JP4053484B2 - Synchronous motor speed control device and magnetic pole position estimation method - Google Patents

Description

  The present invention relates to a synchronous motor speed control apparatus and a magnetic pole position estimation method for estimating an initial magnetic pole position of a rotor from only a rotation angle signal without using a magnetic pole position sensor in a synchronous motor including a linear machine and a rotary machine. It is.

  In Patent Document 1, a voltage corresponding to an arbitrary initial magnetic pole position is applied to a synchronous motor, and the correct magnetic pole position is determined while correcting the magnetic pole position according to the rotation direction of the rotor and the current value of the motor. A magnetic pole position estimation method for searching is disclosed.

Japanese Patent Application Laid-Open No. 6-153576 (summary, others overall)

  The prior art is a trial and error iterative search method that estimates the true value of the initial magnetic pole position of the synchronous motor while approaching a narrow estimated error range from a wide estimated error range of the initial magnetic pole position. For this reason, in order to increase the estimation accuracy of the initial magnetic pole position, it is necessary to increase the number of operations, and there is a problem that the time required for estimation becomes long.

  In addition, the closer to the true value of the magnetic pole position, the clearer the difference in the electric current of the motor used for estimation, and it was not suitable for high-precision estimation.

  An object of the present invention is to provide a synchronous motor speed control device and a magnetic pole position estimation method capable of quickly estimating a highly accurate initial magnetic pole position.

In one aspect of the present invention, relational data of a plurality of q-axis currents (command value iq ^* or detection value iq) respectively corresponding to a plurality of magnetic pole position error angles (θ _err ) are expressed as error angle (θ _err ) vs. q-axis. Approximating a concave quadratic curve on the current (iq) coordinates, and calculating the true magnetic pole position error angle (θ _err ) based on the X coordinate indicating the minimum point of the quadratic curve .

In another aspect, the present invention approximates a quadratic curve using the least square method on the coordinates of error angle (θ _err ) vs. q-axis current (iq), and indicates the minimum point of the quadratic curve. Based on the above, the true magnetic pole position error angle (θ _err ) is calculated to correct the initial magnetic pole position.

In the preferred embodiment, the present invention performs a speed control or a driving operation based on position control using a speed command pattern or position command pattern prepared in advance. At this time, three-phase / two-phase coordinate conversion and three-phase conversion used for control are performed on the temporary dq axes. On the other hand, the actual position of the dq axis is unknown, but if the error angle (θ _err ) can be estimated, the temporary dq axis is rotated by this estimated error angle (θ _err ). It is possible to match the actual dq axes.

For estimation, the q-axis current (command value iq ^* or detection) is detected while changing the magnetic pole position error angle (θ _err ) for each operation with respect to the same speed command pattern or the same position command pattern prepared in advance. The value iq) is measured. The relationship of about four points measured is plotted on the XY coordinates with the horizontal axis being (θ _err ) and the vertical axis being q-axis current (command value iq ^* or detection value iq). This is approximated to a quadratic curve using the method of least squares, and the magnetic pole position error angle (θ _err ) is estimated from the x coordinate of the minimum point.

In the present invention, a true magnetic pole position error angle (θ _err) is obtained by quadratic curve approximation from a relational data of a limited number of magnetic pole position error angles (θ _err ) and q-axis current (command value iq ^* or detection value iq). ), It is possible to estimate the magnetic pole position with high accuracy and speed control based on this with a small number of operations.

  Other objects and features of the present invention will become apparent from the description of the following examples.

  Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is an overall schematic configuration diagram of a speed control apparatus for a synchronous motor according to a first embodiment of the present invention. A three-phase AC voltage is applied from the power converter 2 to the synchronous motor 1. A three-phase alternating current I1 flowing through the synchronous motor 1 is detected by a current detector 3. Rotation angle detector 4 generates a pulse train corresponding to a rotation angle theta _M of the synchronous motor 1. Therefore, when counting the output pulses of the rotation angle detector 4 by an electrical angle calculation unit 5, it is possible to know the magnetic pole position theta _E of the motor rotor. Based on this detection electrical angle theta _E, the magnetic pole position estimation unit 6, a later-described magnetic pole position estimation method, the corrected (estimated) outputs an electrical angle theta ^. Based on the estimated electrical angle θ ^, the dq converter 7 converts the three-phase detection current I1 into the two-phase detection currents iq and id. The speed calculator 8 calculates the rotational speed ω _M from the rotational angle θ _M described above. The current controller 9 matches the q-axis current command value iq ^* and the d-axis current command value id ^* with the q-axis current detection value iq and the d-axis current detection value id, which are output values of the dq converter 7, respectively. Therefore, proportional-integral control is performed for iq ^* -iq and id ^* -id, and two-phase command voltages vq ^* and vd ^* are output. The subtractor 10 calculates the deviation between the speed command value ω _M ^* and the speed detection value ω _M, and the speed controller 11 calculates the q-axis current command value iq ^* . The three-phase conversion unit 12 converts the two-phase command voltages vq ^* and vd ^* into the three-phase command voltage V1 ^* using the estimated electrical angle θ ^. The changeover switch 13 performs a speed command value selection operation according to an instruction from the magnetic pole position estimation unit 6. The load 14 is driven by the synchronous motor 1 via the drive shaft 15.

In this configuration, the initial value of the detected electrical angle theta _E output from the electrical angle calculation unit 5 is assumed to be zero. That is, the output θ _E of the electrical angle calculation unit 5 immediately after the power supply of the speed control device including the magnetic pole position estimation unit 6 is turned on is zero, and then the electric motor 1 of the electric motor 1 according to the rotation angle Δθ _M of the electric motor 1. A value of θ _E = P × Δθ _M is output in proportion to the number P of pole pairs.

