Hello, GridLAB-D team
I have a microgrid that goes to the islanding mode and a diesel generator should supply microgrid. the microgrid load size is: 640 + 460j = 868VA. And the Diesel_DG specifications are as follows:
object diesel_dg {
parent node_151;
name Gen_151;
Rated_V 4160;
Rated_VA 1800000;
TotalRealPow 1500000; // 640 × 2≈1500
// pf 1;
flags DELTAMODE;
Gen_type DYN_SYNCHRONOUS;
rotor_speed_convergence $ {rotor_convergence};
Exciter_type SEXS;
Governor_type GGOV1;
object recorder {
property rotor_speed, rotor_angle, pwr_electric.real, pwr_electric.imag, pwr_mech;
flags DELTAMODE;
// interval -1;
interval 1;
file "Gen_151_Speed.csv";
};
}
Due to the (arbitrarily setting to 50%!?)
Rated_VA = micrigrid load × 2 = 868 × 2 ≈ 1800 kW. but with this, the microgrid becomes unstable.
If the Rated_VA of the diesel_DG changed to 4000 kW instead of 1800 kW (which is much more than the amount of microgrid load), the microgrid becomes stable.
............... Generally, how to determine the Rated_VA or power_out of the diesel_DG?
I actually want the output power of the Diesel_DG (TotalRealPow) to be as large as the microgrid load.
I do not understand the difference between (Rated_VA [VA]; // nominal capacity in VA) and (TotalRealPow [W]; // total real power generated), Because (TotalRealPow) is ignored in simulation, and simulation runed only according to to the (Rated_VA).
1. Rated_VA 1800000;
TotalRealPow 1500000; // 640 × 2≈1500
2. Rated_VA 4000000;
TotalRealPow 1500000; // 640 × 2≈1500
Thank you
-aro
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Sorry for the delay in getting back to you. This got lost in the pile of things to answer last week.
Sizing the generator can be difficult, especially for overall transient stability. While the sizing according to system load is a good start, it isn't always sufficient. Many of the parameters within the generator are given in per-unit, so theoretically, they'll adjust for the size of the generator, but that isn't always true. So along with the Rated_VA property, you may need to adjust things like the internal impedance of the generator, or possibly change some of the control constants/gains/delays. If you have an actual generator you can leverage, that helps a lot. Otherwise, it can be a case of trial and error to find the right VA setting and parameters that will work. We too have had generators that start stable, but then go unstable, and adjusting the parameters to a "better set that matches that generator size" improved the simulation.
In regards to the Rated_VA versus TotalRealPow setting, the latter is only used in a steady-state version of the model. Quite frankly, that model isn't very accurate, and is actually in the process of being removed. So Rated_VA is the size of the generator, but if you want the power output, I'd record the values in pwr_electric, or capture the three individual values in power_out_A, power_out_B, and power_out_C. For what you were looking at, the pwr_electric value is probably the one you want to use.
Hopefully that helps.
-Frank
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
the shaded regions represent the boundaries of the three microgrids. End-use loads in these regions can be supplied by the microgrids when the substation voltage source is not present, those in the non-shaded region cannot.
The distributed energy resources (DER) of the three microgrids are shown in Table I. For each of the nine generation sources, Table I indicates which microgrid they are located in, which node they are connected to, the generator type, the rated active power, and the controller type.
Due to dynamic behaviors, it is not possible to energize Microgrid 3 using the existing combination of generation resources of Microgrid 1 and Microgrid 2. Scenario 2: Re-Energizing Microgrid 3 (Oversized Generators on Microgrid 2)
In Scenario 2, the size of the generator G4 is increased from 600 kW to 750 kW. My simulations:
I made some simplifications. For example, I deleted the PV's and put only two generators G_151 and G_300 (with the following specifications) in microgrid1 and microgrid2.
At t=10, microgrid 1 and 2 are disconnected from the grid. In t=15 the switch 151 to 300 is closed and the microgrid 3 is disconnected from the grid and connected to two other microgrids.
my outputs are as follows:
Question's:
1. Why the output of the generator and the frequency (t=15 to end) have a lot of oscillation , but in the article, these values are smooth and without oscillation ?
2. And While I have chosen the VA (both generator 10 MVA) much more than the amount of microgrids load (2.5 MVA) the simulations Hardly could be stabled. (for VA=5 MVA leads to collapse!!)
In the paper, the total power of generators is approximately equal to the load size.
really I don't understand how to set up the amount of generator power.
Node_300 : bustype SWING;
Node_151 : bustype SWING;
Node_150 : bustype SWING;
Could you please take a look at the attached .glm file and see if there is anything wrong with it? Thank you & Best wishes
-aro
I've attached a version of the IEEE 123-node feeder used for this paper. The attached is actually oriented towards scenario 4, but it should be similar to scenario 1 and scenario 2. This may not be the exact system from the paper, but it is producing results that look almost identical (Fig. 8).
