six phase motor - inprogress #267
Conversation
|
sixphasePMSM_ode (1).pdf |
| I_SY_IDX = 3 | ||
| CURRENTS_IDX = [0, 1, 2, 3] | ||
| CURRENTS = ["i_sd", "i_sq", "i_sx", "i_sy"] | ||
| VOLTAGES = ["u_sd", "u_sq", "u_sx", "u_sx"] |
There was a problem hiding this comment.
Typo: last voltage component should be u_sy
| state[self.I_SD_IDX], | ||
| omega * state[self.I_SQ_IDX], | ||
| u_dqxy[0], | ||
| state[self.I_SQ_IDX], | ||
| omega * state[self.I_SD_IDX], | ||
| omega, | ||
| u_dqxy[1], | ||
| state[self.I_SX_IDX], | ||
| omega * state[self.I_SY_IDX], | ||
| u_dqxy[2], | ||
| state[self.I_SY_IDX], | ||
| omega * state[self.I_SX_IDX], | ||
| u_dqxy[3] |
There was a problem hiding this comment.
Order of these regressors is rather unsorted. Please implement the same ordering as in other motors, i.e. group the current components, the voltage components and the current * omega components.
The same reordering must then be applied to the parameter matrix (self._model_constants).
| Called internally when the motor parameters are changed or the motor is initialized. | ||
| """ | ||
| mp = self._motor_parameter | ||
| self._model_constants = np.array([ |
There was a problem hiding this comment.
It would be helpful to specify via comment which matrix column belongs to which regressor (please see other motor types for reference).
| Returns: | ||
| The converted quantities in the dq representation like ''[i_sd, i_sq, i_sx, i_sy, i_s0+, i_s0-]''. | ||
| since 2N topology is considered (case where the neutral points are not connected) i_s0+, i_s0- will not be taken into account | ||
| """ | ||
| t_vsd = 1/ 3 * np.array([ | ||
| [1, -0.5, -0.5, 0.5 * np.sqrt(3), -0.5 * np.sqrt(3), 0], | ||
| [0, 0.5 * np.sqrt(3), -0.5 * np.sqrt(3), 0.5, 0.5, -1], | ||
| [1, -0.5, -0.5, -0.5 * np.sqrt(3), 0.5 * np.sqrt(3), 0], | ||
| [0, -0.5 * np.sqrt(3), 0.5 * np.sqrt(3), 0.5, 0.5, -1], | ||
| [1, 1, 1, 0, 0, 0], | ||
| [0, 0, 0, 1, 1, 1] |
There was a problem hiding this comment.
For me this seems inconsistent: the function doc says that i_s0 components will not be considered, but the subsequent matrix computation does still compute them / contains the corresponding lines. I think it should be possible to:
- either get rid of these lines if the definition is to only hold for the 2N topology,
OR - cover the general case, whereas it might as well be desired to simulate a 1N topology.
| tp_alphaBetaXY = np.array([ | ||
| [cos, sin, 0, 0, 0, 0], | ||
| [-sin, cos, 0, 0, 0, 0], | ||
| [0, 0, cos, sin, 0, 0], | ||
| [0, 0, -sin, cos, 0, 0], | ||
| [0, 0, 0, 0, 1, 0] | ||
| [0, 0, 0, 0, 0, 1] | ||
| ]) |
There was a problem hiding this comment.
The first diagonal block entry has a positive sense of rotation, but the second one should have a negative. (As of now, both are coded the same).
For easier comprehension, I would suggest to define the 2D rotation matrix first as a function T(epsilon), and then compound this 6D rotation matrix from blkdiag[T(epsilon), T(-epsilon), eye(2)].
There was a problem hiding this comment.
The sense of rotation is still documented in the wrong way here
| u_sd V Direct axis voltage | ||
| u_sq V Quadrature axis voltage | ||
| u_sx V | ||
| u_sx V |
| [ 0, -mp['r_s'], 0, 0, 0, 1, 0, 0, 0, 0, mp['l_q'], 0, 0], | ||
| [-mp['psi_PM'], 0, -mp['r_s'], 0, 0, 0, 1, 0, 0, -mp['l_d'], 0, 0, 0], | ||
| [ 0, 0, 0, -mp['r_s'], 0, 0, 0, 1, 0, 0, 0, 0, -mp['l_y']], | ||
| [ 0, 0, 0, 0, -mp['r_s'], 0, 0, 0, 1, 0, 0, mp['l_x'], 0] |
There was a problem hiding this comment.
Looks very good to me. Good job!
| tp_alphaBetaXY = np.array([ | ||
| [cos, sin, 0, 0, 0, 0], | ||
| [-sin, cos, 0, 0, 0, 0], | ||
| [0, 0, cos, sin, 0, 0], | ||
| [0, 0, -sin, cos, 0, 0], | ||
| [0, 0, 0, 0, 1, 0] | ||
| [0, 0, 0, 0, 0, 1] | ||
| ]) |
There was a problem hiding this comment.
The sense of rotation is still documented in the wrong way here
| t1 = rotation_matrix(epsilon) | ||
| t2 = rotation_matrix(-epsilon) | ||
| tp_alphaBetaXY = np.block([ | ||
| [t1, np.zeros(t1.shape[0],t1.shape[1])], | ||
| [np.zeros(t2.shape[0],t2.shape[1]), t2], | ||
| ]) |
There was a problem hiding this comment.
scipy provides a command that should make this snippet more clear
https://docs.scipy.org/doc/scipy/reference/generated/scipy.linalg.block_diag.html
Development of six phase machines - new feature