JPH0442909B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a control method for a pulse width modulation type inverter that reduces torque pulsation in a low speed region, and particularly relates to a method for controlling a pulse width modulation type inverter to reduce torque pulsation in a low speed region, and in particular, to reduce torque pulsation in a low speed region. This is a pulse width modulation type inverter that is turned ON and OFF alternately between 30° and 30°. The element that transitions to the conductive region has a pulse width peak value on both sides of the output frequency electrical angle of 30° as the base point, and the element that transitions to the non-conductive region , the pulse width peak values on both sides of the output frequency electrical angle of 150° are turned on differently.
The present invention attempts to provide a control method that averages the torque during the commutation overlap period by performing OFF control, increases the torque average value, and enables smooth low-speed operation.

When variable speed control of induction motors etc. is performed using a current type inverter, the torque pulsation is large, especially in the low speed altitude range, and if the pulsation frequency matches the torsional resonance frequency of the shaft system, it will not only have a large impact on the mechanical system, but also cause The instability phenomenon is well known, such as "unevenness" becoming so large that stable operation cannot be expected at all. To describe this phenomenon concretely, for example, when an induction motor is driven at variable speed by a 120° square wave inverter, the characteristics of the generated instantaneous electrical torque with respect to the progression of the output frequency electrical angle ωt in the extremely low frequency region are shown in Figure 1. The torque characteristics are as shown in . This characteristic shows the change in instantaneous torque when the inverter output frequency is taken as a parameter.As is clear from this characteristic diagram, instantaneous torque is expressed in electrical angle.
It pulsates every 60 degrees, and the torque pulsation will be reduced especially if the output frequency becomes high, but the relationship between the torque pulsation reduction effect and the actual generated torque is dichotomous, such as when the torque is significantly reduced. This is contradictory, and a 120° square wave inverter is not practical as its application fields are limited. As a method to solve this problem, as shown in FIG. 2, the electrical angle π/6 is
(rad) as the base point, 0 to π/6 and π/6 to
Period with π/3 and 2/3π to 5/6π and 5/6π
Chopping control is performed to generate multiple pulses each with a constant pulse width during the period of ~π,
π/3 to 2/ based on electrical angle π/2 (rad)
A pulse width modulation type inverter using an equal pulse width method that generates a fixed pulse width during a period of 3π is well known. When the induction motor is driven at variable speed by this pulse width modulation type inverter, the characteristics of the generated instantaneous electrical torque with respect to the progression of the output frequency electrical angle ωt are as shown in FIG. 3. This torque characteristic diagram shows that the inverter output frequency is 0.5
(Hz).The solid line shows the torque characteristic with a pulse width modulation type inverter with equal pulse width, and the broken line shows the torque characteristic with a 120° square wave inverter.From this characteristic diagram, As is clear, when the current is subjected to a predetermined pulse width modulation, the torque pulsation frequency is so high that it has no effect on the shaft system. On the other hand, the torque pulsation in the ultra-low speed range is not much improved compared to the 120° square wave inverter. A solution to the problem of torque pulsation is an unequal pulse width modulation type inverter that performs pulse width modulation as shown in the current waveform pattern of FIG. This inverter has a range of 0 to π/with respect to the progression of the output frequency electrical angle ωt.
In the period up to 3, time ratio control is performed to gradually increase the pulse width, and in the period from 2/3π to π, time ratio control is performed to gradually decrease the pulse width. Then, the characteristics of the instantaneous electric torque generated when the inverter output frequency is 0.5 Hz are as shown in FIG. As is clear from comparing this torque characteristic diagram with the torque characteristic diagram in Figure 3, the maximum value of the instantaneous torque is
In contrast to the one in the figure, which is approximately constant, in the one in Figure 5, the maximum value of the instantaneous torque changes in an arc shape as the output frequency electrical angle progresses, and the width of the drop in torque gradually decreases over time. Therefore, overall, control with unequal pulse widths has a greater effect of reducing torque pulsation than control with equal pulse widths. However, the problem with the unequal pulse width modulation method is that, as is clear from the generated electrical instantaneous torque characteristics in Figure 5, the problem with the progression of the output frequency electrical angle is particularly in the period from 0 to π/6. When comparing the average torque in the period from π/6 to π/3, it can be seen that the latter average torque is larger than the former average torque. What this means is that if it were ideal, the average torque in each period would be approximately the same, so it cannot be said that the effect of reducing torque pulsation is improved. For such an unequal pulse width modulation method, for example, an element transitioning to a conducting region and an element transitioning to a non-conducting region may be alternately switched such that when one is ON, the other is OFF. Phase-to-phase commutation is performed through ON-OFF control, and elements that transition to a conductive region gradually increase their pulse width, whereas elements that transition to a non-conductive region gradually decrease their pulse width. , there is an unequal pulse width modulation method that also uses time ratio control. In this modulation method, the control in the commutation overlap period and the control in other periods are performed completely independently, so it is possible to stop at any point even in the ultra-low speed region, which is compared to the pulse width modulation method shown in Figure 4. This enables stable low-speed operation. However, the problem is, for example, when controlling during the commutation overlap period,
As mentioned above, since the time ratio control is performed to increase or decrease the pulse width between the element transitioning to the conducting region and the element transitioning to the non-conducting region, the electrical angle of 30° and There is "velocity unevenness" in the ultra-low speed region, as seen in the unequal pulse width modulation method shown in Figure 4, where the average torque in the sections before and after 150 degrees is different, which is extremely Smooth low-speed driving cannot be expected at all.

