We are doing a project in neutral point potential balance in three level diode clamped inverter. We were struck with a few issues.
1.We use RL Load in closed loop when we get voltage Vabc as feedback and then it is converted to dq transformation.When compared Vdq with reference Vdq normally in drives d is taken as 1 and q is taken as 0.
2.We had a doubt that the same strategy is followed in our work or take d=1 and q=1 and also we need an explanation that why d is taken as 1 and q is taken as 0
I have also attached my simulation file.
Ok I think I might see what the problem is. If your controller is in the DQ or synchronous reference frame then all quantities should be DC or zero Hz in steady state.
If you run your simulation for 0.02 or 20msec then you should see one cycle. If the phase sequence and phase angle between the inverter output voltages and the PLL is incorrect you will see 100Hz instead of 0Hz.
You are using PLL which synchs to what the phase voltages or the line voltages?
Line-line and phase-neutral voltages are phase shifted by 30 deg so you must account for this.
PWM normally controls the pole voltage or the voltage measured from the inverter leg to the -ve of the DC bus. You are feeding back the line-line voltages which is the difference between each pole voltage.
Hence you have to ensure that the reference voltages the PLL represents are being used on the correct feedback quantities.
It appears you have both incorrect phase and phase sequence and thus you are generating a voltage reference to the PMW generator which are ~200Hz instead of 50 Hz.
To solve this generate the voltages in open-loop and ensure that you have one to one corresponds with the filtered voltages and the reference voltages supplied to the PWM Generator.
Your ABC/DQ produces alternating feedback which is then passed through the voltage controller and these too have their frequency even further increased by the DQ/ABC transform which are then passed to the PWM controller.
This will act like +ve feedback and your output frequency will increase. When I simulate at t=0.2 the voltages are already 250Hz. You did not see this because on the scopes you limited the buffering to the last 5000 points. Remove this when debugging.
I suggest first omit the controller and just generate the voltage references in open loop.
Then make sure the voltages you are measuring and feeding back are the ones to be controlled (ie phase on line voltages). In general inverters always generate the phase to neutral or pole voltages and you are feeding back line -line. This is ok but then account for phase shift in the ABC/DQ.
Hope this helps but it appears you have multiple issues to resolve. Draw these out by hand and do one step at a time.
Thank you for your answer sir i have struck with this point for last two weeks but finally i get the clear view of my problems.I hope your sugge
stions will help to complete my work sir.
It is important to note that the DQ or two axis theory is derived using the phase voltage and current s as per a STAR connected system.
In STAR line and phase currents are equal but line-line and phase voltages are not.
In a DELTA system the line-line and phase voltages are the same but not the line and phase currents.
This is often left out or overlooked in many books.
That is why you must develop any simulation one step at a time.
In closed loop system why d-axis quantity always taken 1 and q-axis always taken to 0 .May i follow the same assumption for my work.
It does not have to. With Vd=1 and Vq=0 you are fixing phase A voltage with d-axis. You can create phase shifting by varying Vq.
This is in effect what current controllers do.
In your simulation you are using a fixed PLL. What you need do to is to design a PLL which controls Vq such that it creates the Sin/Cos to lock the d axis with phase A. In open loop there will always be a phase difference between your reference and output voltages and your PLL cannot compensate for this. You need to design a dynamic PLL that forces your closed controller to align Vd with phase A. This may be done by forcing Vq=0. You have just set Vq=0.
I have run the simulation in open loop and when you use phase voltages as feed back the output voltages you can see the phase difference
Read up how to do this, it is quite a standard technique and you need to do it from first principles.
This problem is basically related to phase and coordinate transformation,
please find Clark and Park transformation,
1) ABC to alpha-beta and then 2) alpha-beta to d-q frame (with theta)
please find correct equations for the transformations. The theta can be obtained
using a simulink FUNC block that contains formula atan2 (u_beta, u_alpha)
with the coordinate transform voltage space vector join sides with the d-axis of the selected coordinate frame that is the reason d=magnitude q=0
if it is not in your simulation then possibly your theta is wrong
Check these two models for the diode clamped and flying capacitor three level inverter...lots of prob will solved
- Badges
- Science topic
- Similar topics
- Engineering
- Control Theory