally reaches each load terminal. Follow that path before assigning a load model. The correct model is the one that reproduces measured changes in feeder power and current for the voltage that actually reaches the loads, not merely the voltage commanded at the source.
Constant power, constant current, and constant impedance describe different steady-state responses. They do not describe transient motor starting, protection operation, or delayed control actions. For a power-flow study, select or combine them according to the load composition, regulation path, study objective, and field response.
Where Does the Voltage Change Stop?
Start with simultaneous voltage readings at the source bus, regulated bus, and representative feeder locations. A 5% source-bus increase does not prove that customer-terminal voltage increased by 5%. A transformer tap controller or feeder voltage regulator can absorb the change and hold the downstream voltage near its target.
| Reading | Outcome | Meaning | Next check |
|---|---|---|---|
| Source voltage changes; load-bus voltage does not | Voltage change stops at regulation | Downstream load power can remain nearly unchanged because its terminal voltage did not change | Check regulator or tap response and model its control state |
| Source and load-bus voltages change together | Voltage reaches the load | The selected voltage-dependence model controls the calculated power and current response | Compare measured power and current before and after the change |
| Voltage changes differ by feeder location | Regulation and feeder drop divide the change | A single feeder-wide load assumption may hide local behavior | Partition loads by regulated section, customer class, or measurement area |
This distinction explains why constant power is often useful for an upstream transmission study. Distribution substations with active voltage regulation can isolate customer voltage from a transmission-voltage change. The transmission model then sees approximately unchanged downstream demand even though individual customer devices are not inherently constant-power loads.
What Does the Current Reading Mean?
At a load bus, apparent power follows S = V I*. For magnitude comparisons with unchanged power factor, use normalized voltage v = V/V0, power p = P/P0, and current i = I/I0. The three basic models then produce distinct branches:
| Model | Power law | Current law | Response to rising voltage | Response to falling voltage |
|---|---|---|---|---|
| Constant power or constant kVA | p = 1 |
i = 1/v |
Current decreases | Current increases |
| Constant current | p = v |
i = 1 |
Power increases; current remains fixed | Power decreases; current remains fixed |
| Constant impedance | p = v^2 |
i = v |
Power and current increase | Power and current decrease |
The expectation that current must rise with voltage applies only when load impedance remains fixed. A controlled load or a normally loaded motor can draw less current when voltage rises because it is maintaining approximately the same power within its operating range.
For a 5% voltage increase at the load terminal, a constant-power model calculates I/I0 = 1/1.05 = 0.9524, a 4.76% current decrease. Constant current produces no current change and a 5% power increase. Constant impedance produces a 5% current increase and 1.05^2 = 1.1025, a 10.25% power increase.
For a 5% voltage reduction, constant power produces I/I0 = 1/0.95 = 1.0526, a 5.26% current increase with unchanged power. Constant current produces a 5% power reduction. Constant impedance produces a 5% current reduction and 0.95^2 = 0.9025, a 9.75% power reduction. These calculations assume unchanged power factor and apply to steady-state magnitude response.
Which Model Matches the Measured Power Response?
Calculate the percentage change in voltage at the load location and compare it with the simultaneous percentage change in power. For small voltage changes, a 1% voltage reduction gives the following identifying responses:
| Observed power change for a 1% voltage drop | Matching model | Current behavior |
|---|---|---|
| Approximately 0% | Constant power | Current rises to maintain power |
| Approximately 1% decrease | Constant current | Current remains approximately fixed |
| Approximately 2% decrease | Constant impedance | Current falls approximately with voltage |
A measured one-for-one change between voltage and power points toward constant-current behavior. A stronger, approximately two-for-one response points toward constant impedance. Little power change points toward constant power, provided the voltage measurement is taken downstream of any active regulator.
Use both real and reactive power when the software models them separately. A fit based only on real power can reproduce transformer kW while missing feeder current, voltage drop, or power factor. Compare calculated current as an independent check rather than accepting a power-only match.
Does Load Composition Change the Decision?
Yes. A distribution feeder aggregates devices with different voltage responses. Commercial and industrial feeders commonly contain a larger constant-power component because motors can reduce current as voltage rises while maintaining mechanical output within their operating range. Lighting and other passive elements can contribute constant-impedance behavior. Residential feeders may therefore warrant a higher impedance fraction, while motor-heavy industrial feeders may warrant a higher constant-power fraction.