Figure 2 is an electric detection angle theta _E output from the electrical angle calculation unit 5 as an input value is an internal block diagram of a magnetic pole position estimation unit 6 to use the q-axis current command iq ^* from t1 transient data. The adder 34 outputs a corrected estimated electrical angle θ ^ obtained by adding the value of the magnetic pole position error angle θ _err to the detected electrical angle θ _E. The management unit 30 manages the entire magnetic pole position estimation process according to a flowchart described later. The command value generator 31 outputs a speed command pattern 50 as shown in FIG. The memory 32 stores the q-axis current command value iq ^* acquired at the timing of time t1 and the value of the magnetic pole position error angle θ _err at that time in accordance with the rules described later for the speed command pattern 50. The least square method calculation unit 33 performs a quadratic curve approximation calculation by the least square method when the relational data between θ _err and iq ^* stored in the memory 32 satisfies a predetermined condition to be described later, and provides a highly accurate magnetic pole The position error angle θ _err is calculated. When the detected q-axis current value iq is used instead of the q-axis current command value iq ^* , the detected value iq contains more noise components than the q-axis current command value iq ^*. is required.

FIG. 3 is another internal block diagram of the magnetic pole position estimation unit 6 having a function equivalent to that of FIG. 2 using an integral value of the absolute value of the q-axis current command iq ^* . The absolute value calculator 35 outputs the absolute value of the input signal iq ^* . The integrator 36 starts integration of the absolute value of the q-axis current command value iq ^* from the initial value = 0 simultaneously with the start of operation according to the operation pattern 50 of FIG. 4, and has an integral value clear function. If the magnetic pole position estimation unit 6 of FIG. 3 is used, iq ^* data over the entire operation period according to the operation pattern 50 of FIG. 4 is used. Therefore, compared with the case where the magnetic pole position estimation unit 6 of FIG. 2 is used, the estimation accuracy is improved under the condition that the value of iq ^* is very small. The reason why the absolute value calculator 35 is used is that when the speed pattern 50 of FIG. 4 is given as an input of the speed control device, the waveform 52 of the q-axis current (command iq ^* or detected value iq) is simply changed from t0 to t7. This is because when it is integrated up to zero, it becomes zero.

Here, the principle of estimating the magnetic pole position error angle θ _err in the preferred embodiment of the present invention will be described in detail with reference to FIGS. The outline is as described in the item of the disclosure of the invention.

FIG. 5 is a diagram illustrating the phase relationship between the true torque current I and the provisional q-axis current command iq ^* . First, an operation based on speed control or position control of the synchronous motor is performed using a speed command pattern or position command pattern prepared in advance. At this time, three-phase / two-phase coordinate conversion and three-phase conversion used for control are performed on the temporary dq axes shown in FIG. On the other hand, although the actual position of the dq axis is unknown, if the error angle θ _err can be estimated, the actual dq axis is rotated by the estimated error angle θ _err to thereby _calculate the actual dq axis. It is possible to match the −q axis.

For the estimation, the thrust required by the synchronous motor is the same for the same driving operation regardless of the value of the magnetic pole position error angle θ _err , and it is necessary at this time if the thrust is T _e and the thrust constant is K _T. The current I in the actual q-axis direction also uses a relationship that is equal to I = T _e / K _T. Further, as shown in FIG. 5, the relationship between the control q-axis current command value iq ^* existing on the temporary dq axis and the current I in the actual q-axis direction is expressed by equation (1). The

iq ^* = I / cos θ _err (1)
FIG. 6 is a graph showing the relationship between the magnetic pole position error angle θ _err and the q-axis current (command iq ^* or detected value iq). Since I is equal when performing the same driving operation, equation (1) is shown with I as a constant. For the same speed command pattern 50 or the same position command pattern 51 prepared in advance, the q-axis current command value iq ^* is measured while changing the magnetic pole position error angle θ _err for each operation, and the error angle θ _err is measured. And a graph plotting the relationship between the current iq ^* and a current function are concave functions. Therefore, it can be seen that the condition that the provisional dq axis and the actual dq axis coincide with each other is the minimum point of the current iq ^* . Based on this, in the present embodiment, the relational data of θ _err and iq ^* acquired for the same operation pattern is expressed by a quadratic curve (a, b and c are constants). Then, the true magnetic pole position error angle θ _err is estimated using Equation (3) representing the x-axis coordinate of the local minimum point.

f (x) = ax ² + bx + c (2)
x _min = −b / (2a) (3)
In the present embodiment, the change width of the magnetic pole position error angle θ _err that is changed for each operation is equal to 30 degrees, and the number of measurement points is four. Hereinafter, the estimation error in the measurement interval and the number of measurement points will be evaluated for several measurement cases with reference to FIG.

  FIG. 7 is a graph showing the relationship between measurement data points and estimated error angles. In (a) to (f) of FIG. 7, the value obtained from the measurement data is plotted by solving the equation (4) in which the current I in the real q-axis direction that can be regarded as a constant is 1 in the equation (1). Yes. Hereinafter, ideal data obtained from the equation (4) will be referred to as ideal measurement data or ideal measurement points.

iq ^* = 1 / cos θerr (4)
Therefore, the value of the magnetic pole position error angle θ _{err to} be obtained in FIG. 7 is zero in this case. Also, the curves drawn in the graphs (a) to (f) of FIG. 7 are quadratic approximate curves obtained by the method of least squares based on the four plot points that simulate the measurement points described above. Therefore, the value of the estimated error angle described in each graph coincides with the distance between the x-axis coordinate of the minimum point of the quadratic curve and the origin. Further, in each measurement case shown in FIG. 7, it is assumed that the range of θ _{err is in} the range of −90 degrees to +90 degrees including 0 degrees. However, at other values of θ _err , the speed control device is actually a divergent system. Accordingly, by detecting divergence by a method described later, the measurement is performed so that the range of θ _{err is in} the range of −90 degrees including +0 degrees to +90 degrees. The x coordinate of the measurement data at each θ _err is data_x [k] (equivalent to θ _err ), the y coordinate is data_y [k], k = 1 to 4, and data_x [1] to data_x [4] are arranged in ascending order. Assuming that Then, all the measurement results can be classified into any of cases 1 to 5 shown in Table 1.

  Each classification case shown in Table 1 includes mutually overlapping conditions, and the measurement results cannot be uniquely classified only with this table. However, it is possible to classify uniquely by using the flowchart of FIG. Table 2 shows ideal measurement data corresponding to the graphs (a) to (f) in FIG.

