4.5 | 2 ratings Rate this file 54 Downloads (last 30 days) File Size: 387 KB File ID: #35929 Version: 1.0
image thumbnail

Sequential AC-DC load flow method for two-terminal HVDC networks



A HVDC load flow algorithm in Simulink treated separately from an AC load flow in SimPowerSystems.

| Watch this File

File Information

This demo illustrates a sequential AC-DC load flow method in SimPowerSystems that involves iterations between DC and AC solutions.

The AC load flow is performed by the load flow tool of the Powergui. The DC load flow accommodates a two-terminal HVDC link taking into account tap changers of converter transformers. It is performed by a Simulink model containing the DC power flow algorithm. HVDC link and converter controls parameters are specified in an initialization file.

A script controls the whole process and iterates until the convergence criteria have been reached. This solution technique can be extended to any number of HVDC links connected inside the same AC system or between isolated AC systems.

Required Products SimPowerSystems
MATLAB release MATLAB 7.14 (R2012a)
Tags for This File   Please login to tag files.
Please login to add a comment or rating.
Comments and Ratings (10)
13 Jun 2015 winson woods

Now, I find a method to solve oscillation in rotor mechanical angle difference. They can be eliminated by add a “Unwrap” block before the scope. Anyone who encounters the problem can try this method.

Comment only
09 Jun 2015 winson woods

Report Another Problem.
After I use fixed-step-discrete and Backward Euler method to solve the "LF_AC29bus_HVDCdemo.mdl", I can’t watch abnormal oscillations in the AC/HVDC waveforms pages.
However, when I try to watch the rotor mechanical angle difference between two machines, such as “B_5 MTL-5000MVA” and “LG31-200 MVA”, I can still find the numerical oscillation. When I use a “Rate Transition” block or a “Zero Order Hold” block (multiplication is 333) before the output, the angle-difference curve seems smoother, but there are still much difference between the result and with that given by “Phasor” method. When I use the “Phasor” method to solve the "LF_AC29bus"(without HVDC) model, the angle-difference curve seems much smoother and correct.
There may be some mistakes in my understanding. How to solve the problem? Could anyone give me some suggestions?

Comment only
01 Jun 2015 winson woods

Thanks Silvano Casoria for timely response. After I follow your suggestion to change the solver from “Tustin” to “Backward Euler”, the abnormal oscillation disappears, and when I carry out other simulation, it seems the results are all correct.
However there is still one small problem, if I execute the “LF_ACDC_solve.m“ in mat lab 2013b version, I can get same power flow results as you give in the help file. When I execute the “LF_ACDC_solve.m“ in matlab 2014b version, I get a different load flow results at the HVDC system’s inverter side:
The helpfile and Matlab 2013b: Tap = 1.0000 (pu), alpha = 145.50 (deg), gamma = 21.20 (deg).
Matlab 2014B results: Tap = 0.9875 (pu), alpha = 144.56 (deg), gamma = 22.83 (deg).
Both results won’ introduce the abnormal oscillation. The differences are small and it seems both of the results can be accepted. But I wonder why this would happen.
Thanks Silvano Casoria again to provide such a good tool. Looking forward to the upgrade version.

Comment only
28 May 2015 Silvano Casoria

Until a revision is issued, if the simulation results are not accurate (abnormal oscillations), please use the Backward Euler method to discretize the electrical system (instead of Tustin) in the Powergui block dialog box (Configure parameters:Solver type: Backward Euler).

Comment only
27 May 2015 Silvano Casoria

In response to Winson Woods comment.
Thank you for pointing out the problem. We are working to find a solution.

Comment only
27 May 2015 winson woods

Thanks Silvano Casoria for sharing the "Sequential AC-DC load flow method for two-terminal HVDC networks" and the simpowersystem models. I think it provide a high efficient tool to carry out AC/DC simulation in matlab.

However, after I run the "LF_ACDC_solve.m" and start the "LF_AC29bus_HVDCdemo.mdl" in matlab 2014b, I find that it can't reach a stable operation point. When I watch the AC/DC waveforms (voltages, powers, and angles), they oscillation even in the end of whole 6 seconds’ simulation. What's the problem? Could you fixed the problems and provide an upgrade version of the “HVDC_Load_Flow’’ example?

19 Nov 2014 Silvano Casoria

In response to yusuf comment.
It is possible to have only two AC busses in the AC system. I suppose that the two AC busses are the rectifier and the inverter commutating busses.

The sequential method implies executing successfully AC system and a DC system load flow analysis. Refer to the SimPowerSystems documentation regarding the usage of the load flow analysis function.

Comment only
13 Nov 2014 yusuf

yusuf (view profile)

I would like to learn that I can use this method for 2 AC bus load flow analysis or not?
If I can use, how? Thank you.

Comment only
02 Dec 2012 Sameetr  
26 Oct 2012 yq meng


Comment only

Contact us