MATLAB Central Discussions - Join the conversation!
Contenido principal

Resultados de


Hassan Ali
Hassan Ali
Última actividad el 27 de Dic. de 2023

Modelling BLDC motor losses

Hello Everyone,
I want to model an electric vehicle in simscape electrical, I have few quiries regarding it.
  1. I have modelled an 3-Phase inverter, and used ee_getPowerLossSumary to get switching losses, and the results are okay as i was expecting, but now is there any other function to calculate conduction losses?
  2. I want to connect a BLDC motor, I have few parameters from manufacturers datasheet, but not all the parameters, so what would the best way according to your understanding, to model motor losses (Copper + Core).
I'm trying to calculate major fundamental losses of an EV. Looking for your inputs on this.
Thank you!
Kindly help me correct this code to function properly. I am just learning MATLAB. i cannot get the output in abc frame. This is the code:
%----------- Define input and state parameters-----------------------------
clc
v_dc = 350; % DC input voltage in V
m = 0.841; % modulation index
C = 4000e-6; % DC buss capacitance in uf
L_1 = 2.5e-3; % Inverter side inductance in mH
L_2 = 2.5e-3; % Load side inductance in mH
L = 0; % load inductance
C_f = 10e-6; % filter capacitance in uf
R_f = 0.7; % damping resistance in ohms
R_L = 20; % load resistance in ohms
f_s = 10e3; % switching frequency
f = 60; % System frequency
R_s = 0.01; % Capacitance of the DC circuit
I_d = 8.594; % steady state current
w = 2*pi*f; % System angular Frequency
% Define initial steady state values
v_c = 349.4; i_d = 8.594; i_q = -0.213; v_df = 285; v_qf = -120; i_Ld = 8.594; i_Lq = 0.85;
%------------------S V P W M Generator-------------------------------------
% Define reference vector Uref
U_mag = m*v_dc/2; % Magnitude of Uref
% Define switching vectors
U1 = [v_dc/2;0]; % Vector Q1
U2 = [v_dc/4;sqrt(3)*v_dc/4]; % Vector Q2
U3 = [-v_dc/4;sqrt(3)*v_dc/4]; % Vector Q3
U4 = [-v_dc/2;0]; % Vector Q4
U5 = [-v_dc/4;-sqrt(3)*v_dc/4]; % Vector Q5
U6 = [v_dc/4;-sqrt(3)*v_dc/4]; % Vector Q6
% Define sector angles
theta1 = pi/6;
theta2 = pi/2;
theta3 = 5*pi/6;
theta4 = 7*pi/6;
theta5 = 3*pi/2;
theta6 = 11*pi/6;
% Define duty cycles for each switch using a for loop
for t=0:1/f_s:1/f % Time variable from 0 to one cycle of system frequency with steps of switching frequency
U_phase = w*t; % Phase of Uref (t is time variable)
U_alpha = U_mag*cos(U_phase); % Alpha component of Uref
U_beta = U_mag*sin(U_phase); % Beta component of Uref
if (0 <= U_phase) && (U_phase < theta1) % Sector 1
T1 = (sqrt(3)*U_beta + U_alpha)/(2*v_dc);
T2 = (-sqrt(3)*U_beta + U_alpha)/(2*v_dc);
T0 = 1 - T1 - T2;
d_a(round(t)+1) = T1 + T0/2;
d_b(round(t)+1) = T2 + T0/2;
d_c(round(t)+1) = T0/2;
elseif (theta1 <= U_phase) && (U_phase < theta2) % Sector 2
T3 = (sqrt(3)*U_beta - U_alpha)/(2*v_dc);
T2 = (sqrt(3)*U_beta + U_alpha)/(2*v_dc);
T0 = 1 - T3 - T2;
d_a(round(t)+1) = T0/2;
d_b(round(t)+1) = T2 + T0/2;
d_c(round(t)+1) = T3 + T0/2;
elseif (theta2 <= U_phase) && (U_phase < theta3) % Sector 3
T3 = (sqrt(3)*U_beta - U_alpha)/(2*v_dc);
T4 = (-sqrt(3)*U_beta - U_alpha)/(2*v_dc);
T0 = 1 - T3 - T4;
d_a(round(t)+1) = T0/2;
d_b(round(t)+1) = T0/2;
d_c(round(t)+1) = T3 + T0/2;
elseif (theta3 <= U_phase) && (U_phase < theta4) % Sector 4
T5 = (-sqrt(3)*U_beta + U_alpha)/(2*v_dc);
T4 = (-sqrt(3)*U_beta - U_alpha)/(2*v_dc);
T0 = 1 - T5 - T4;
d_a(round(t)+1) = T5 + T0/2;
d_b(round(t)+1) = T0/2;
d_c(round(t)+1) = T4 + T0/2;
elseif (theta4 <= U_phase) && (U_phase < theta5) % Sector 5
T5 = (-sqrt(3)*U_beta + U_alpha)/(2*v_dc);
T6 = (sqrt(3)*U_beta + U_alpha)/(2*v_dc);
