I am doing a project with PV array in Simulink. When I run small section of the total system i do not get an error. But when i integrate full system(my system model`s are completely okay) I get an algebraic error saying
"An error occurred while running the simulation and the simulation was terminated
Caused by:
Simulink cannot solve the algebraic loop containing 'system_approach_first/PV Array/Diode Rsh/Product5' at time 0.0 using the TrustRegion-based algorithm due to one of the following reasons: the model is ill-defined i.e., the system equations do not have a solution; or the nonlinear equation solver failed to converge due to numerical issues."
Can anyone say me the route to solve this irritating problem?
I have made simulation sample time very small but the problem still on. Also i reduced tolerance to a very small amount
to solve this problem, you must either change the sampling step (e.g. Ts=1e-4) or change the solver parameter using (ode 45 ) or (ode23tb) otherwise you try with the fixed step.
Kind regards
Dear Zarzisur Rahman Rony,
I suggest you to see links and and attached file on topic.
https://www.mathworks.com/matlabcentral/fileexchange/71682-pv-array-with-algebraic-loop-broken
https://www.ijarse.com/images/fullpdf/1420817243_183.pdf
https://cyberleninka.org/article/n/1256193
https://www.jnrd.info/2019/12/10-5027-jnrd-v9i0-06/
Best regards
Dear Nasreddine Attou Sir,
I followed your suggestions. Unfortunately my model did not run still then. I used discrete mode in power gui with 1e-6 sample time. Algebraic loop time constant was initially 5e-5s. I tried to reduce it to 1e-6 but no result. Then i increased time constant to 1e-4s. then simulation started and gave good result. I still do not know why increasing time constant makes the simulation to run!!!@
Dear Mohamed-Mourad Lafifi Sir,
Thank you for the materials. I am using built in model. I read the materials and followed that Gert Kruger`s model does not work with discrete simulation. I am working with discrete sample. Also i find that a unit delay around error point may solve the problem. But i do not know the sample time of unit delay block and where to put the unit delay block in the built in PV array model
Dear Zarzisur Rahman Rony,
I suggest you to see the following links.
https://github.com/motahhir/MATLAB-Simulink-Model-of-Photovoltaic-Cell-Panel-and-Array-
https://github.com/motahhir/PV-panel-under-shading-and-implementation-of-DMPPT
Zarzisur Rahman Rony ,Matlab generally takes four digits after the decimal point, and in this case the sampling time Ts is in the denominator of the equation, when we run the simulation with a Ts lower than 1e-4, i.e. 4 digits of value 0 after the decimal point, the system takes the value of denominator equal to 0, to this effect the calculation of the equation gives an infinite value which results in the non-convergence of the system (in my view)
Dear Zarzisur Rahman Rony, it should be enough to add a "memory" block (you can find it in Simulink's Discrete toolbox) where algebraic loop occurs (e.g. after "system_approach_first/PV Array/Diode Rsh/Product5") then try with/without setting "inherit sample time" option. This allows breaking algebraic loop and it should work.
Dear Moussa Labbadi Sir,
Thank you for your link of github simulation files. I am a follower of Saad Motahhir Sir`s work and the works are great..
Dear Nasreddine Attou Sir,
I did not know that matlab takes 4 digit after decimal point. So that can explain why my system worked.
Anyway I faced problem when my system goes to an abrupt change with 1e-4s time constant also. Then i found that PV panel voltage is fluctuating to positive to negative value(I used a input capaitor 1uF in PV terminal) during that abrupt system parameter change. Then i use input capacitor with high capacitance value (1000uF) and algebraic loop did not occur since then.
Dear Alessandro di Gaeta Sir,
Thank you for you valuable direction. It worked. i have connected a memory block after {Vd} from block and ticked inherit sample time. And it solved my problem in any kind of measurement filter time constant value(with using input capacitor of boost converter). But if i do not use input capacitor of boost converter( connected to PV array) then for very small measurement filter time constant (eg 1e-6) does not give any real result (though simulation run). But for measurement filter time constant value > 1e-4 , simulation run and give nice result ( even without input capacitor).
So, I think we should use memory block (shown in following figure) and use measurement filter time constant >1e-4s.
Best regards.
- Badges
- Science topic