  Looking at FIG. 7 by case, the estimated error angle in case 1 in which data_y [1] to data_y [4] are arranged in descending order and case 5 in which data_y [1] to data_y [4] are arranged in ascending order are − It turns out that it becomes large with 20 degrees and 14 degrees. The reason for this is considered to be that a quadratic curve, which is a function having a minimum point, is approximated to monotonously decreasing or monotonically increasing data whose existence of the minimum point is unknown. In case 2 and case 4, the estimated error angles are -3.9 degrees and 0.7 degrees, respectively, which are greatly improved as compared to cases 1 and 5. The reason for this is considered that the data of (data_x [3], data_y [3]) is the minimum point in case 2, and the data of (data_x [2], data_y [2]) is the minimum point in case 4. . However, the estimated error angle of Case 2 is still close to 4 degrees. In case 3a and case 3b, the estimated error angles are 1.7 degrees and -1.1 degrees, respectively, and it is likely that the estimated error angle can be reduced stably. The reason is that under the conditions of Case 3a and Case 3b, in addition to the existence of the minimum point, the line connecting each measurement point has a shape close to a left-right symmetric shape, making it easier to approximate a quadratic curve that is a left-right symmetric system. It is thought that it was because of.

As described above with reference to the ideal measurement data of FIG. 7, in order to estimate the value of the magnetic pole position error angle θ _err with high accuracy using the least square method, in this embodiment, the measurement data is shown in Table 1. The data is acquired so as to satisfy the condition of case 3a or case 3b in the classification table. In the following, a detailed description will be given of a process from efficiently obtaining measurement data that satisfies such conditions to performing a least squares calculation.

FIG. 8 is a diagram schematically illustrating the configuration of the memory 32 illustrated in FIGS. 8, 60 is a read-reference pointer referenced by pointer name Dat_1, stores the address of the memory always stores the value of the minimum of the magnetic pole position error angle theta _err. A write pointer 61 is referred to by a pointer name of dat_w, and is used for measurement data write control. Reference numeral 62 denotes a memory for recording measurement data, which can record eight numerical values. Further, the addresses of the memory 62 are from mem_st to mem_end (= mem_st + 7), and it is possible to refer to adjacent data every time +1 or 1 is added to the address.

  Next, a method for making the memory having the configuration shown in FIG. 8 a ring memory so that data can be efficiently discarded and reacquired will be described. The ring memory is a memory structure that can access data corresponding to the minimum address of the memory with an address obtained by adding +1 to the maximum address of the memory, and is called in this way because the structure makes you imagine a circle. Similarly, in the ring memory, it is possible to access data corresponding to the maximum address of the memory with an address obtained by adding −1 to the minimum address of the memory. Although details of the ring memory are omitted, the realization thereof can be realized by performing address calculation of the read reference pointer dat_1 and the write pointer dat_w according to the flowchart shown in FIG.

  FIG. 9 is a flowchart showing an address calculation method for ring memory access. In the figure, reference numeral 70 denotes a start point of the address calculation, and the process proceeds to the determination process 71 unconditionally. In the determination process 71, it is determined whether dat_x> mem_end is satisfied for dat_x indicating dat_1 or dat_w. When it is established, the process 72 substitutes the calculation result of (dat_x−mem_end) + mem_st for dat_x, and the determination process 71 is performed again. In the determination process 71, when dat_x> mem_end is not satisfied, the process proceeds to the determination process 73. In the determination process 73, it is determined whether dat_x <mem_st is satisfied. When it is established, the process 74 substitutes the calculation result of mem_end− (mem_st−dat_x) for dat_x, and the determination process 73 is performed again. In the determination process 73, when dat_x <mem_st is not satisfied, the process proceeds to 75 and the process is terminated.

FIG. 10 is a flowchart showing an initial four-point data acquisition method in the magnetic pole position estimation unit 6 of the type that acquires the integral value of the absolute value of the q-axis current command value iq ^* . In the figure, reference numeral 130 denotes the starting point of the initial four-point data acquisition process, and the process proceeds to the initialization process 131 unconditionally. In the initialization process 131, first, the contact 2 of SW1 shown in FIG. 1 is selected. Accordingly, the operation pattern 50 of FIG. 4 generated by the command value generation unit 31 in the magnetic pole position estimation unit 6 can be given as the speed command value ω _M ^* of the speed control device. Next, the read reference pointer dat_1 and the write pointer dat_w are initialized with mem_st which is the start address of the memory 62. Next, the magnetic pole position error angle θ _err is initialized to θ _err = 0, the counter variable count is initialized to count = 0, and the flag variable flag_up is initialized to flag_up = + 1. When the above initialization process is completed, in the determination process 132, it is determined whether or not the counter variable count satisfies the condition of count <4. If measurement data for four points is not obtained, this condition is satisfied, and the process shifts to processing 133. In process 133, the integrator 36 is initialized with zero, and the process proceeds to process 134. The process 134 instructs the command value generation unit 31 in FIG. 3 to start generating the speed pattern. Next, in the determination process 135, it is determined whether or not normal rotation has occurred. If the deviation of the magnetic pole position is less than ± 90 degrees, in principle, rotation starts in the same direction as the speed command value ω _M ^* . Specifically, it is confirmed whether the rotation direction matches the command value. Become. If the directions match, that is, if the rotation is normal, the process proceeds to the determination process 136. In the determination process 136, it is determined whether or not the synchronous motor 1 has stopped. If it has stopped, the process proceeds to process 137. If not, the determination process 136 is repeatedly executed until it stops. In the process 137, the current value of the magnetic pole position error angle θ _err is saved in the memory corresponding to the address stored in the write pointer dat_w, and the process proceeds to the process 138. In the process 138, the integral value of the absolute value of the q-axis current command value iq ^* output from the integrator 36 in FIG. 3 is stored in the memory corresponding to the address obtained by adding +1 to the value (address) stored in the write pointer dat_w. Then, the process proceeds to process 139. In process 139, the value of <dat_w + flag_up × 2> is substituted for the write pointer dat_w in preparation for the next measurement, the counter variable count is incremented by 1, and the process proceeds to process 140. In process 140, the value of the magnetic pole position error angle θ _err is updated with θ _err + (30 ° × flag_up) in preparation for the next measurement, and the process proceeds to determination process 132. If the normal rotation is not determined in the determination process 135, the process proceeds to the determination process 141. In the determination process 141, it is determined whether or not the control of the synchronous motor 1 has diverged. When the divergence occurs, the process proceeds to process 142, the speed command value ω _M ^{* is set} to zero, and the process proceeds to determination process 143. In the determination process 143, if the counter variable count satisfies count ≧ 1, the process proceeds to process 144, and if not, the process proceeds to process 147. In the process 144, since the previous measurement has been rotated normally for the number of counts and data has been acquired, the value of the magnetic pole position error angle θ _err is updated by θ _err − {30 ° × (count + 1)} in preparation for the next measurement. The processing shifts to processing 145. As a result, it is possible to set θ _err that is 30 ° smaller than the minimum measured θ _err . In process 145, the flag variable flag_up is changed to -1. Thus, the process 140 is a process of subtracting the magnetic pole position error angle θ _err by 30 degrees from the value at that time. Further, the value of the write pointer dat_w is set to <dat_w− (count + 1) × 2>, the value of the read reference pointer dat_1 is updated to <dat_1 + count × 2>, and the process proceeds to the determination process 146. In the determination process 146, it is determined whether or not the synchronous motor 1 has been stopped. If the synchronous motor 1 has stopped, the process proceeds to the process 132, and if it has not stopped, the determination process 146 is repeatedly executed until it stops. When the counter variable count does not satisfy count ≧ 1 in the determination process 143 described above, it is a case where the counter variable count has not been normally operated at all. In process 147, the value of the magnetic pole position error angle θ _err is advanced by 180 degrees. The magnetic pole position is updated to a normal operating position, and the process proceeds to the process 146. When the control does not diverge in the determination process 141, it is a case where there is almost no torque because the deviation of the magnetic pole position is near +90 degrees or near 90 degrees, and the process proceeds to the determination process 149. In the determination process 149, when the counter variable count satisfies count ≧ 1, the process 150 and the process 151 are sequentially executed, and the process proceeds to the determination process 132. The process 150 is the same as the process 144, and the process 151 is the same as the process 145. When the counter variable count does not satisfy count ≧ 1 in the determination process 149, the value of the magnetic pole position error angle θ _err is advanced by 90 degrees in the process 153 to update the magnetic pole position where normal operation is possible or the control is divergent. The process proceeds to the determination process 132. In the determination process 132, when the condition of count <4 is not satisfied, the measurement data for four points is obtained, and the process proceeds to the node 154. The node 154 corresponds to the node 220 in the flowchart shown in FIG.