1) The oscillations can be caused by a number of things. This can be different constants in the control, it may be in how generators are interacting, or could even just be the system being marginally stable under the conditions you've provided. The simple answer for comparison of the paper is "it is a different system".
2) The VA rating of the generator is only one parameter of the simulations. This is actually why dynamic simulation is so important and why simple "energy balance" scenarios don't always capture the full picture. Some of the parameters may not be compatible with the other portions of the system, or cause an unstable-looking response.
Hello dear Frank Thank you very much for your help
I checked my codes. They had two problems.
1. In my codes: (node_151 and node_300 were SWING) but in your codes: (node_151 and node_300 are SWING_PQ) (multiple microgrids nodes)
2. And in my codes the governor also had default values but you have changed the default values (GGOV1).
I have two questions:
1. How these GGOV1 control values be calculated according to the microgrid (with trial and error or set by a special procedure).
2. How can I create a sensor in the gridlabd? for example, if voltage is increased to a certain extent, an arbitrary switch be opened?
Is there such a function in the gridlabd? Or do I have to make it myself?
thanks again for your great guidance & best wishes
-aro
Last edit: aro 2018-04-05
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
For item 1, the values either came from a data sheet for an existing generator, or they were done by trial and error. There are likely special procedures to tune them, but for "non-utility" studies like these, we typically just do the "trial and error" approach. This isn't completely unguided -- a lot of the time, you can look at a block diagram of the control (GGOV1) and see which parameters might be ones you'd need to adjust to get the response you're looking to get. For example, if something is responding too quickly (and causing an oscillation), decreasing a gain or changing a delay may help.
For the second item, there is no already-created object to do this. It should be pretty straight-forward to do, and you could potentially do it in runtime code (though that hasn't been thoroughly tested with deltamode, so I'm not sure if it would work). We do have some work that is going to create an interface mechanism for "general protection/action switches", but we won't have that for a little while. If you do implement it, I'd recommend either creating a new switch-based object, or add it as an operating mode for one of the other devices. See the GLD_PROPERTY pages for how to interface with the nodes to get the voltages (there are examples of how to use them in regulator.cpp inside powerflow).
-Frank
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hello, GridLAB-D team
I have a microgrid that goes to the islanding mode and a diesel generator should supply microgrid. the microgrid load size is: 640 + 460j = 868VA. And the Diesel_DG specifications are as follows:
object diesel_dg {
parent node_151;
name Gen_151;
Rated_V 4160;
Rated_VA 1800000;
TotalRealPow 1500000; // 640 × 2≈1500
// pf 1;
flags DELTAMODE;
Gen_type DYN_SYNCHRONOUS;
rotor_speed_convergence $ {rotor_convergence};
Exciter_type SEXS;
Governor_type GGOV1;
object recorder {
property rotor_speed, rotor_angle, pwr_electric.real, pwr_electric.imag, pwr_mech;
flags DELTAMODE;
// interval -1;
interval 1;
file "Gen_151_Speed.csv";
};
}
Due to the (arbitrarily setting to 50%!?)
Rated_VA = micrigrid load × 2 = 868 × 2 ≈ 1800 kW. but with this, the microgrid becomes unstable.
If the Rated_VA of the diesel_DG changed to 4000 kW instead of 1800 kW (which is much more than the amount of microgrid load), the microgrid becomes stable.
............... Generally, how to determine the Rated_VA or power_out of the diesel_DG?
I actually want the output power of the Diesel_DG (TotalRealPow) to be as large as the microgrid load.
I do not understand the difference between (Rated_VA [VA]; // nominal capacity in VA) and (TotalRealPow [W]; // total real power generated), Because (TotalRealPow) is ignored in simulation, and simulation runed only according to to the (Rated_VA).
1. Rated_VA 1800000;
TotalRealPow 1500000; // 640 × 2≈1500
2. Rated_VA 4000000;
TotalRealPow 1500000; // 640 × 2≈1500
Thank you
-aro
Hello Aro,
Sorry for the delay in getting back to you. This got lost in the pile of things to answer last week.
Sizing the generator can be difficult, especially for overall transient stability. While the sizing according to system load is a good start, it isn't always sufficient. Many of the parameters within the generator are given in per-unit, so theoretically, they'll adjust for the size of the generator, but that isn't always true. So along with the Rated_VA property, you may need to adjust things like the internal impedance of the generator, or possibly change some of the control constants/gains/delays. If you have an actual generator you can leverage, that helps a lot. Otherwise, it can be a case of trial and error to find the right VA setting and parameters that will work. We too have had generators that start stable, but then go unstable, and adjusting the parameters to a "better set that matches that generator size" improved the simulation.
