Questions on Simulation and Sign Convention about VSCShunt

[YiFeeiWang]

Dear Developers of ANDES,

I hope you are doing well, and thank you for your great project!

When I tried to examine the 'kundur_vsc.xlsx' case ANDES provided, the power flow analysis was successful, but the simulation would stop at the specified fault instant and report "Time step reduced to zero. Convergence is not likely". Neither reducing the step size or the number of iterations helped. Similarly, I built a simple test case myself where a DC link connects two AC areas with VSCs as intefaces. The load flow analysis was also successful, and the system could be simulated if disturbances were not added. However, I got the same convergence issue when I tried to add another load by toggling, to alter the load directly, or to toggle some lines. Were such behaviours expected? Could the lack of dynamic models for VSCs be the reason?

Thank you again and best regards,
Yifei

[YiFeeiWang]

Dear ANDES Developers,

I found that with certain control settings and network configuration it is still possible to simulate the hybrid AC-DC system. However, there might exist a more fundamental problem: the ExtAlgeb part of VSCShunt might have not followed the sign conventions. For example, I believe psh should correspond to the active power flowing from the AC bus to the interal shunt bus (the one with vsh and ash), so e_str of a should be 'psh' instead of '-psh'. In the power flow solutions, this even made the total generation smaller than the total load in some cases. Could you please have a look?

Best,
Yifei

[cuihantao]

Thanks for reporting it. I will check it over the weekend.