CH662019A5 - INVERTER ARRANGEMENT WITH A TWO-PULSE INVERTER WITH A TOLERANCE BAND CONTROL DEVICE. - Google Patents

Description

The invention relates to an inverter arrangement with a two-pulse inverter with a device for tolerance band control with two partial inverters, between which a load is connected, and by comparing the setpoint current and the actual value current to achieve a predetermined current profile.

A tolerance band control of the type mentioned above is already known and is e.g. described in the book "Electric traction vehicles", Vol. 2 (2nd edition) by Karl Sachs, Springer-Verlag Vienna - New York 1973, p. 883 as a two-point control. However, if both partial inverters are reversed simultaneously in a two-pulse inverter, there are some disadvantages. The switching frequency can become impermissibly high or low, or only one of the two inverters can participate in the switching processes over a long period of time.

It is the object of the invention to eliminate the disadvantages described in a particularly simple manner and to provide an inverter with a device for tolerance band control of the type mentioned at the beginning, in which the switching levels for both partial inverters can be easily shifted, in which e.g. the tolerance bands can be swapped in a meaningful way and expanded with a high switching frequency.

The above-mentioned object is achieved according to the invention in that each of the two partial inverters is connected to its own comparator and in that the switching levels of the individual comparators with the associated switching arrangements for the hysteresis setting and formation for both partial inverters are mutually shifted.

The advantage of the invention can be seen in particular in the fact that the switching frequency of the individual parts either becomes lower or the ripple of the output current decreases.

The advantages of the further developments contained in the dependent claims are mentioned later in the description of the mode of operation.

A tolerance band control, which is also called a two-point controller, allows an inverter with a DC link to get by with relatively little control electronics. The instantaneous setpoint is compared with the instantaneous current actual value on a comparator. The output of this comparator switches the load to a positive or negative voltage, depending on whether the current actual value is too small or too large. The size of the hysteresis of the comparator and the time behavior of the load result in a free switching frequency. The hysteresis width on the comparator represents the tolerance band in which the actual current value fluctuates if the time behavior of the load is given by a simple time constant. In the case of a two-pulse circuit with two partial inverters per load circuit (one phase of a multi-phase load), expediently in a single-phase bridge circuit (common voltage intermediate circuit), both parts of the inverters can be switched simultaneously (of course in push-pull mode, i.e. if one switches to plus, the other switches to minus and vice versa) ). In this case, the device works with only one inverter per load circuit, as mentioned above. However, a two-pulse circuit includes other options that are not used by the two-point control described in this paragraph. In addition to positive and negative voltage at the load, this circuit also enables zero voltage at the load, in two ways, either both partial inverters to plus or minus.

According to the present invention, the two-pulse circuit can be used in a simple manner by comparing the same instantaneous current setpoint and actual values with two separate comparators, each of which controls a partial inverter. The tolerance bands of both comparators are slightly shifted against each other, i.e. one receives a small additional positive voltage, the other a negative one, in the order of the tolerance range. If now e.g. with positive voltage at the load (one partial inverter at plus, the other at minus) the current becomes more positive, it first reaches the switching threshold of one comparator, e.g. for one partial inverter and switches it to minus. The voltage at the load becomes zero (free-wheeling, both poles at minus). If the current then decreases as a result of this, this comparator will now take over the current regulation on its own; the second comparator switches continuously to minus. However, if the current still increases, the switching threshold of the second comparator is reached and this switches the other part of the inverter to plus, so that at the

5

10th

15

20th

25th

30th

35

40

45

50

55

60

65

3rd

662 019

Reverse polarity load, i.e. negative voltage appears. If the current now decreases again, the other partial inverter switches to minus (freewheeling, both poles to minus), which can result in the current rising again. In this case, the comparator takes over the current control for the other partial inverter and one partial inverter remains permanently switched to minus. The freewheel (voltage = 0) was on the minus rail side in both cases. Depending on the control and load conditions, it can also happen that the freewheel is on the side of the plus rail. Since, in addition to positive and negative voltage, zero can also be applied to the load, we are dealing with a three-point controller. With the same load conditions, the current ripple becomes smaller with the same switching frequency or the switching frequency becomes lower with the same current ripple (tolerance range), which is an advantage.