Thus, when the magnetic pole position error angle θ _err of the measurement data is data_x [k] and the corresponding q-axis current command value iq ^* is data_y [k], k = 1 to 4, data_x [1 arranged in ascending order. ] To data_x [4] can be taken out in order. That is, it is possible to sequentially retrieve data_x [1] to data_x [4] arranged in ascending order by accessing data obtained by adding +2 addresses by using the read reference pointer dat_1 as a base point. Similarly, by accessing data obtained by adding +2 to the address starting from data_1 + 1, the corresponding data_y [1] to data_y [4] can be sequentially extracted.

FIG. 11 is a flowchart showing a check and re-acquisition method after 4-point data acquisition in the magnetic pole position estimation unit 6 of the type that acquires the integral value of the absolute value of the q-axis current command value iq ^* . In this flowchart, it is checked whether or not the measured four points of data correspond to cases 3a and 3b shown in Table 2 above. If yes, nothing is done. Realize the function to do. In the process 221 following the node 220, the local variable k is initialized to 1, and the process proceeds to the process 222. In process 222, the value read from the memory at address <dat_1 + (2k-1)> is substituted into data_y [k], and the process proceeds to determination process 223. In the determination process 223, it is determined whether or not k ≧ 4 is established. If not, the process proceeds to process 224. In the process 224, the local variable k is incremented by 1, and the process proceeds to the process 222. When k ≧ 4 is satisfied in the determination process 223, the process proceeds to the determination process 225. In the determination process 225, it is determined whether or not the case 3a is met. If applicable, the process proceeds to the node 176, and if not, the process proceeds to the determination process 227. In the determination process 227, it is determined whether or not the case 3b is applicable. If applicable, the process proceeds to the node 176, and if not, the process proceeds to the determination process 228. In the determination process 228, it is determined whether or not the case 2 is applicable. If applicable, the process proceeds to process 229, and if not, the process proceeds to determination process 230. In the determination process 230, it is determined whether or not the case 1 is applicable. If applicable, the process proceeds to process 229, and if not, the process proceeds to determination process 239. In the process 229, the already measured maximum θ _err value read from the memory at the address <dat_1 + 6> is substituted for data_x [4], and the process proceeds to the process 231. In the process 231, the value of θ _err is set to data_x [4] + 30 °, and the process proceeds to the process 232. In process 232, the integrator 36 is initialized with zero, and the process proceeds to process 233. The process 233 instructs the command value generation unit 31 in FIG. 5 to start generating the speed pattern. Next, in the determination process 234, it is determined whether or not the synchronous motor 1 has been stopped. If the synchronous motor 1 has stopped, the process proceeds to the process 235. If not, the determination process 234 is repeatedly executed until the stop. In process 235, the value of the write pointer dat_w is updated with the value of the read reference pointer dat_1, and the process proceeds to process 236. In process 236, the current value of the magnetic pole position error angle θ _err is stored in the memory corresponding to the address stored in the write pointer dat_w, and the process proceeds to process 237. In the process 237, the integral value of the absolute value of the q-axis current command value iq ^* output from the integrator 36 of FIG. 3 is stored in the memory corresponding to the address obtained by adding +1 to the value (address) stored in the write pointer dat_w. Then, the process proceeds to process 238. In process 238, the value of the read reference pointer dat_1 is updated to <dat_1 + 2>, and the process proceeds to process 221. In the determination process 239, it is determined whether or not the case 4 is applicable. If applicable, the process proceeds to process 241; otherwise, the process proceeds to determination process 240. In the determination process 240, it is determined whether or not the case 5 is applicable. When it corresponds, it transfers to the process 241, and when it does not correspond, it transfers to the process 250 and is set as an estimation error. In the process 241, the already measured minimum θ _err value read from the memory at the address dat_1 is substituted for data_x [1], and the process proceeds to the process 242. In the process 242, the value of θ _err is set to data_x [1] −30 °, and the process proceeds to the process 243. In the process 243, the integrator 36 is initialized with zero, and the process proceeds to the process 244. The process 244 instructs the command value generation unit 31 in FIG. 5 to start generating the speed pattern. Next, in the determination process 245, it is determined whether or not the synchronous motor 1 has stopped. If the synchronous motor 1 has stopped, the process proceeds to the process 246. If not, the determination process 245 is repeatedly executed until it stops. In process 246, the value of the write pointer dat_w is updated with the value of <dat_1-2>, and the process proceeds to process 247. In process 247, the current magnetic pole position error angle θ _err is stored in the memory corresponding to the address stored in the write pointer dat_w, and the process proceeds to process 248. In the process 248, the integral value of the absolute value of the q-axis current command value iq ^* output from the integrator 36 in FIG. 3 is stored in the memory corresponding to the address obtained by adding +1 to the value (address) stored in the write pointer dat_w. , The process proceeds to processing 249. In process 249, the value of the read reference pointer dat_1 is updated to <dat_1-2>, and the process proceeds to process 221. As described above, the data of the four measurement points acquired by executing the flowchart of FIG. 11 corresponds to the cases 3a and 3b shown in Table 2 above.

