You need to do the following:

1. Prepare three equivalent bus admittance matrices corresponding to the pre-fault, during-fault, and post-fault structures of the system. The pre-fault matrix represents the original system. The during-fault matrix represents the system with the faulty line model according to the fault type and location. The post-fault matrix represents the system with the fault clread.

2. Decide the times of the fault occurance, and fault clearing.

3. Perform the time domain simulation such that the simulation changes the system equivalent admittance matrix of the system at the appropriate times. Knowledge of the use of the 'if... then' logic is required.

You can refer to the following book

https://www.researchgate.net/publication/307584190_Dynamic_Security_of_Interconnected_Electric_Power_Systems_-_Volume_1?_sg=qg62isJ5w3GPztObGoN5KY1r-209pJLHCmCoSF1-jMDQyi155ylqcdcPdaZutev4UVb4x5j8x09RnbObXwnvhM9ss_dyDjj4RaWsUuFO.vU3V1vTfSm-wXWAEbhMcF6rwx2nbvk5rMe9bdaJkpel6Xy2rBTT9njRm7-VoaeTKfWo2wN25Mn6rMH-qKYxI3A

You can also refer to my 2004 PhD thesis which available also at my RG page.

https://www.researchgate.net/publication/307585561_Electromechanical_System_Equivalent_and_Dynamic_Load_Aggregation_for_Transient_Stability_of_Electric_Power_System?_sg=FCDnlYFWmMmh6m1iCtrBIbphNV3ze-pFtcQlDhPEbLFCHuUjKpq-fKCEpofTqasNHGKRU2_9-S6sIO6852Q-UOd2V5AUuJUCe06nq4si.KzlL4sZiuNDwbIxxqD0XX42WT_rGkxJM_LdltJbJTKJHEwhsB9yWYpHpAllgOVFYnUPhcXSEB1zqB2sOBjt0Vw

Good luck

Fortunately, I found the code of setting the disturbance. It is a modification of the PST codes. Here it is :) Here it is :)

%**** setup disturbance times (all in seconds)****

% simulation

t_switch(1) = 0; %start time

t_switch(2) = 0.10; % time to apply the disturbance

fault_dur = input('Disturbance duration in sec =');

t_switch(3) = t_switch(2)+fault_dur; % fault clearing time

t_switch(4) = input('Total simulation time in seconds ='); % end time

h = input('time step size in sec ='); % integration stepsize

% N.B. according to PAS-97, no 4, 1978, h may be taken 0.1 or 0.2 seconds

stepsize = h;

k_switch(1) = round((t_switch(2)-t_switch(1))/stepsize)+1;

k_switch(2) = round((t_switch(3)-t_switch(1))/stepsize)+1;

k_switch(3) = round((t_switch(4)-t_switch(1))/stepsize)+1;

%*****Disturbance *****

%Fault at bus 2 cleared after fault_dur by removing line 2-19

% *****Step#1: Formuation of reduced Ybus****

%(1) Prefault: construct reduced Y matrix - all loads are assumed

% to be of constant impedance load

[Y_red,V_rec] = red_ybus(bus,line); % pre-fault reduced Ybus

%(2) first change and Y_red_f

bus_f = bus;

dist_bus = input('input disturbance bus no.=');

dist1 = input('input the load first change in p.u=')

bus_f(dist_bus,6) = bus_f(dist_bus,6)*(1-dist1); %busload sudden drop

[Y_red_f,V_rec_f] = red_ybus(bus_f,line); % during fault reduced Ybus

%(3) second change and Y_red_pf

bus_pf = bus_f;

disp('For one disturbance, the second = 0')

dist2 = input('input the load second change in p.u=')

bus_pf(dist_bus,6) = bus_pf(dist_bus,6)*(1+dist2); %busload disturbance

[Y_red_pf,V_rec_pf] = red_ybus(bus_pf,line); % post-fault reduced Ybus

%****Step#2 intializing the linearized model********

%*********by calculating intial conditions *********

% step 2: initialization

flag = 0;

%all machines are represented by constant emf behined transient reactance

%linearized model

f = mac_em(0,1,bus,flag);

%******Step#3: intergartion by "Trapizoidal rule"

% or "Recurrsive convolution process"

% also "Predector -Corrector method" *******

for k = 1:k_switch(3)

% step 3a: network solution

mach_ref(k) = 0;

flag = 1;

f = mac_em(0,k,bus,flag); % network-machine interface

psi = psi_re(:,k) + jay*psi_im(:,k);

if k >= k_switch(2)

cur = Y_red_pf*psi; % network solution

bus_v(:,k) = V_rec_pf*psi; % bus voltage reconstruction

elseif k >= k_switch(1)

cur = Y_red_f*psi; % network solution

bus_v(:,k) = V_rec_f*psi; % bus voltage reconstruction

else

cur = Y_red*psi; % network solution

bus_v(:,k) = V_rec*psi; % bus voltage reconstruction

end

cur_re(:,k) = real(cur); cur_im(:,k) = imag(cur);

% step 3b: compute dynamics and integrate

flag = 2; %Dynamic calculation

pmech(:,k) = pmech(:,1); % constant mechanical input power

f = mac_em(0,k,bus,flag); % dynamics calculation

if k ~=k_switch(3)

j = k+1;