Especially at low fundamental frequencies, including direct current operation, the circuit described may mean that only one inverter cycles and the other remains in one switch position for a long time. On the one hand, this leads to an uneven distribution of the switching losses between the two partial inverters.On the other hand, depending on the design of the inverters, the commutation capability can be lost if an inverter stops clocking for a longer period (e.g. more than 1 s) (discharge of the quenching capacitors). In order to remedy this deficiency, according to a further development of the invention, e.g. every second circuit exchanged the shift of the two tolerance bands. This role reversal of the two partial inverters ensures that both switch alternately. As a result, everyone only switches at half the switching frequency, even if the voltage at the load changes only between 0 and plus or 0 and minus for a long time. With DC operation (basic frequency = 0)

it can happen that voltage zero is constantly demanded. In this case, switching no longer takes place despite the tolerance band exchange, which can lead to a loss of commutation capability. This can be remedied by a device that initiates a forced changeover whenever a partial inverter has not received a changeover command for a certain period of time. As a result, the control will respond and perform another correction circuit. Since the condition for this time depends on the inverter properties, this dual switching device should expediently be implemented in the control set upstream of the inverter (pulse generation), where compliance with the other limit (switching back too quickly) must also be ensured.

If the hysteresis of the comparators (tolerance range) is fixed, an instantaneous switching frequency is set which runs parabolically from zero with full negative voltage with a minimum with half negative voltage. It reaches zero again at zero voltage and then runs after a further parabola with maximum at half the positive voltage and again reaches zero at maximum positive voltage. If the voltage is controlled sinusoidally over time (which is necessary for sinusoidal current with a linear load), a modulated clock frequency is set, the mean value of which depends on the modulation. This mean value also has a maximum approximately in the middle of the control range, here the maximum modulation is to be understood as that where there is no overdriving (cutting off the peaks) in the sinus domes. Trials have shown this. In order not to exceed the switching frequency permitted for the inverters in the entire operating range, the hysteresis must be made so large that this half-modulation does not exceed this permissible frequency. Then it becomes relatively small at the control limits. Especially if you want to use the overdrive, this leads to disadvantages.

By overdriving (i.e. cutting off the sinus domes in the medium voltage), the undershoot clocking of the normal tolerance band control can be gradually converted into a basic wave clocking (voltage proportional to the basic frequency, as required for induction machines). With increasing overload, the clocking shrinks more and more around the zero crossings of the sinusoidal modulation until only one circuit remains per zero crossing. From here, the output voltage is of course constant (no longer frequency-proportional) and rectangular. In order to have the transition area as constant as possible (no large jumps), the switching frequency should be as high as possible. In order to utilize the average switching frequency permitted by the inverter as far as possible in the entire control range, the instantaneous current tolerance band control can be subordinated to a clock frequency control. An adjustable setpoint is compared with the measured average clock frequency via a controller. The hysteresis of the comparators serves as the manipulated variable. The actual value of the clock frequency is obtained via frequency-voltage converters. In three-phase circuits, the filter time constant required for averaging can be reduced by using the mean value or maximum value of the three phases. In this case, only a clock frequency control is necessary for all three phases with a common hysteresis specification as the manipulated variable.

The invention is explained in more detail below with reference to the drawing.

In the drawing shows

1 shows an embodiment of a device for tolerance band control for a two-pulse inverter,

2 is a block diagram of a device according to FIG. 1, expanded for three-phase operation,

Fig. 3 is a voltage-time diagram to illustrate the operation of the invention.