The present invention was developed by paying attention to this point, and in particular, the present application is directed to alternating between specific phases.
This is a pulse width modulation method that performs ON-OFF control and phase-to-phase commutation, and the element transitioning to the conduction region has an electrical angle
By making the peak values of the equal pulse width trains in the sections before and after 30 degrees different, and making the peak value in the previous section large, on the other hand, the element transitioning to the non-conducting region has the exact opposite. The main feature is that the pulse width modulation method is performed, and will be described in detail below based on the embodiment shown in FIG.

In this embodiment, 1 is a forward converter formed by connecting thyristors in a pure bridge and converts AC input power into DC power, 2 is a DC reactor, and 3 is generally called a series diode type inverter. An inverter converts direct current power into alternating current power and supplies power to a load motor 4 (an induction motor is shown in the embodiment); 5 is a small generator for speed detection to extract an actual speed signal; 6 is a setting device for giving a speed command signal, 7 is a first comparator for comparing the speed setting command signal and the actual speed detection signal, and 8 is a speed control device for once amplifying the speed error voltage. 9 is an absolute value conversion circuit for extracting the absolute value of the input signal; 10 is a DC current detector such as a shunt resistor or DC transformer for extracting the DC voltage of the DC intermediate circuit; and 11 is a DC current detector. 12 is a current control amplifier for once amplifying the current error voltage; 13 is an automatic comparator for arbitrarily shifting the phase of the gate signal; In the pulse phase shift circuit, 14 is an amplifier for once amplifying the slip frequency command signal of the speed error voltage.
Reference numeral 15 is an oscillator for oscillating the required pulses based on the input slip frequency command signal, and 16 is an oscillator for counting the pulse signals derived from the oscillator. ,
17 is a pulse width modulation counter that produces a pulse signal train in a specific period, and 17 is a pulse signal train corresponding to the progression of the output frequency electrical angle ωt derived from the pulse width modulation counter, and an F signal indicating normal rotation. A frequency divider that obtains a gate signal that sequentially fires the thyristor group of the inverse converter 3 using three quantities including an R signal indicating reversal, and a clock input terminal of this frequency divider.
A pulse signal train will be input to CL. 1
8 is a counter for counting the pulse signal input from the ring counter and applying amplitude modulation to the DC current during the commutation overlap period, and 19 ₁ -19 ₂ is
It is a NOT gate, 20 ₁ -20 ₂ is an AND gate,
Reference numeral 21 denotes an amplifier for once amplifying a signal obtained by converting a rectangular wave pulse signal into a DC level signal, and guiding this amplified signal to the second comparator 11 as a correction signal.

