Thank you for a very good series of videos on this subject! I have a question though: when implementing the PI-controllers in Simulink and using the ideal regulator formula P(1+I*1/s), how is K_v, K_i, K_p, and T_i related to P and I, respectively? Are K_v and K_i equal to P, or is K_p equal to P? And is I=1/T_i? I am just a bit confused about the names of all parameters :)
Thank you for your feedback. I am using the ISA (International Society of Automation) form of the PI(D) controller in my courses. Your assumption is correct, i.e. P(1+I*1/s) with P = K_p and I = 1/Ti. The correct implementation of the PI controller with feedforward, tracking and saturation in covered in Video 6.
Hello, I'm also having trouble understanding the form of the PI controller. For the current control: If P = Kp = 1/(L1+L2) and the form is P(1 + I*1/s), then what is the purpose of Ki = 2(L1+L2)/Tclp. in other words, where does Ki fit into the actual equation used to form the controller?
@ 21:38, the transfer function Iinv/Vinv={1/(L1+L2)}* 1/S. = Kp/S. But next you design PI controller where transfer changed entirely because ouput is now Vinv = C*(Iref-I). It seems confusing to me. Kindly comment.. same is true for @ 6:11
The first equation is the process (P): input Vinv,dq and output Iinv,dq. The current process output needs to be controlled. The second equation you mention is simply the controller (C): input control error Iinv,dq,SP - Iinv,dq and output Vinv,dq, i.e. a current controller. We have a standard closed-loop system. You should probably also consult video 3.
@@frankydebruyne9842 Thanks Sir! Video 3 tells everything. Your lectures are so natural and spontaneous, very normalized that anybody can understand. Really, your lectures are like art. THANKS
We design the current control based on only 1/s(L1+L2) dynamics, it is fine. However, how can we determine if there is an instability when you include the capacitor?
As long as you do not push bandwidth too far it should not be a problem. If you want more performance, you should go for more advanced control techniques than just PI control.
@@frankydebruyne9842 so if we push the bandwidth too far, it means that we have big big possibility to have to have a LCL resonant peak close to PWM harmonics?
SUPERB!!!!!!! Even reading an entire book did not give such a crystal-clear idea....
Thank you for the nice feedback !
Thank you so much for these wonderful videos! This topic interested me very much. Could you share your model for doing your exercises?
Unfortunately, the models and slides are for ECAM students only. I hope you understand. Thank you for the nice comments.
Thank you for a very good series of videos on this subject! I have a question though: when implementing the PI-controllers in Simulink and using the ideal regulator formula P(1+I*1/s), how is K_v, K_i, K_p, and T_i related to P and I, respectively? Are K_v and K_i equal to P, or is K_p equal to P? And is I=1/T_i? I am just a bit confused about the names of all parameters :)
Thank you for your feedback. I am using the ISA (International Society of Automation) form of the PI(D) controller in my courses. Your assumption is correct, i.e. P(1+I*1/s) with P = K_p and I = 1/Ti. The correct implementation of the PI controller with feedforward, tracking and saturation in covered in Video 6.
@@frankydebruyne9842 Thank you! :)
Hello, I'm also having trouble understanding the form of the PI controller. For the current control: If P = Kp = 1/(L1+L2) and the form is P(1 + I*1/s), then what is the purpose of Ki = 2(L1+L2)/Tclp. in other words, where does Ki fit into the actual equation used to form the controller?
@ 21:38, the transfer function Iinv/Vinv={1/(L1+L2)}* 1/S. = Kp/S. But next you design PI controller where transfer changed entirely because ouput is now Vinv = C*(Iref-I). It seems confusing to me. Kindly comment.. same is true for @ 6:11
The first equation is the process (P): input Vinv,dq and output Iinv,dq. The current process output needs to be controlled. The second equation you mention is simply the controller (C): input control error Iinv,dq,SP - Iinv,dq and output Vinv,dq, i.e. a current controller. We have a standard closed-loop system. You should probably also consult video 3.
@@frankydebruyne9842 Thanks Sir! Video 3 tells everything. Your lectures are so natural and spontaneous, very normalized that anybody can understand. Really, your lectures are like art. THANKS
We design the current control based on only 1/s(L1+L2) dynamics, it is fine. However, how can we determine if there is an instability when you include the capacitor?
As long as you do not push bandwidth too far it should not be a problem. If you want more performance, you should go for more advanced control techniques than just PI control.
@@frankydebruyne9842 so if we push the bandwidth too far, it means that we have big big possibility to have to have a LCL resonant peak close to PWM harmonics?
if the bandwidth is pushed too far, current control based on only 1/s(L1+L2) dynamics will be problematic.