Control Design for a Bidirectional Buck Boost Converter

Discussion in 'The Projects Forum' started by enyyubfa, Mar 24, 2009.

  1. enyyubfa

    Thread Starter New Member

    Mar 24, 2009
    4
    0
    Hi there guys,

    I'm trying to design a bidirectional buck boost converter as part of a larger three way solar charge controller project. The buck boost part will be connected to an energy storage device, which will probably be a 2F 5V ultracapacitor.

    The aim is to step down 9v source to charge the ultracap, and then when power is needed the converter will work in reverse and boost the ultracap voltage to provide the load. I plan to use average current mode control for the converter. The only problem is that i'm really bad at control theory.

    I've just about got the converter working in buck mode, but having problems getting it working in boost mode as i can't understand how to get the control setup for this and the capacitors complicate matters.

    At the moment i'm still at the simulation stage using PSim software.

    For the control design I have an outer voltage loop and a inner current loop. I've used two PI block with identical gains and time constants of 1 and 0.001sec. The control voltage is then fed into a comparator with a 20kHz 1V peak to peak sawtooth oscillator.

    Can anyone help please? I've included the circuit diagram.
     
  2. t_n_k

    AAC Fanatic!

    Mar 6, 2009
    5,448
    782
    I can't see the 9V source in your simulation schematic.

    Inner control (current) loops are often designed to be "faster" than the outer (voltage) loop - as in a DC machine control loop. Presumably the same may be true in your case, but you have assigned the same control parameters to both.
     
  3. enyyubfa

    Thread Starter New Member

    Mar 24, 2009
    4
    0
    The 9V supply is supposed to be on the left most side of the circuit. For this schematic i'm considering the 2F capacitor to be fully charged and want to operate the converter in boost mode transferring energy from the right to to the left.
     
  4. t_n_k

    AAC Fanatic!

    Mar 6, 2009
    5,448
    782
    OK - I understand about the 9V.

    My gut feeling is that the forward and reverse dynamics will be quite different and you will therefore need different tuning parameters in each case - that's potentially a problem if you want to do it with just one control loop - albeit with the inner and outer loop design you propose.

    You might consider running the reverse (boost) mode in a current control regime - where the current demand (from the supercap) is dictated by the load requirements. I don't know what your overall design brief is and whether such an option is acceptable.

    It's quite easy to imagine running the system as a controlled current drive in either direction - you'd just need a clamping control to choke off the buck mode when the target 5V is reached on the supercap.

    What other background resources are you using - is this a project for a course or just something of personal interest / necessity you want to design and build?
     
  5. enyyubfa

    Thread Starter New Member

    Mar 24, 2009
    4
    0
    Hmm

    Thanks, thats given me something to think about. Would you just need one current loop to have the bidirectional functionality? And where would you sense the voltage from?

    Its a project for my course and at the moment I'm just trying learn voltage and current mode control for single buck and boost converters. But I can't find any info on the control for bidirectional buck-boosts.

    What i'm aiming for is the converter to respond to changes at the load. If the PV panel can't provide enough power to the load then the ultracap will kick in and when there is an excess of power to the load the ultracap will be charged.

    BTW could you explain how you implent a voltage or current clamp in circuit?

    Thanks for the help. Much appreciated. :)
     
  6. t_n_k

    AAC Fanatic!

    Mar 6, 2009
    5,448
    782
    Hi again,

    I think one of the issues with the control scheme you attached (as the image in your original post), is that the primary feedback control parameter is the node voltage at the junction of the 0.8Ω and the 154μH.

    What is the 0.8Ω representing, BTW?

    The basic question is - what is it that you are attempting to control or regulate?

    In the case shown, you would appear to be attempting control the voltage at that node. The controlling source appears to be a fixed DC level.
    Note: some of the elements in the diagram have no values indicated so it's hard to see what is intended. The more info given - the easier to interpret.

    I suggested earlier that one strategy might be to think about controlling the current that flows in the required direction. At least thinking about that idea - even if it's purely academic - will give you some insights.

    To do this, I would proceed as follows :-

    1. Use the in-line current probe (at the aforementioned node) as the sole feedback signal.
    2. Have a single control PI loop - rather than an inner / outer case.
    3. Add some sort of limiter to the control path (before the PWM comparator) so that the PWM generator part is never in an unbounded state.
    4. Set your ramp (PWM gen) source as 50% duty - so that at 0V at the comparator +ve terminal gives 50% PWM duty - this is the neither bucking nor boosting "idle" condition.
    5. Control the power / current magnitude & flow direction by setting the control demand source as either a positive (forward buck case) or negative (reverse boost case) input at the (sole) summing junction.
    6. Play with the PI parameters until you can get stable current flow in either direction.

    Doing this exercise will be quite informative and hopefully lead on to other ideas. You could also consider doing a directional power controlled feedback simulation - say by multiplying Vsupercap*Isupercap=Psupercap and using this as the feedback signal.

    Once you have mastered a single control loop you may be better equipped to look at the more complex multi-loop case & /or the more challenging task of meeting the load sharing goal. BTW it may not be necessary to use a multi-feedback model - the simpler the better!

    On a slightly different tangent - while the goal is a working project this post may have been more appropriate for the homework forum.

    Good luck!:)
     
  7. wong_jennhwa

    New Member

    Aug 24, 2009
    1
    0
    hi, enyyubfa

    how your work's progress? done it?
    from the circuit diagram, the voltage sense supported is at the load voltage but not supercap volt. and for the energy sharing circuit, first you must create a system that able to control the current or power flow. hope it can give u some idea.:)
     
Loading...