To describe the operation of this embodiment configured as above, first, when a required speed command signal is given from the speed setter 6 to the first comparator 7, this speed command signal is directly used for speed control of the major loop. The forward converter 1 is connected to the forward converter 1 through the path of the amplifier 8 → absolute value converter 9 → second comparator 11 of the current minor loop → current control amplifier 12 → automatic pulse phase shift circuit 13.
The required DC power is applied to the thyristor group 1 from the forward converter 1 to the inverse converter 3. In parallel with this operation, the speed command signal derived from the speed control amplifier 8 is once amplified by the amplifier 14, and this amplified DC level signal is applied to the oscillator 15. Sequentially oscillates a group of pulse signals with frequencies corresponding to
These signal groups are applied to 16 pulse width modulation counters for counting. The operation mode of this oscillation pulse signal group and the pulse width modulation counter is shown in the time chart of FIG. 7. For example, when the pulse width modulation counter counts 5 or more pulses, the count value is When the count value reaches 8, a rectangular wave signal as shown in FIG. 7 rises from the first terminal (not shown) of the pulse width modulation counter. A rectangular wave signal as shown in Fig. 7 is sent from the third terminal when the count value reaches 13, and a rectangular wave signal as shown in Fig. At this point, a rectangular wave signal as shown in Figure 7 is sent from the 4th terminal, and similarly, when the count value reaches 40, a rectangular wave signal as shown in Figure 7 is sequentially sent from the 10th terminal. Then, from these rectangular wave signals indicated by ~, a logic gate (not shown) of the pulse width modulation counter 16 calculates the logical product condition of the signal ~ and the negation signal ~ of the signal ~, as shown in Fig. 7. Obtain a signal with a pulse width as shown in ○A. Similarly, by calculating the logical product condition of the signal and each negation signal ~, we can obtain the signal shown in ○B.
By taking the logical product condition of the signal and each negative signal ~, we can obtain a signal as shown in ○C, and by calculating the logical product condition of the signal ~ and each negative signal ~, we can obtain a signal such as shown in ○d, ~ By calculating the logical product condition of each signal and the negation signal, the signals shown in circles are extracted.

The pulse widths of the signal groups ○I to ○H extracted in this way are taken into consideration in advance. The signals also have equal pulse widths, and the pulse width of the former signal group is narrower than the pulse width of the latter signal group, and these five pulse train signal groups are expressed as an abbreviation of electrical angle with respect to the progression of the output frequency electrical angle ωt. It is generated in a period of π/3 (rad). By performing processing exactly the same as the above operation, the second counter (not shown) of the pulse width modulation counter 16 and the logic gate calculate the electrical angle of about π/3 to about 2 shown in the current pattern of FIG. Similarly, a third counter (not shown) of the pulse width modulation counter 16 and a logic gate convert the pulse signal in the period of /3π into electrical angles of 2/3π to π as shown in the current pattern of FIG. A pulse train signal group is obtained over a period of time. In this way, pulse width modulation counter 1
The reason why the configuration of 6 is composed of three sets of counters and logic gates is that, for example, in this application, the control in the commutation overlap period and the control in other periods are performed independently. There is no need to provide three sets of counters in this way; the desired purpose can be achieved with just one counter.
Furthermore, the current pattern shown in FIG. 7 can be easily obtained by software processing using a microcomputer.

Now, when the current pattern mode shown in FIG. 7 is obtained by the pulse width modulation counter 16, for example, the rise of each of the rectangular wave signals from to in FIG. Take out the F signal and the R signal indicating reverse rotation, respectively.
These F signal group and R signal group are respectively passed through OR gates (not shown) to 17 ring counters, a second NOT gate ₁₉₂ , and a second AND gate 2.
0 and ₂ respectively. Note that since consideration has been made in advance that the F signal is at level "0", for example, the pulse width modulation counter 16
When the signal shown in FIG. 7 and the F signal "1" indicating normal rotation are input to the ring counter 17,
First, on the condition that the 17th ring counter is an F signal, the thyristor of the positive side U arm and the thyristor of the negative side Y arm of the inverse converter 3 are connected to set the phase order of the inverse converter 3 to the forward rotation mode. At the same time, the rising portion of the ◯ mark signal in FIG. 7 is applied to the counter 18 to count it. Therefore, when the U-arm thyristor and the Y-arm thyristor are fired, a load current begins to flow through these thyristors in the path of the U-phase winding and V-phase winding (not shown) of the motor 4. In such a state, the signal shown in FIG.
When the R signal is applied to the NOT gate 19 ₂ and the second AND gate 20 ₂ , the ring counter 17 sets the phase order of the inverse conversion unit to the inversion mode based on the R signal. A gate signal is applied to the thyristor and the thyristor of the negative side Y arm, and at the same time, the load current is caused to flow through the path of the W-phase winding and the V-phase winding (not shown) of the motor 4, and at the same time, the counter 18 Perform the first count at the signal. Note that the counter 18 is 3 from ○I to ○C in Figure 7.
Since consideration has been made in advance so that the output will be "L" at the time the signals are counted, the output of the counter 18 will be at the "L" level at the time the signals ○A and ○RO○C are counted. The output of the NOT gate ₁₉₁ becomes "H" level, and this "H" level signal is given to the first and second AND gates _201-202 , and when the F _signal is input, the When the logical condition of the second AND gate 20 ₂ is satisfied, the gate is opened, and when the R signal is input, the logical product condition of the first AND gate 20 ₁ is satisfied and the gate is opened. In this way, during the period when the 18 counters are counting up to the third pulse train signal group, each
The logical product output of the AND gate is amplified by the amplifier 21, and this amplified signal is given as a correction signal to the comparator 11 of the minor loop. Using the command signals added with the same polarity, the firing phase of each thyristor element of the forward converter 1 is controlled to change the DC output current to ○a, ○b in Fig. 7.
Since control is performed to increase the amplitude of the DC current in synchronization with the pulse period of ○C, in the normal rotation mode where the F signal is input, the load current supplied to the motor 4 is as shown in ○I in Fig. 7. The level is temporarily increased as shown in the pulse signals shown in and ○Ro○C, and as a result, the electrical angle becomes approximately π/with respect to the transition of the output frequency electrical angle ωt.
The instantaneous torque in the period up to 6 becomes large. In this state, when the pulse width modulation counter 16 inputs a signal as shown in ○D in FIG. At the same time as the ignition, the pulse signal of ○2 is sent to the counter 18.
to set the output to "H" level. In this way, when the output of the counter 18 goes to the "H" level, the output of the first NOT gate ₁₉₁ goes to the "L" level, and the AND conditions of the AND gates _201-202 are no longer satisfied _. Based on this condition, the output of the amplifier 21 drops to zero, and the current command signal given from the absolute value conversion circuit 9 and the current detection signal given from the DC current detector 10 are used to control each thyristor element of the forward conversion section 1. The load current supplied to the motor 4 in the normal rotation mode in which the ignition phase is controlled to perform predetermined current control and the F signal is input is approximately π/6 to approximately π in electrical angle as shown in FIG. It goes down temporarily as shown in the period of /3.

