请教，v0.5.3版本的odrive的foc.cpp里，这个乘数0.8是什么原理？

天下乌鸦一般黑 · 发表于 2022-6-1 23:50:22

float mod_scalefactor = 0.80f * sqrt3_by_2 * 1.0f / std::sqrt(mod_d * mod_d + mod_q * mod_q);
这一行代码。实在不懂这个0.8是啥意思，有大牛指点下我不。。。糊涂了
整个函数如下：

ODriveIntf::MotorIntf::Error FieldOrientedController::get_alpha_beta_output(
uint32_t output_timestamp, std::optional<float2D>* mod_alpha_beta,
std::optional<float>* ibus) {
if (!vbus_voltage_measured_.has_value() || !Ialpha_beta_measured_.has_value()) {
// FOC didn't receive a current measurement yet.
return Motor::ERROR_CONTROLLER_INITIALIZING;
} else if (abs((int32_t)(i_timestamp_ - ctrl_timestamp_)) > MAX_CONTROL_LOOP_UPDATE_TO_CURRENT_UPDATE_DELTA) {
// Data from control loop and current measurement are too far apart.
return Motor::ERROR_BAD_TIMING;
}
// TODO: improve efficiency in case PWM updates are requested at a higher
// rate than current sensor updates. In this case we can reuse mod_d and
// mod_q from a previous iteration.
if (!Vdq_setpoint_.has_value()) {
return Motor::ERROR_UNKNOWN_VOLTAGE_COMMAND;
} else if (!phase_.has_value() || !phase_vel_.has_value()) {
return Motor::ERROR_UNKNOWN_PHASE_ESTIMATE;
} else if (!vbus_voltage_measured_.has_value()) {
return Motor::ERROR_UNKNOWN_VBUS_VOLTAGE;
}
auto [Vd, Vq] = *Vdq_setpoint_;
float phase = *phase_;
float phase_vel = *phase_vel_;
float vbus_voltage = *vbus_voltage_measured_;
std::optional<float2D> Idq;
// Park transform
if (Ialpha_beta_measured_.has_value()) {
auto [Ialpha, Ibeta] = *Ialpha_beta_measured_;
float I_phase = phase + phase_vel * ((float)(int32_t)(i_timestamp_ - ctrl_timestamp_) / (float)TIM_1_8_CLOCK_HZ);
float c_I = our_arm_cos_f32(I_phase);
float s_I = our_arm_sin_f32(I_phase);
Idq = {
c_I * Ialpha + s_I * Ibeta,
c_I * Ibeta - s_I * Ialpha
};
Id_measured_ += I_measured_report_filter_k_ * (Idq->first - Id_measured_);
Iq_measured_ += I_measured_report_filter_k_ * (Idq->second - Iq_measured_);
} else {
Id_measured_ = 0.0f;
Iq_measured_ = 0.0f;
}
float mod_to_V = (2.0f / 3.0f) * vbus_voltage;
float V_to_mod = 1.0f / mod_to_V;
float mod_d;
float mod_q;
if (enable_current_control_) {
// Current control mode
if (!pi_gains_.has_value()) {
return Motor::ERROR_UNKNOWN_GAINS;
} else if (!Idq.has_value()) {
return Motor::ERROR_UNKNOWN_CURRENT_MEASUREMENT;
} else if (!Idq_setpoint_.has_value()) {
return Motor::ERROR_UNKNOWN_CURRENT_COMMAND;
}
auto [p_gain, i_gain] = *pi_gains_;
auto [Id, Iq] = *Idq;
auto [Id_setpoint, Iq_setpoint] = *Idq_setpoint_;
float Ierr_d = Id_setpoint - Id;
float Ierr_q = Iq_setpoint - Iq;
// Apply PI control (V{d,q}_setpoint act as feed-forward terms in this mode)
mod_d = V_to_mod * (Vd + v_current_control_integral_d_ + Ierr_d * p_gain);
mod_q = V_to_mod * (Vq + v_current_control_integral_q_ + Ierr_q * p_gain);
// Vector modulation saturation, lock integrator if saturated
// TODO make maximum modulation configurable
float mod_scalefactor = 0.80f * sqrt3_by_2 * 1.0f / std::sqrt(mod_d * mod_d + mod_q * mod_q);
if (mod_scalefactor < 1.0f) {
mod_d *= mod_scalefactor;
mod_q *= mod_scalefactor;
// TODO make decayfactor configurable
v_current_control_integral_d_ *= 0.99f;
v_current_control_integral_q_ *= 0.99f;
} else {
v_current_control_integral_d_ += Ierr_d * (i_gain * current_meas_period);
v_current_control_integral_q_ += Ierr_q * (i_gain * current_meas_period);
}
} else {
// Voltage control mode
mod_d = V_to_mod * Vd;
mod_q = V_to_mod * Vq;
}
// Inverse park transform
float pwm_phase = phase + phase_vel * ((float)(int32_t)(output_timestamp - ctrl_timestamp_) / (float)TIM_1_8_CLOCK_HZ);
float c_p = our_arm_cos_f32(pwm_phase);
float s_p = our_arm_sin_f32(pwm_phase);
float mod_alpha = c_p * mod_d - s_p * mod_q;
float mod_beta = c_p * mod_q + s_p * mod_d;
// Report final applied voltage in stationary frame (for sensorless estimator)
final_v_alpha_ = mod_to_V * mod_alpha;
final_v_beta_ = mod_to_V * mod_beta;
*mod_alpha_beta = {mod_alpha, mod_beta};
if (Idq.has_value()) {
auto [Id, Iq] = *Idq;
*ibus = mod_d * Id + mod_q * Iq;
power_ = vbus_voltage * (*ibus).value();
}
return Motor::ERROR_NONE;
}

复制代码

lonely9391 · 发表于 2022-6-2 08:59:59

只能开到80%的开度？

天下乌鸦一般黑 · 发表于 2022-6-2 22:57:02

lonely9391 发表于 2022-6-2 08:59
只能开到80%的开度？
(引用自2楼)

我也觉得是这个意思。但是这个80%是怎么来的，为啥不是81%。。。为啥不是90%。。。

天下乌鸦一般黑 · 发表于 2023-8-16 21:26:19

这个问题还是没有答案。
我感觉做了圆周功率限制，就没必要用这个0.8了吧

qwe2231695 · 发表于 2023-8-16 23:10:17

天下乌鸦一般黑发表于 2023-8-16 21:26
这个问题还是没有答案。
我感觉做了圆周功率限制，就没必要用这个0.8了吧 ...
(引用自4楼)

看注释就能知道，是用于判断力矩输出较大，Iq较大时，关闭积分计算，是积分抗饱和策略。

天下乌鸦一般黑 · 发表于 2023-8-16 23:20:49

qwe2231695 发表于 2023-8-16 23:10
看注释就能知道，是用于判断力矩输出较大，Iq较大时，关闭积分计算，是积分抗饱和策略。 ...
(引用自5楼)

大佬在啊。
抗积分饱和看懂了。
主要是这个0.8。我在学习（抄袭）它代码的时候，主要是看不懂这个0.8~很早的版本也有这个0.8

请教，v0.5.3版本的odrive的foc.cpp里，这个乘数0.8是什么原理？

阿莫论坛20周年了！感谢大家的支持与爱护！！