Next, an algorithm using the least square method used when calculating the value of the magnetic pole position error angle θ _err that minimizes the value of iq ^* will be described. First, a method of solving a quadratic curve approximation by the least square method and a basic formula will be described.

In quadratic curve approximation by the method of least squares, when (x _k , y _k ) (k = 1, 2,... N: natural number) is used as measured data, d _k = y _k − (a * x _k ² _Dk is defined as + b * x _k + c), and the coefficients a, b, and c of the quadratic curve in Equation (2) that minimize the sum of the squares of d _k are _obtained . Since what is required now is the x-coordinate that gives the minimum point obtained by the above equation (3), only the necessary coefficients are a and b. This solution is generally known, and when determining Smy, Smxy, Smx2y, Smx, Smx2, Smx3, and Smx4 as shown in equations (5) to (11), coefficients a and b are expressed by equations (12) and (12), respectively. 13) It is given by the equation.

Therefore, the value of x _min to be obtained can be calculated by the equation (14) by substituting the equations (12) and (13) into the equation (3).

By placing (x _k , y _k ) = (data_x [k], data_y [k]) and substituting the equations (5) to (11) directly into the equation (14), the value of iq ^* is minimized. The magnetic pole position error angle θ _err can be calculated as x _min = θ _err . However, in this embodiment, when the change width of the magnetic pole position error angle θ _err to be changed for each operation is equal to Δθ, each measurement point (data_x [k], data_y [k]) (k = 1, 2,..., N (n: natural number)) are mapped to ((data_x [k] −data_x [1]) / Δθ, data_y [k]), respectively, and a least-squares operation is performed on the coordinates after the mapping. Therefore, the calculation amount and the number of handling digits are reduced.

FIG. 12 is a diagram for explaining the mapping of coordinates used for the calculation of the method of least squares. Hereinafter, the coordinates before mapping are referred to as a coordinate plane A, and the coordinates after mapping are referred to as a coordinate plane B. In the coordinate plane B, the value of the x coordinate of each measurement point is an integer starting from 0. Thereby, it is possible to greatly reduce the number of digits of the memory for storing the calculation results of the equations (8) to (11) used in the least square method calculation. Further, when the number of measurement data is known in advance as in this embodiment, the amount of calculation can be reduced by obtaining equations (8) to (11) in advance. n = 4, Δθ = 30, (x _k , y _k ) = ((data_x [k] −data_x [1]) / Δθ, data_y [k]) are substituted into equations (5) to (11). Thereby, the values of Smy, Smxy, Smx2y, Smx, Smx2, Smx3, and Smx4 on the coordinate plane B are expressed by equations (15) to (21).

  Next, when Expressions (15) to (21) are substituted into Expression (14), Expression (22) is obtained.

Since x _min obtained by the equation (22) is an x coordinate that gives a minimum point on the coordinate plane B, it is inversely mapped onto the original coordinate plane A. Since the inverse mapping of the x coordinate can be expressed by the equation (23), the magnetic pole position error angle that is the x coordinate that gives the minimum point on the coordinate plane A by substituting the equation (22) into the equation (23). The value of θ _err can be obtained as in equation (24).

(X _min × Δθ) + data_x [1] -------- (23)

Therefore, in the flowchart of FIG. 13 to be described next, after calculating Smy, Smxy, Smx2y by the equations (15) to (17), the value of the magnetic pole position error angle θ _err is calculated by calculating the equation (24). Calculated.

FIG. 13 is a flowchart of the least square method calculation following the flowchart of FIG. In the figure, a node 270 means the same node as the lowermost node 226 in FIG. 11, and shifts to the process 271 unconditionally. In the process 271, the minimum θ _err value in the four measurement data read from the memory at the address dat_1 is substituted into data_x [1], and the process proceeds to the process 272. In process 272, Smy, Smxy, and Smx2y are calculated from equations (15) to (17), and the process proceeds to process 273. In the process 273, the value of the magnetic pole position error angle θ _err is calculated based on the equation (24), and the process proceeds to the process 274. In process 274, the contact 1 of SW1 is selected, and the speed control device returns the synchronous motor 1 to a state in which the speed can be controlled based on the speed command value ω _M ^* given from the outside. The magnetic pole position estimation process is terminated.

  FIG. 14 is a graph in which the estimation error between the magnetic pole position estimation result and the true magnetic pole position according to the above embodiment is obtained by simulation. In the figure, the horizontal axis represents the initial magnetic pole position error, which means the error angle between the true magnetic pole position and the control magnetic pole position that occur when no magnetic pole position estimation is performed. FIG. 14 shows that according to the present embodiment, in principle, the magnetic pole position can be obtained with an accuracy within an estimation error of 2 degrees.

FIG. 15 is an overall schematic configuration diagram of a speed control device for a synchronous motor according to a second embodiment of the present invention. Each component is the same as that of the first embodiment shown in FIG. 1, but the q-axis current signal input to the magnetic pole position estimation unit 6 is changed from the q-axis current command value iq ^* to the q-axis current detection value iq. Only the changes are different. Although functionally equivalent to the first embodiment, the q-axis current detection value iq includes more noise components than the q-axis current command value iq ^* , so the internal configuration of the magnetic pole position estimation unit 6 is 2 and 3 is different from the first embodiment only in that the low-pass filter 37 shown in FIG. 2 and FIG. 3 is essential. If iq ^* in the first embodiment is replaced with iq, the flowcharts of FIGS. 12 to 16 can be similarly used in the second embodiment.