Those classifications are starting hypotheses, not fixed allocations. A candidate 50% constant power / 50% constant current mix has been used for industrially influenced demand. A 50% constant impedance / 50% constant current mix has also been used for distribution feeders. Neither split should be applied without checking measured feeder response.
A composite ZIP representation expresses the active-power response as P/P0 = aZ v^2 + aI v + aP, where aZ + aI + aP = 1. The coefficients represent constant-impedance, constant-current, and constant-power fractions. Apply the same concept separately to reactive power if the program supports it; do not copy active-power coefficients into reactive power without comparing measured kvar behavior.
Does Voltage Regulation Hide Load Behavior?
Active regulation changes the path between the commanded voltage and the load. If a source-bus voltage reduction causes a regulator to raise its tap and restore the customer voltage, the underlying voltage-sensitive load receives little or no reduction. The upstream study may then observe nearly constant demand even when the devices downstream contain substantial current or impedance components.
Record regulator position or control state with voltage and power. If the tap changes during the test, separate the source-voltage response from the terminal-voltage response. For a voltage-reduction load-management study, model the control action that will actually occur. A constant-power load at every bus predicts no load-power reduction and raises current as voltage falls, so it cannot represent a measured reduction in feeder demand unless regulation prevents the voltage change from reaching the loads.
For voltage-regulation assessment, constant power is generally the more demanding basic case because falling voltage drives current upward, increasing feeder voltage drop. Constant current maintains the original current as voltage falls. Constant impedance is self-relieving because both current and power fall. Run the demanding case for equipment and voltage limits, but use the calibrated composite case for the expected operating result.
Which Field Measurements Calibrate the Model?
Install synchronized voltage recorders at the source or transformer secondary and at representative feeder locations. Pair those readings with feeder kW, kvar, current, regulator state, and the modeled load allocation for the same operating interval. The model topology and load distribution must first reproduce the base condition; otherwise an incorrect conductor, transformer, switch state, or load placement can be mistaken for a bad load model.
- Capture a stable pre-change operating point with bus voltage, feeder current, kW, kvar, and control state.
- Apply or identify a steady voltage change, such as the 3% or 5% reductions being studied.
- Wait for the selected steady condition and use a measurement window that matches the intended study horizon. Do not mix an immediate response with a later controlled response.
- Calculate normalized voltage, power, and current at each measured location.
- Compare the response with the constant-power, constant-current, and constant-impedance laws.
- Adjust the composite fractions by feeder section or load class until voltage, kW, kvar, and current move in the measured directions and by comparable amounts.
Repeat the comparison at more than one loading level when measurements permit. One operating point can hide a compensating error between load allocation and voltage-dependence coefficients.
How Do You Configure and Verify the Resolving Model?
- Map the electrical path from the source bus through transformers, regulators, feeder impedance, and load buses.
- Confirm which downstream voltages change during the study condition. Treat buses held by active regulation separately from buses that follow the source.
- Group loads by location and dominant behavior. Assign candidate power, current, and impedance fractions rather than one feeder-wide model when measurements show different responses.
- Enter the base kW and kvar at the reference voltage, then enter the selected load-type fractions. Check that each set of fractions totals
1.0. - Solve the base case and reconcile bus voltage, branch current, transformer loading, feeder kW, and kvar with field readings.
- Run the required 3% and 5% voltage-change cases. Compare results at the load buses, not only at the source bus.
- For a reduction case, verify the expected direction: constant-power current rises, constant-current power falls in direct proportion to voltage, and constant-impedance current and power both fall.
- Run a constant-power sensitivity case where low voltage or equipment loading is the concern, then retain the field-calibrated composite case as the expected operating case.
FAQ
Can I model every feeder load as constant power?
You can use it as a demanding voltage-drop sensitivity case, but it predicts no demand reduction during a 3% or 5% terminal-voltage reduction. Use a calibrated current and impedance contribution when field kW falls with voltage.
Does a 5% voltage increase always increase feeder current?
No. It increases current by 5% for constant impedance, leaves current unchanged for constant current, and decreases current by 4.76% for constant power, assuming unchanged power factor.
Can I use a 50/50 load mix without measurements?
Use a 50/50 mix only as a candidate case. Compare measured voltage, kW, kvar, and current before assigning either a power/current or impedance/current split to the study model.
Does a voltage regulator make the downstream load constant power?
No. It can make upstream demand appear constant by holding the load-terminal voltage steady. Verify the final model by matching the measured load-bus voltage, feeder kW, kvar, and current after the regulator reaches its study-state position.