In regards to the Rated_VA versus TotalRealPow setting, the latter is only used in a steady-state version of the model. Quite frankly, that model isn't very accurate, and is actually in the process of being removed. So Rated_VA is the size of the generator, but if you want the power output, I'd record the values in pwr_electric, or capture the three individual values in power_out_A, power_out_B, and power_out_C. For what you were looking at, the pwr_electric value is probably the one you want to use.
Hopefully that helps.
-Frank
Frank, thanks for your help
actually, I want to simulate the first and second scenarios of this article
http://ieeexplore.ieee.org/document/7857787/
that done by gridla-d.
the shaded regions represent the boundaries of the three microgrids. End-use loads in these regions can be supplied by the microgrids when the substation voltage source is not present, those in the non-shaded region cannot.
The distributed energy resources (DER) of the three microgrids are shown in Table I. For each of the nine generation sources, Table I indicates which microgrid they are located in, which node they are connected to, the generator type, the rated active power, and the controller type.
Scenario 1: Re-Energizing Microgrid 3 (with Microgrid 1&2)
Due to dynamic behaviors, it is not possible to energize Microgrid 3 using the existing combination of generation resources of Microgrid 1 and Microgrid 2.
Scenario 2: Re-Energizing Microgrid 3 (Oversized Generators on Microgrid 2)
In Scenario 2, the size of the generator G4 is increased from 600 kW to 750 kW.
My simulations:
I made some simplifications. For example, I deleted the PV's and put only two generators G_151 and G_300 (with the following specifications) in microgrid1 and microgrid2.
At t=10, microgrid 1 and 2 are disconnected from the grid. In t=15 the switch 151 to 300 is closed and the microgrid 3 is disconnected from the grid and connected to two other microgrids.
my outputs are as follows:
Question's:
1. Why the output of the generator and the frequency (t=15 to end) have a lot of oscillation , but in the article, these values are smooth and without oscillation ?
2. And While I have chosen the VA (both generator 10 MVA) much more than the amount of microgrids load (2.5 MVA) the simulations Hardly could be stabled. (for VA=5 MVA leads to collapse!!)
In the paper, the total power of generators is approximately equal to the load size.
really I don't understand how to set up the amount of generator power.
Node_300 : bustype SWING;
Node_151 : bustype SWING;
Node_150 : bustype SWING;
Could you please take a look at the attached .glm file and see if there is anything wrong with it?
Thank you & Best wishes
-aro
Last edit: aro 2018-04-03
Hello Aro,
I've attached a version of the IEEE 123-node feeder used for this paper. The attached is actually oriented towards scenario 4, but it should be similar to scenario 1 and scenario 2. This may not be the exact system from the paper, but it is producing results that look almost identical (Fig. 8).
1) The oscillations can be caused by a number of things. This can be different constants in the control, it may be in how generators are interacting, or could even just be the system being marginally stable under the conditions you've provided. The simple answer for comparison of the paper is "it is a different system".
2) The VA rating of the generator is only one parameter of the simulations. This is actually why dynamic simulation is so important and why simple "energy balance" scenarios don't always capture the full picture. Some of the parameters may not be compatible with the other portions of the system, or cause an unstable-looking response.
Hopefully the attached system helps.
-Frank
Hello dear Frank
Thank you very much for your help
I checked my codes. They had two problems.
1. In my codes: (node_151 and node_300 were SWING) but in your codes: (node_151 and node_300 are SWING_PQ) (multiple microgrids nodes)
2. And in my codes the governor also had default values but you have changed the default values (GGOV1).
I have two questions:
1. How these GGOV1 control values be calculated according to the microgrid (with trial and error or set by a special procedure).
2. How can I create a sensor in the gridlabd? for example, if voltage is increased to a certain extent, an arbitrary switch be opened?
Is there such a function in the gridlabd? Or do I have to make it myself?
thanks again for your great guidance & best wishes
-aro
Last edit: aro 2018-04-05
Hello Aro,
For item 1, the values either came from a data sheet for an existing generator, or they were done by trial and error. There are likely special procedures to tune them, but for "non-utility" studies like these, we typically just do the "trial and error" approach. This isn't completely unguided -- a lot of the time, you can look at a block diagram of the control (GGOV1) and see which parameters might be ones you'd need to adjust to get the response you're looking to get. For example, if something is responding too quickly (and causing an oscillation), decreasing a gain or changing a delay may help.
For the second item, there is no already-created object to do this. It should be pretty straight-forward to do, and you could potentially do it in runtime code (though that hasn't been thoroughly tested with deltamode, so I'm not sure if it would work). We do have some work that is going to create an interface mechanism for "general protection/action switches", but we won't have that for a little while. If you do implement it, I'd recommend either creating a new switch-based object, or add it as an operating mode for one of the other devices. See the GLD_PROPERTY pages for how to interface with the nodes to get the voltages (there are examples of how to use them in regulator.cpp inside powerflow).
-Frank
Hello dear Frank
Thank you for your suggestions and advice.
best wishes
-aro