FIG. 16 is an overall schematic configuration diagram of a speed control apparatus for a synchronous motor according to a third embodiment of the present invention. The third embodiment is different from the first and second embodiments in that a position control system is provided. The change-over switch 20 is a switch SW3 that performs a selection operation of the speed command value ω _M ^* according to an instruction from the magnetic pole position estimation unit 6, and normally uses the output of the position controller 21 as the speed command value with the contact 1 selected. . When control of the synchronous motor 1 described later diverges, the function of setting the speed command value ω _M ^* = 0 is achieved by selecting the contact 2. The position controller 21 has a function of amplifying a deviation value between the position command value θ _M ^* obtained by the subtractor 22 and the position detection value θ _M. The change-over switch 23 is a switch SW2 that performs a position command value selection operation according to an instruction from the magnetic pole position estimation unit 6, and is normally in a state in which the contact 1 is selected. At this time, the position control system is in a state in which the position of the synchronous motor 1 can be controlled based on the position command value θ _M ^* given from the outside.

When performing the magnetic pole position estimation operation, an operation pattern corresponding to the speed command curve 51 of FIG. 4 generated by the command value generation unit 31 in the magnetic pole position estimation unit 6 by selecting the contact 2 of the changeover switch 23. Can be given as the position command value θ _M ^* . In this embodiment, the operation pattern is a waveform as shown by the speed command curve 51 in FIG. 4, and the integrated value is given as the position command value θ _M ^* , so that the flowcharts of FIGS. It is.

The above embodiment is summarized as follows. First, a synchronous motor 1, a power converter 2 that supplies a three-phase alternating current to the electric motor, and a current detection means 3 that detects a three-phase alternating current flowing through the electric motor are provided. Then, with the rotation angle detecting means 4 for outputting a signal theta _M corresponding to the angle at which the motor is rotated, the electrical angle calculation unit 5 for calculating an electrical angle theta _E of the electric motor with the rotation angle signal theta _M ing. Based on the calculated electrical angle θ _E , dq conversion means 7 for performing three-phase / two-phase coordinate conversion from the three-phase detection current to the two-phase current, and speed calculation means for calculating the rotation speed from the rotation angle signal θ _M 8 is provided. In addition, current control means 9 is provided for controlling the q-axis current command iq ^* and the d-axis current command id ^{* so} that the q-axis current detection value iq and the d-axis current detection value id, which are the outputs of the dq conversion means, are matched. ing. Furthermore, speed control means 10 and 11 for calculating a q-axis current command value iq ^*, which is an input value of the current control means, from a deviation between the speed command value ω _M ^* and the detected speed ω _M are provided. Further, the output to the two-phase command voltage of the current control means with an arithmetic electrical angle theta _E vq ^*, the speed control apparatus of a synchronous motor and a 3-phase converting means 12 for converting the vd ^* to 3-phase command voltage Vl ^* Is targeted. Here, means (6:30) for approximating relational data of a plurality of q-axis currents (command value iq ^* or detection value iq) respectively corresponding to the plurality of magnetic pole position error angles θ _err to a concave quadratic curve. I have. Next, an error angle calculating means for calculating a magnetic pole position error angle theta _err based on the X-coordinates indicating the minimum point of the quadratic curve (6:30), the magnetic pole position error angle theta _err in calculating the electrical angle theta _E Magnetic pole position estimation means (6: FIGS. 2 and 3) for estimating the estimated electrical angle θ ^ by addition / subtraction is provided.

  FIG. 17 is a configuration diagram of a hardware system in which the linear motor common to the above embodiments is the synchronous motor 1. In the figure, a position command value or a speed command value is issued from the host controller 291 to the servo amplifier 290. The linear type synchronous motor 292 drives the load 295. The movement of the movable element 296 of the synchronous motor is detected by the position detector 297 as a pulse train corresponding to the moving distance. Position information to be detected by the position detector 297 is recorded on the linear scale 298. The electric motor drive power output from the servo amplifier 290 is transmitted by the cable 299. The position detection signal of the electric motor 292 is transmitted to the servo amplifier 290 through the cable 301. A position command signal generated by the host controller 291 is transmitted to the servo amplifier 290 via the cable 302. Power is supplied to the servo amplifier 290 through the power cable 303.

  Next, the correspondence between the codes in FIG. 17 and the codes in the above-described embodiment will be described. The synchronous motor 1 corresponds to the synchronous motor 292, the rotation angle detector 4 of the synchronous motor 1 corresponds to the position detector 297, the load 14 corresponds to the load 295, and the drive shaft 15 has no corresponding portion. The power converter 2 and the following and the reference numerals 3 to 13 and 20 to 23 are included in the servo amplifier 290.

  In this configuration, the scale on the linear scale 298 fixed to the side surface of the linear type synchronous motor 292 is read by the position detector 297, so that the number of pulses proportional to the amount of movement of the mover 296 can be output. As described above, the first to third embodiments can be realized by the hardware configuration of FIG.