T0 = 1 - T5 - T6;
d_a(round(t)+1) = T5 + T0/2;
d_b(round(t)+1) = T6 + T0/2;
d_c(round(t)+1) = T0/2;
elseif (theta5 <= U_phase) && (U_phase < theta6) % Sector 6
T1 = (sqrt(3)*U_beta + U_alpha)/(2*v_dc);
T6 = (sqrt(3)*U_beta - U_alpha)/(2*v_dc);
T0 = 1 - T1 - T6;
d_a(round(t)+1) = T1 + T0/2;
d_b(round(t)+1) = T0/2;
d_c(round(t)+1) = T6 + T0/2;
end
end
%-------------------------Define system matrices---------------------------
% Create Three-phase SVPWM VSI Inverter
% System matrix Nx-by-Nx matrix
A = [-1/(C*R_s),-sqrt(3)*m/(2*C),0,0,0,0,0;
sqrt(3)*m/(3*L_1),-R_f/(3*L_1),w,-1/(2*L_1),-sqrt(3)/(6*L_1),-R_f/(3*L_1),0;
0,-w,-R_f/(3*L_1),-sqrt(3)/(6*L_1),-1/(2*L_1),0,R_f/(3*L_1);
0,1/(2*C_f),-sqrt(3)/(6*C_f),0,w,-1/(2*C_f),sqrt(3)/(6*C_f);
0,sqrt(3)/(6*C_f),1/(2*C_f),-w,0,-sqrt(3)/(6*C_f),-1/(2*C_f);
0,R_f/(3*(L_2+L)),0,1/(2*(L_2+L)),sqrt(3)/(6*(L_2+L)),((-3*R_L-R_f)/(3*(L_2+L))),w;
0, 0, R_f/(3*(L_2+L)), -sqrt(3)/(6*(L_2+L)), 1/(2*(L_2+L)), -w, ((-3*R_L-R_f)/(3*(L_2+L)))];
% Define input matrix
B = [1/(C*R_s),-sqrt(3)*i_d/(2*C);d_a*v_dc,(sqrt(3)*v_c)/L_1;d_b*v_dc,0;d_c*v_dc,0;0,0;0,0;0,0]; % Nx-by-Nu input matrix
% Define output matrix
C = [0 1 0 0 0 0 0; % Ny-by-Nx matrix
0 0 1 0 0 0 0;
0 0 0 1 0 0 0;
0 0 0 0 1 0 0;
0 0 0 0 0 1 0;
0 0 0 0 0 0 1];
% Feedthrough matrix
D = zeros(6, 2); % Ny-by-Nu matrix
% create state-space model object
sys = ss(A,B,C,D);
% Define initial conditions and input
x0 = [v_c; i_d; i_q; v_df; v_qf; i_Ld; i_Lq]; % Initial state vector
t = 0:1e-6:0.5; % Time vector for simulation
u = repmat([v_dc;m],1,length(t)); % repeat u for each time step
% Simulate the system
[y, ~, x] = lsim(sys, u, t, x0);
% Extract the states
v_c_sim = x(:, 1);
i_d_sim = x(:, 2);
i_q_sim = x(:, 3);
v_df_sim = x(:, 4);
v_qf_sim = x(:, 5);
i_Ld_sim = x(:, 6);
i_Lq_sim = x(:, 7);
% Extract the outputs
v_abc_sim = y(:, 1:3);
i_abc_sim = y(:, 4:6);
v_dq_sim = y(:, 4:5);
i_dq_sim = y(:, 2:3);
% Plot the variables
figure;
subplot(4, 2, 1);
plot(t, v_c_sim);
xlabel('Time');
ylabel('v_c');
title('Capacitor Voltage');
subplot(4, 2, 2);
plot(t, i_d_sim);
xlabel('Time');
ylabel('i_d');
title('d-Axis Current');
subplot(4, 2, 3);
plot(t, i_q_sim);
xlabel('Time');
ylabel('i_q');
title('q-Axis Current');
subplot(4, 2, 4);
plot(t, v_df_sim);
xlabel('Time');
ylabel('v_df');
title('d-Component Filter Voltage');
subplot(4, 2, 5);
plot(t, v_qf_sim);
xlabel('Time');
ylabel('v_qf');
title('q-Component Filter Voltage');
subplot(4, 2, 6);
plot(t, i_Ld_sim);
xlabel('Time');
ylabel('i_Ld');
title('d-Axis Load Current');
subplot(4, 2, 7);
plot(t, i_Lq_sim);
xlabel('Time');
ylabel('i_Lq');
title('q-Axis Load Current');
% Perform coordinate transformation from dq frame to abc frame for currents
i_a_sim = cos(w*t)*i_d_sim - sin(w*t)*i_q_sim;
i_b_sim = cos(w*t - 2*pi/3)*i_d_sim - sin(w*t - 2*pi/3)*i_q_sim;
i_c_sim = cos(w*t + 2*pi/3)*i_d_sim - sin(w*t + 2*pi/3)*i_q_sim;
% Perform coordinate transformation from dq frame to abc frame for voltages
v_a_sim = cos(w*t)*v_df_sim - sin(w*t)*v_qf_sim;
v_b_sim = cos(w*t - 2*pi/3)*v_df_sim - sin(w*t - 2*pi/3)*v_qf_sim;
v_c_sim = cos(w*t + 2*pi/3)*v_df_sim - sin(w*t + 2*pi/3)*v_qf_sim;
Many thanks
Ver publicación completa
I'm trying to simulate EV drive system using FEM PMSM and IGBT block of simscape electrical.
The simulation results (magnitude of the current in time domain) is quite similar with the experimental results.
But in frequency domain analysis using FFT (fast fourier transform), amplitude of current harmonics(specially 5, 7 th of fundamental frequency) are way to different.
Is there any way that I could increase my simulation accuracy?
Go to top of page