• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Answers 2 answers
  • Subscribers 82 subscribers
  • Views 731 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TMS320F2800135: MTPA for IPMSM motor

Ken Anderson
Prodigy 40 points
Part Number: TMS320F2800135
Other Parts Discussed in Thread: C2000WARE-MOTORCONTROL-SDK, C2000WARE

Tool/software:

Hi I have a motor with a saliency ratio of about 2. I determined this by using an LCR meter and finding the minimum and maximum Ls through a full electrical rotation. TI documentation (SPRACF3) states that Ld > Lq, whereas all other sources state Lq > Ld. Using TI's Ld > Lq results in an MTPA kconst value that is negative (obj->kconst = 0.25 * (flux_Wb /(Ls_q_H - Ls_d_H));) which does not play well with the subsequent calculations when Is_ref_A is positive: obj->gconst = obj->kconst / obj->Is_ref_A; and then obj->angleCurrent_rad = acosf(obj->gconst - sqrtf(obj->gconst * obj->gconst + 0.5));

So, I swapped the motor parameters to make Lq the larger inductance value and Ld the smaller one, and this fixes the NaN issue. However, even though my IPMSM motor has a saliency ratio of around 2 which is fairly high, the inductance is quite small (0.000010uH to 0.000020uH) compared to rated flux (0.02 V/Hz or about 0.0032 Wb). The calculations result in gconst around 62, and this results in only about a one degree change in angle, and so extremely small values of Id_ref compared to Iq_ref. Is it true that my IPMSM simply will not benefit noticeably from MTPA due to high flux-to-inductance ratio?

  • 0
    TI__Guru** 116090 points
    Ken Anderson said:
    TI documentation (SPRACF3) states that Ld > Lq,

    It could be a typo. The MTPA algorithm in the example lab in motor control SDK uses the Lq>Ld to calculate the injection current or vector angle.

    Ken Anderson said:
    Is it true that my IPMSM simply will not benefit noticeably from MTPA due to high flux-to-inductance ratio?

    Not only MTPA, you may have to add a decoupling for this.

  • 0 in reply to Yanming Luo
    Prodigy 40 points

    Thanks for the response. I will work on implementing DQ decoupling. Is there any sample code for me to use as a reference?

  • 0 in reply to Ken Anderson
    TI__Guru** 116090 points

    Not yet for reference. You can find a lot of the papers on website, and implement it in your project.

  • 0 in reply to Yanming Luo
    Prodigy 40 points

    Understood. I am surprised this is not included in the base code since it is easy to add and provides a more 'complete' control solution with the potential for improved performance on some motors. It could be enabled with a compiler predefined symbol as is done with other features. Also, the code appears to be 'pre-wired' for it with the existing feed-forward terms. In any case, does this look correct to you?

    // obj->Vdq_ffwd_V.value[0] = 0.0f;
    // obj->Vdq_ffwd_V.value[1] = 0.0f;

    obj->Vdq_ffwd_V.value[0] = obj->speed_rps * objUser->Ls_q_H * obj->Idq_in_A.value[1];
    obj->Vdq_ffwd_V.value[1] = -obj->speed_rps * (objUser->Ls_d_H * obj->Idq_in_A.value[0] + objUser->motor_ratedFlux_Wb);

    Thanks.

  • +1 in reply to Ken Anderson
    TI__Guru** 116090 points

    The dq decoupling method has been implemented in the Universal Motor Lab as below, but it was not validated and tested on an IPM motor. You can refere to it.

    #if defined(MOTOR1_DECOUP)
    if((obj->flagMotorIdentified == true) && (obj->flagEnableDecoupleCtrl == true))
    {
    float32_t speed_rps = obj->speed_Hz * MATH_TWO_PI; // unit is rad/s

    obj->V_decoup_max_V = obj->adcData.VdcBus_V * objSets->V_decoup_sf;
    obj->V_decoup_min_V = -obj->V_decoup_max_V;

    // -We*Lq*Iq, SI unit: We(rad/s)*L(H)*I(A)
    obj->Vdq_decoup_V.value[0] =
    __fsat(-speed_rps * objSets->Ls_q_H * obj->Idq_in_A.value[1],
    obj->V_decoup_max_V, obj->V_decoup_min_V);

    // +We*(Ld*Id+flux), SI unit: We(rad/s)*L(H)*I(A), Flux(wb)
    obj->Vdq_decoup_V.value[1] =
    __fsat(speed_rps * (objSets->Ls_d_H * obj->Idq_in_A.value[0] + obj->flux_Wb),
    obj->V_decoup_max_V, obj->V_decoup_min_V);
    }
    else
    {
    obj->Vdq_decoup_V.value[0] = 0.0f;
    obj->Vdq_decoup_V.value[1] = 0.0f;
    }

    obj->Vdq_ffwd_V.value[0] += obj->Vdq_decoup_V.value[0];
    obj->Vdq_ffwd_V.value[1] += obj->Vdq_decoup_V.value[1];
    #endif // MOTOR1_DECOUP

    The detailed introduction about the lab can be found in the lab user’s guide as the links below.

    C2000WARE-MOTORCONTROL-SDK: https://www.ti.com/tool/C2000WARE-MOTORCONTROL-SDK

    Universal Project and Lab User’s Guide: https://www.ti.com/lit/spruj26

    Example lab project at the folder: C:\ti\c2000\C2000Ware_MotorControl_SDK_<version>\solutions\universal_motorcontrol_lab\f28002x

    or                                                         C:\ti\c2000\C2000Ware_MotorControl_SDK_<version>\solutions\universal_motorcontrol_lab\f28003x

    or                                                         C:\ti\c2000\C2000Ware_MotorControl_SDK_<version>\solutions\universal_motorcontrol_lab\f280013x

  • 0 in reply to Yanming Luo
    Prodigy 40 points

    Thank you for the code snippet. I downloaded the latest MotorControl SDK (C2000Ware_MotorControl_SDK_5_02_00_00) and don't see that in the universal_motorcontrol_lab for F280013x, but I will add what you posted. Thanks!