1 is an overall schematic configuration diagram of a speed control device for a synchronous motor according to a first embodiment of the present invention. The internal block diagram of the magnetic pole position estimation part using the data at the time of t1 of q-axis current command iq ^* . The internal block diagram of the magnetic pole position estimation part using the integral value of the absolute value of q-axis current command iq ^* . The speed command used at the time of magnetic pole position estimation, a position command pattern, and the waveform figure of q-axis current. The figure explaining the phase relationship of the true torque current I and temporary q-axis current command iq ^* . The graph which shows the relationship between magnetic pole position error angle (theta) _err and q-axis current command iq ^* . The graph which shows the relationship between a measurement data point and an estimation error angle. FIG. 3 is a diagram schematically illustrating a configuration of a memory 32. The flowchart which shows the address calculation method for ring memory access. The flowchart which shows the acquisition method of initial 4-point data. The flowchart which shows the check after 4 points | pieces data acquisition, and the reacquisition method. The figure explaining the mapping of the coordinate used for the calculation of the least squares method. FIG. 12 is a flowchart of least squares calculation following the flowchart of FIG. 11. FIG. The graph which calculated | required the estimation error of the magnetic pole position estimation result and the true magnetic pole position by simulation. The whole schematic block diagram of the speed control apparatus of the synchronous motor by the 2nd Example of this invention. The whole schematic structure figure of the speed control device of the synchronous motor by the 3rd example of the present invention. The block diagram of the hardware system which uses a linear motor as the synchronous motor 1. FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Synchronous motor, 2 ... Power converter, 3 ... Current detector, 4 ... Rotation angle detector, 5 ... Electric angle calculation part, 6 ... Magnetic pole position estimation part, 7 ... dq conversion part, 8 ... Speed calculator, DESCRIPTION OF SYMBOLS 9 ... Current controller, 11 ... Speed controller, 12 ... Three-phase converter, 13 ... Input signal changeover switch SW1, 14 ... Load, 15 ... Drive shaft or coupling member, 20 ... Input signal changeover switch SW2, 21 ... Position Controller 23... Input signal selector switch SW 1, 33... Least square method computing unit 35. Absolute value computing unit 36. Integrator with integral value clear function 50. Speed command pattern 51. Position command pattern 52. q-axis current (torque) waveform, ω _M ^* ... speed command value, ω _M ... speed detection value, iq ^* ... q-axis current command value, iq ... q-axis current detection value, id ^* ... d-axis current command value, id ... d-axis current detection value, vq ^* ... q-axis voltage command value, v ^* ... d-axis voltage command value, V1 * ^... 3-phase command voltages, I1 ... 3-phase alternating current flowing through the synchronous motor 1, θ ^ ... estimated electrical angle after correction, theta _{E ...} electric detection angle, theta _{M ...} rotor Position, θ _err ... Magnetic pole position error angle, θ _M ^* ... Position command value.

Claims (20)