As described above, during the period when the electrical angle reaches approximately π/3 with respect to the progression of the output frequency electrical angle ωt (this period indicates the commutation overlap period), the F signal causes a transition to the conduction region. The thyristor of the U-arm to be activated is fired, and the thyristor of the W-arm to be transitioned to a non-conductive area is activated by the R signal, and the elements to be transitioned to a conductive area and the elements to be transitioned to a non-conductive area are alternately switched on. In the process of ON-OFF control, the present invention particularly focuses on the level of the current pulse width supplied to the motor during the period up to approximately π/6 electrical angle for the element transitioning to the conductive region, and the The level of the current pulse width supplied to the motor in the period from π/6 to approximately π/3 is made different, and a predetermined pulse width control is performed so that the former level is greater than the latter level, On the other hand, for elements transitioning to a non-conducting region, a predetermined pulse width control with a completely opposite relationship is performed, so the conventional unequal pulse width modulation method or unequal pulse width using time ratio control is used. Compared to the method that uses both the modulation method and the commutation control between phases, as shown by the broken line in the generated electrical instantaneous torque characteristic diagram in Fig. 5, especially in the period when the electrical angle is up to π/6, As the peak value of the instantaneous torque increases, the average torque increases, and the average torque for the same period and the average torque for the period from electrical angle π/6 to π/3 are approximately at the same level. It can be understood. A feature of the present application is that the average torques of the first period and the second period in the commutation overlapping period are made to be approximately at the same level. Note that when the electrical angle reaches approximately π/3, the commutation from the W phase to the U phase side is completely completed, so after the electrical angle reaches approximately π/3, the constant pulse width signal is This operation is continued until the electrical angle reaches approximately 2/3π, and the load current is alternately applied to the Y arm and Z arm on the negative side. The current pattern mode during this period becomes the pattern mode shown in FIG.
The reason why the level of the constant pulse width differs in time with respect to the transition of the output frequency electrical angle ωt in the pattern mode shown in the figure is that, for example, the Y arm on the negative side shifts to the non-conducting region, whereas the Z arm on the negative side In the process of the arm transitioning to the conduction region, as mentioned above, the Y arm and Z arm on the negative side are alternately ON-OFF controlled, and the width and level of the current pulse supplied to each arm are controlled in each period. be different from each other,
This is to perform predetermined pulse width control so that the average value of instantaneous torque in each period is approximately at the same level. In this way, by applying the current pattern mode signal shown in FIG. 7 to the ring counter, the F signal or R
When the elements of the inverse conversion section are controlled ON-OFF in a predefined time width based on the signal, (A) to (C) in Fig. 8 are obtained.
A pulse width modulated current as shown in is supplied to the motor. In the current waveform diagram of Fig. 8, (a) indicates the U-phase current, and (b) indicates the V-phase phase current.
At the same time, (c) shows the phase current of the W phase.

