Load in modelling 1

[pajce]

Hi guys,
how do you model load on the distribution feeder for steady condition Power Load Flow?
program I work with have options:
1. Constant Power load
2. Constant Current load
3. Constant Impedance load.

problem I have is at my company, we are so used to presenting loads (Power Flow Studies) as Constant Power.
Well, when changing the voltage on the TS bus, say increasing by 5%, my current in the model will go down ( due to S=const). Well I humbly don't believe that is the case in the "real life". I would expect current to follow voltage, 5% voltage increase on line will bring increase in current and vice versa. (in steady condition)
Am I right?
If so, what determines type of load for my study?

Thank you for any comment

 

[saladhawks]

The engineering model that I use for my distribution system (Milsoft WindMil) has load mix types of:

Constant kVA
Constant Impdance
Constant Current

I have attempted to vary the % of each load type based on the loads served for each individual distribution feeder. For example, feeders serving mostly residentail type loads have a high % of Constant Impedance load type. Feeders serving mostly industrial loads have a high % of Constant kVA load type. As a check, I have installed voltage recorders at various locations in the distribution system to check the model's results. You can use these field measurements to help fine tune your load type allocations (assuming of course that your model is 100% accurate).

 

[dpc]

On a distribution feeder, you're going to be somewhere between Constant Impedance and Constant current I would guess. We sometimes model as a 50/50 mix.

Constant current or constant power are more conservative if you are worried about voltage regulation.

 

[jghrist]

Studies done for voltage reduction load management programs have indicated that a 1% drop in voltage on most distribution feeders will result in a 1% decrease in power. This indicates that a constant current model is appropriate. Sorry, I can't cite any particular sources.

I would expect industrial loads to have more of a constant power component because of motor loads, so you might use a mix of 50% constant power and 50% constant current if the software allows it. You might also use a mix of constant impedance with constant current for residential loads.

The reason that a constant power load model is appropriate for transmission studies is that most distribution substations have voltage regulators. Changes in the transmission voltage will not affect the customer voltage, so the power will be constant.

 

[pajce]

Dpc, I am not sure what you mean when you say
"Constant current or constant power are more conservative if you are worried about voltage regulation."?
Can you please put some more details

 

[DanDel]

pajce, to answer your question, current will only increase with increased voltage if your load impedance does not change, which is rare in 'real life'.

Loads such as motors, which usually are responsible for a high percentage of most commercial/industrial loads, will decrease current when the voltage is increased (within certain parameters).

Your motor load should be modeled as 'Constant Power', and probably most of the rest, (like lighting, etc.) should be modeled as 'Constant Impedance'.

 

[dpc]

Right - if you model as constant impedance, as the voltage decreases, the current drops and this reduces the voltages drops in your system.

With a constant current model, the current stays the same as voltage drops, and with constant kVA it increases as voltage drops. So the most conservative approach is to model as constant power or constant kVA.

 

[pajce]

Thank you guys. I guess if I go with Constant Power for all the loads on the feeder (not just motor loads) I won't be able to simulate 3% and 5% voltage reduction that Utilities sometimes perform in emergency in order to offload Power transformers during peak summer loads.

 

[dpc]

Sure you can. But your load current will increase.

 

[jghrist]

I won't be able to simulate 3% and 5% voltage reduction that Utilities sometimes perform in emergency in order to offload Power transformers during peak summer loads.

Which is why the constant power model is not appropriate. If it were, the utilities' voltage reduction would do no good.

 

[dpc]

It depends on what you are trying to accomplish with the model.

 

[stevenal]

I've heard the same as Jghrist, constant currrent is the best model for typical utility distribution circuits. Here is a slightly different take:

http://www.beckwithelectric.com/infoctr/appnotes/App16.pdf

 

[jghrist]

After a little research, I came across this from an EPRI PQ Commentary at

http://www.epri-intelligrid.com/intelligrid/docs/comm_41_final.pdf

Table 2. Voltage reduction test results among four utilities (Source: Power System Voltage Stability by C.W. Taylor, McGraw-Hill, Inc.)
[tt]
Utility Percent Demand Reduction for
1% Voltage Reduction Utility
Residential Commercial
American Electric Power 1 0.80 0.78
American Electric Power 2 0.90 0.86
Consumers Power Company 0.83 1.38
San Diego Gas & Electric Co. 1.14 0.08
[/tt]

A constant current model would give 1%, a constant impedance model would give 2% for small changes, and a constant power model would give 0%.