A synchronous motor, a power converter that supplies three-phase alternating current to the electric motor, current detection means for detecting the three-phase alternating current flowing through the electric motor, and a signal (θ _M ) corresponding to the angle at which the electric motor rotates A rotation angle detecting means for calculating the electric angle (θ _E ) of the rotating shaft of the electric motor using the rotation angle signal (θ _M ), the calculated electric angle (θ _E ) and the electric angle Q-axis current calculation means for calculating a q-axis current based on the output of the current detection means; current command means for outputting a q-axis current command (iq ^* ); and the q-axis current calculated from the q-axis current command; Current control means for controlling the power converter so as to reduce the deviation of the electric power, error angle calculation means for calculating a magnetic pole position error angle (θ _err ) included in the calculated electric angle (θ _E ), and the electric angle (theta _E) and the magnetic pole position error angle The control apparatus for a synchronous motor provided with a magnetic pole position estimation means for calculating an estimated magnetic pole position (theta ^) based on theta _err) and,
Based on speed control or position control, using the same speed command pattern or the same position command pattern prepared in advance and performing three-phase / two-phase coordinate conversion and three-phase conversion used for control on the temporary dq axis Driving means for performing a plurality of driving operations;
Means for measuring the q-axis current while changing the magnetic pole position error angle (θ _err ) for each of these operations ;
The relationship between the magnetic pole position error angle (θ _err ) and the corresponding q-axis current measured for each operation is plotted on the XY coordinates with the horizontal axis (θ _err ) and the vertical axis the q-axis current. Means to
The relational data of a plurality of q-axis currents (command value or detection value) respectively corresponding to a plurality of magnetic pole position error angles (θ _err ) based on the plurality of plots are quadratic on the coordinates of error angle vs. q-axis current. A means of approximating a curve,
The control apparatus for a synchronous motor, wherein the error angle calculation means calculates the magnetic pole position error angle (θ _err ) based on an X coordinate indicating a minimum point of the quadratic curve.   2. The speed control apparatus for a synchronous motor according to claim 1, wherein said means for approximating the quadratic curve employs a least square method. 2. The speed control device for a synchronous motor according to claim 1, wherein the plurality of magnetic pole position error angles (θ _err ) are about 30 degrees apart. 2. The synchronous motor according to claim 1, wherein the number of data of the relationship data with respect to the magnetic pole position error angle (θ _err ) for quadratic curve approximation and the q-axis current command value or the q-axis current detection value is four. Speed control device. 2. The speed control apparatus for a synchronous motor according to claim 1, further comprising means for measuring an integral value of an absolute value of the q-axis current (command value iq ^* or detection value iq). According to claim 1, wherein the relational data of the magnetic pole position error angle versus the q-axis current (command value or detection value) is monotonously decreasing (or increasing) with respect to the change of the magnetic pole position error angle (θ _err ). And a means for setting a magnetic pole position error angle (θ _err ) larger (or smaller) than the already measured maximum magnetic pole position error angle (θ _err ) and re-measuring the relational data. Synchronous motor speed control device. A synchronous motor, a power converter that supplies three-phase alternating current to the electric motor, current detection means for detecting the three-phase alternating current flowing through the electric motor, and a signal (θ _M ) corresponding to the angle at which the electric motor rotates A rotation angle detection means for calculating the electrical angle (θ _E ) of the motor using the rotation angle signal (θ _M ), and a three-phase based on the calculated electrical angle (θ _E ). Dq conversion means for performing a three-phase / two-phase coordinate conversion from a detected current to a two-phase current, a speed calculation means for calculating a rotation speed from the rotation angle signal (θ _M ), a q-axis current command (iq ^* ) and Current control means for controlling the d-axis current command (id ^* ) so that the q-axis current detection value (iq) and the d-axis current detection value (id), which are outputs of the dq conversion means, are matched with each other, and the current control Q-axis current command value (i ^*) A speed command value (omega _{M ^*)} and the rotational speed (omega a speed control means for computing on the basis of the deviation between _M), calculation electrical angle (theta _E) 2 to output of said current control means with In a speed control device for a synchronous motor comprising three-phase conversion means for converting a phase command voltage (vq ^* , vd ^* ) into a three-phase command voltage (Vl ^* ),
Driving means for performing a plurality of driving operations based on speed control while performing three-phase / two-phase coordinate conversion and three-phase conversion used for control on a temporary dq axis using the same speed command pattern prepared in advance. When,
Means for measuring the q-axis current while changing the magnetic pole position error angle (θ _err ) for each of these operations ;
The relationship between the magnetic pole position error angle (θ _err ) and the corresponding q-axis current measured for each operation is plotted on the XY coordinates with the horizontal axis (θ _err ) and the vertical axis the q-axis current. Means to
A relational data of a plurality of q-axis current command values or q-axis current detection values (iq ^* or iq) respectively corresponding to a plurality of magnetic pole position error angles (θ _err ) based on the plurality of plots is a concave quadratic curve. Means to approximate
Error angle calculation means for calculating the magnetic pole position error angle (θ _err ) based on the X coordinate indicating the minimum point of the quadratic curve;
A speed control apparatus for a synchronous motor, comprising magnetic pole position estimation means for estimating an estimated electrical angle (θ ^) by adding / subtracting the magnetic pole position error angle (θ _err ) to / from the calculated electrical angle (θ _E ). .   8. The speed control apparatus for a synchronous motor according to claim 7, wherein the means for approximating the quadratic curve employs a least square method. 8. The position control means according to claim 7, further comprising a position control means for calculating a speed command value (ω _M ^* ) as an input value of the speed control means from a deviation between the position command value (θ _M ^* ) and the detected position (θ _M ). A speed control device for a synchronous motor, characterized in that 10. The magnetic pole position estimation means according to claim 9, wherein the magnetic pole position estimation means outputs an operation pattern prepared in advance as the position command value (θ _M ^* ) to the position control means, and the magnetic pole position error angle (θ _err ) for each operation. A speed control device for a synchronous motor, comprising means for measuring a q-axis current (command value iq ^* or detection value iq) for each operation while changing a value. 8. The synchronous motor speed control device according to claim 7, wherein the plurality of magnetic pole position error angles (θ _err ) are about 30 degrees apart. _{8. The} synchronous motor according to claim 7, wherein the number of data of the relationship data with respect to the magnetic pole position error angle (θ _err ) for quadratic curve approximation and the q-axis current command value or the q-axis current detection value is four. Speed control device. 8. The method according to claim 7, wherein the relationship data of the magnetic pole position error angle versus the q-axis current (command value or detection value) is monotonously decreased (or increased) with respect to a change in the magnetic pole position error angle (θ _err ). And a means for setting a magnetic pole position error angle (θ _err ) larger (or smaller) than the already measured maximum magnetic pole position error angle (θ _err ) and re-measuring the relational data. Synchronous motor speed control device. A step of supplying a three-phase alternating current from a power converter to a synchronous motor, a step of detecting a three-phase alternating current flowing through the motor, a step of detecting a rotation angle of the motor, and a motor using the rotation angle detection output a step of computing an electrical angle theta _E of the rotary shaft, a step of computing the q-axis current based on the calculated electrical angle theta _E and the detected current, and generating a q-axis current command, the q-axis current command Controlling the power converter to reduce the deviation from the calculated q-axis current and calculating the magnetic pole position error angle (θ _err ) included in the calculated electrical angle (θ _E ); In the magnetic pole position estimation method for a synchronous motor, including the step of calculating an estimated electrical angle (θ ^) by adding and subtracting the electrical angle (θ _E ) and the magnetic pole position error angle (θ _err ),
Based on speed control or position control, using the same speed command pattern or the same position command pattern prepared in advance and performing three-phase / two-phase coordinate conversion and three-phase conversion used for control on the temporary dq axis An operation step for performing a plurality of driving operations;
A measurement step for measuring the q-axis current while changing the magnetic pole position error angle (θ _err ) for each of these operations ;
The relationship between the magnetic pole position error angle (θ _err ) and the corresponding q-axis current measured for each operation is plotted on the XY coordinates with the horizontal axis (θ _err ) and the vertical axis the q-axis current. Plot step to
The relational data of a plurality of q-axis current command values (or q-axis current detection values) corresponding to each of a plurality of magnetic pole position error angles (θ _err ) based on the plot of the plurality of points is expressed as error angle vs. q-axis current coordinates. Approximating a concave quadratic curve above,
A method for estimating a magnetic pole position of a synchronous motor, comprising: calculating the magnetic pole position error angle (θ _err ) based on an X coordinate indicating a minimum point of the quadratic curve.   15. The method for estimating a magnetic pole position of a synchronous motor according to claim 14, wherein a least square method is applied to the approximation of the quadratic curve. 15. The point according to claim 14, wherein each point according to the relational data is represented by (data_x [k], data_y [k]) (k = 1, 2,..., N (n: natural number)), and the change width of the magnetic pole position error angle θ _err. The magnetic pole of the synchronous motor is characterized in that, when Δθ, the least-squares method calculation is performed on the coordinates in which each point is mapped to ((data_x [k] −data_x [1]) / Δθ, data_y [k]). Position estimation method. 15. The step of outputting the operation pattern prepared in advance as a position command value (θ _M ^* ) according to claim 14, and the electric angle (θ _M ) of the motor rotating shaft coincide with the position command value (θ _M ^* ). To output a speed command value (ω _M ^* ) and to control the rotation speed (ω _M ) of the motor so as to coincide with the speed command value (ω _M ^* ). ^* ) And a step of measuring a q-axis current (command value iq ^* or detection value iq) while changing the value of the magnetic pole position error angle (θ _err ). Magnetic pole position estimation method. 15. The synchronous motor magnetic pole position estimation method according to claim 14, wherein the plurality of magnetic pole position error angles (θ _err ) are set at intervals of about 30 degrees. _15. The magnetic pole of a synchronous motor according to claim 14, wherein the number of data of relational data with respect to the magnetic pole position error angle θ _err for approximating the quadratic curve to the q-axis current command value or the q-axis current detection value is four. Position estimation method. 15. The method according to claim 14, wherein the relationship data of the magnetic pole position error angle versus the q-axis current command value or the q-axis current detection value is monotonously decreased (or increased) with respect to a change in the magnetic pole position error angle (θ _err ). Te, comprising the steps of: setting a greater than the maximum of the magnetic pole position error angle previously determined (theta _err) (or smaller) magnetic pole position error angle (theta _err), the step of performing remeasurement of the related data based on this A magnetic pole position estimation method for a synchronous motor, characterized in that:

Images