As described above, in the present invention, the pulse width modulation method performs phase-to-phase commutation by alternately controlling ON-OFF of elements transitioning to a conductive region and elements transitioning to a non-conducting region within a commutation overlap period. In particular, for elements transitioning to the conduction region, the level of the current pulse width during the period of electrical angle 0 to π/6 is set to electrical angle π/6 to π/3.
The current pulse width is controlled to be larger than the level during the period of , and a signal with a completely opposite current pulse width is given to the element transitioning to the non-conducting region. It has various effects.

The electrical angle is 0 with respect to the progression of the output frequency electrical angle.
The period up to π/6 and the electrical angle are π/6 to approximately π/
Since the average torque is set to approximately the same level as in the period up to 3, torque pulsation in the ultra-low speed region and the low speed region can be effectively suppressed.

Since amplitude modulation is applied to the DC current of the inverter during the commutation overlap period, speed unevenness in the low speed region is eliminated and smooth low speed operation can be performed.

Since amplitude modulation is applied to the DC current of the inverter in the low speed region, it is possible to increase the average torque in each period of the generated instantaneous electrical torque, which not only improves the starting characteristics but also greatly improves controllability. Can provide good PWM inverter.

[Brief explanation of drawings]

Figure 1 is a torque characteristic diagram showing the generated electrical instantaneous torque with respect to the progression of the output frequency electrical angle ωt by a conventional 120° square wave inverter, Figure 2 is a current pattern waveform diagram by the conventional equal pulse width modulation method, and Figure 3 The figure is a torque characteristic diagram showing the relationship between the electric instantaneous torque characteristics generated by a conventional 120° square wave inverter and the equal pulse width modulation method. Figure 4 is a current pattern waveform diagram by the conventional unequal pulse width modulation method. Figure 5 is a torque characteristic diagram showing the relationship between the generated electrical instantaneous torque characteristics between the conventional unequal pulse width modulation method and the pulse width modulation method according to the present invention, and Figure 6 is a pulse width modulation diagram showing an embodiment of the present invention. Specific circuit diagram of type inverter,
FIG. 7 is a time chart showing the operation of the pulse width modulation counter, which is the main part of the present application, with respect to the progression of the output frequency electrical angle ωt, and FIG. 8 is a current pattern waveform showing the phase current of the pulse width modulation type inverter according to the present invention. figure. 1 is a forward conversion section, 2 is a DC reactor, 3 is an inverse conversion section, 4 is a load motor, 6 is a speed setting device, 7-1
1 is a comparator, 8 is a speed control amplifier, 9 is an absolute value conversion circuit, 10 is a DC current detector, 12 is a current control amplifier, 13 is an automatic pulse phase shift circuit, 14-
21 is an amplifier, 15 is an oscillator, 16 is a pulse width modulation counter, 17 is a ring counter, 18 is a counter, 19 ₁ -19 ₂ is a NOT gate, 20 ₁ -2
0 ₂ is an AND gate.

Claims (1)

[Claims] 1 A current source inverter consisting of a thyristor pure brittle connection forward conversion section, a DC reactor, a series diode type inverse conversion section, and a control section, which applies pulse width modulation to the AC output current supplied to the load from the inverse conversion section. In this case, phase-to-phase commutation is performed by alternately controlling ON and OFF the elements that transition to the conduction region and the elements that transition to the non-conduction region during the commutation overlap period of the inverse conversion section, and control the elements during the commutation overlap period. In a device in which the control in other periods is performed independently, the control in the low speed region where the output frequency of the inverse conversion section is below a predetermined value is shifted to the conduction region side within the commutation overlap period. The transition element changes the level of the signal train with equal pulse width (pulse width A) during the period when the electrical angle is from 0 to π/6 (T1), and the level of the signal train with equal pulse width (pulse width A) during the period when the electrical angle is from π/6 to π/3 (T2). The level of the equal pulse width (pulse width B) signal train is set to be greater than the level of the equal pulse width (pulse width B) signal train in the period T2, and the element that moves to the non-conducting region side makes the level of the equal pulse width (pulse width B) signal train in the period T1 higher than the level of the signal train in the period T2. The element that moves to the conduction region side of the current minor loop that controls the forward conversion section is set so that the level of the equal pulse width (pulse width A) signal train is smaller than the level of the signal train.
Each ON period is synchronized with the ON period of the T1 period and the ON period of the T2 period of the element transitioning to the non-conducting region side.
A method for controlling a pulse width modulation type inverter, characterized in that a correction signal is added only during a period to control the firing phase of each element of a forward conversion section and apply current amplitude modulation to a direct current.