% following statements are predictor steps

mac_ang(:,j) = mac_ang(:,k) + h*dmac_ang(:,k);

mac_spd(:,j) = mac_spd(:,k) + h*dmac_spd(:,k);

edprime(:,j) = edprime(:,k) + h*dedprime(:,k);

eqprime(:,j) = eqprime(:,k) + h*deqprime(:,k);

flag = 1;

mach_ref(j) = 0;

f = mac_em(0,j,bus,flag);

psi = psi_re(:,j) + jay*psi_im(:,j);

if k >= k_switch(2)

cur = Y_red_pf*psi;

elseif k >= k_switch(1)

cur = Y_red_f*psi;

else

cur = Y_red*psi;

end

cur_re(:,j) = real(cur); cur_im(:,j) = imag(cur);

pmech(:,j) = pmech(:,k);

flag = 2;

f = mac_em(0,j,bus,flag);

% following statements are corrector steps, "Trapiziodal rule"

mac_ang(:,j) = mac_ang(:,k) +...

h*(dmac_ang(:,k)+dmac_ang(:,j))/2.;

mac_spd(:,j) = mac_spd(:,k) +...

h*(dmac_spd(:,k)+dmac_spd(:,j))/2.;

edprime(:,j) = edprime(:,k) +...

h*(dedprime(:,k)+dedprime(:,j))/2.;

eqprime(:,j) = eqprime(:,k) +...

h*(deqprime(:,k)+deqprime(:,j))/2.;

end

end

t = [0:stepsize:t_switch(4)]'; % time

If you have the dynamical model of your system and simulated it using Euler method, the easiest way is to do the following:

1. In your time domain simulation identify the voltage parameter for the bus that you need to apply the three phase fault to.

2. Make simple for-loop to determine the duration of fault. Usually we take a three cycles fault.

3. Within the loop make the bus voltage=0

4. Run the simulation

I also agree with Syed comment, but the mentioned method is suitable for disturbances associated with no topological, or structural changes in the system.

Dear Dr. EL-Shimy ,

First of all, I'd like to thank you for your prompt and useful answer and suggestions. It gave me an insight into fault occurrence simulation.

As my academic major is control engineering, I'm not familiar with the power network simulation.

By the way, I considered your book; It was great.

I would like to apply the control strategy to the popular WSCC 3-machine, 9-bus power system. I have the admittance matrix of it's Internal node model which corresponds to the pre-fault state.

the question is how should I construct the during-fault and the post-fault matrices according to the different values of landa?

As you know, landa is the fraction of the transmission line to the left of the fault. If landa=0, the fault is on the bus bar of generator #1, and landa =0.5 puts the fault in the center point of the transmission line between generator #1 and generator #2, and so on.

Thank you in advance.

Dear Syed,

Could you please explain more?

I think this is true for landa equals to zero. Am I right?

The voltage is assumed to be constant . The state variables are: relative rotor angle, angular velocity, relative input mechanical power and the steam valve opening.

Dear Saeed,

I truly appreciate your feedback; thank you. If the construction of the matlab model of the system is not among the main tasks of your study & you are also familiar with matlab codes, I suggest you to use either the Power System Analysis (PST) toolbox, or the Power System Analysis Toolbox (PSAT) in the basic simulation requirements. They are open-source software tools i.e. you can change thier original codes, and/or add new models of components or controllers. With these facilities, more efficient and focused studies can be made in a reasonable time provided that you have a sufficient background of understanding and using the matlab codes. Good luck

Dear Dr. EL-Shimy ,

You are welcome. Really, I should do some optimization problem and I have written some other programs in the same code. So, I would like to do the transient stability analysis within m-file codes.

I should say that I'm familiar with matlab code, but as I'm new to power systems, It think that it will take long to recognize the toolbox.

Do you mean that we can construct the during-fault and the post-fault matrices according to the different values of landa within (PSAT) automatically?

The main challenge is the construction of during-fault and the post-fault admittance matrices, and how landa should be considered in matrix construction?

I greatly appreciate the information you have provided me with on the topic .

Dear Saeed,

Sorry couldn't back to you quickly. As per your problem, you will need to do the following:

1. Since you need to study the affect of faults at different positions of the transmission line, your resistance and inductance is a measure of distance, i.e. given in (/m).

2. Now consider the fault location as a imaginary bus and follow the steps that I have stated in my previous answer.

3. In this way, each fault location will constitute a different system with different system parameters.

Hope this will answer your query.

Good luck.

Dear Saeed,

For a specified fault loaction according to your choice, either the PST, or the PSAT can perform the whole time-domain simulation i.e. construction of the models, solving the ADEs of the system, and execute and plot the results. They also have some cobtrollers within their libraries. You may extract the state space matrices of the system, then perform controller tuning based on your control criteria, then edit the parameters of the controller(s) within any of the mentioned software tools. Finally, inspect the relative performances. I am sure that you will find the PSAT simple in use, and helpful even if you didn't use its command line capabilities.

In my profile, you will find some relevant publications of simulation and control design.

Good luck

Dear Syed,

The method you mentioned is straightforward and excellent, but it is limited to situations of temporary faults where no tripping is required; right?

Yes Dr. Shimy, the method is for temporary faults. I suppose what Saeed is trying to do is study the affect of these different faults and design a unified optimal controller for his system. As per your suggestions, PST can definitely perform enhanced fault studies but since he was working on an m-file, I had to suggest an easy method that can provide him atleast a starting point for his work.

Indeed your method for fault simulation is acceptable for evaluating the impact of controllers, unless disconnection of one or more elements is desired.

Dear All,

Thank you for all your comments and also my apologies for the slow reply.

As dear Syed said, I want to study the affect of these short-circuit faults and design a unified optimal controller for my system.

Also he said thet " the resistance and inductance is a measure of distance, i.e. given (/m) "

But, the question is how to construct the imaginary bus corresponding to the fault location. I have a 3*3 Admittance matrix for my 3-machine 9-bus power network. what changes should be done on this matrix?

I hope I've not bothered you with my questions.

Anyone has a Matlab code or script to reconstruct a short circuits signal to be used for Discrete Wavelet Transform