The circuit arrangement shown in FIG. 1 is roughly divided into the following modules, each with dashed lines:

An inverter power section WL, comprising a two-pulse inverter 1, a tolerance band control device TB, a device for tolerance band reversal TU, the device per se for setting the switching level NI being shown as a separate assembly due to the expansion to be described later on multi-phase operation and a proportional / integral controller PI.

The two-pulse inverter 1 contains two partial inverters 2, 3, between which a load 4 is connected. The actual current ij flowing through the load is detected by a current transformer 4 '. A capacitor 5 serves as a smoothing capacitor. Two comparators 6, 7 are provided with resistors 8 to 15 for setting the operating point. A potentiometer 16 is connected to an analog inverter 16 'as a DC voltage source for shifting the switching level. A changeover switch 17 is connected to a bistable multivibrator, flip-flop 18, which is connected to two inverters 19, 20 and an exclusive-OR circuit 22 via a clock input 21. Via the switch 17, alternating positive or negative voltages, which can be set by means of the potentiometer 16, are fed to both comparators 6 and 7 in time with the output signal of the flip-flop 18.

A controller 23 is used to set the hysteresis and contains resistors 24 to 28, a capacitor 29 and a diode 30, which has the limiting function for minimal hysteresis. The output of the controller 23 leads directly to first changeover contacts of two changeover switches 34, 35 in the same direction. The second changeover contacts are connected to the output of the controller 23 via an analog inverter 31 '. A potentiometer 31 is provided for the second variant of setting the hysteresis (indicated by the dashed line). A setting wi

s

10th

15

20th

25th

30th

35

40

45

50

55

60

65

662 019

4th

the status 32 is intended for setting the minimum hysteresis. Another potentiometer 33 serves as a DC voltage source. Two changeover switches 34, 35 are connected to the outputs of the comparators 6 and 7. The reference number 36 denotes an f / U converter, is the setpoint current and ij the actual value current.

The tolerance band control of the switching arrangement described and drawn is carried out in the following way:

The instantaneous current setpoint and actual value are compared with one another in the two comparators 6 and 7. The outputs switch the inverter 1 using control sets (not shown). A "1" signal at the comparator outputs means that both partial inverters 2 and 3 switch to plus, while the "0" signal switches to minus. In order for the control sense to be correct (negative feedback), the setpoint and actual value must be connected to each other at the inputs of the comparators 6 and 7. The outputs of the comparators 6 and 7 switch the hysteresis, which can be set together at the potentiometer 31, back to the inputs via the changeover switches 34 and 35 (positive feedback). The adjustment of the tolerance bands (Bj, B2 in Fig. 3) that can be adjusted on the potentiometer 16 is carried out via the switch

17 either positive or negative on the minus inputs of comparators 6 and 7. Because the setpoint and actual value are connected in a crossed manner to both comparators 6 and 7, the same polarity at the minus inputs causes the tolerance bands Bi, B2 to shift in opposite directions. Reversing this polarity causes the function of both partial inverters 2 and 3. The flip-flop 18 controls this polarity change. If both comparator outputs have the same signal, then the preparation inputs of the flip-flop

18 set. The next changeover is then detected via the exclusive-OR circuit 22 and the switching through to the output is effected via the clock input 21 of the flip-flop 18. When operating with clock frequency control, the clock frequency setpoint is compared by potentiometer 33 with the actual value by f / U converter 36. The difference influences the hysteresis input of the rest of the circuit via the controller 23, which is designed as a proportional-integral controller, the hysteresis being limited at the bottom.

The diagram shown in FIG. 3 illustrates the highly schematic relationship between the control deviation and the clocking of the partial inverters 2 and 3. The upper polygon (Fig. 3a) represents the time course of the control deviation Ai (difference between ii and is) with associated tolerance bands Bi and B2. The two middle pulse-time diagrams b and c give the associated time course 5 of each Terminal of the load 4 applied potential U2 or U3 with respect to the minus busbar of the inverter 1 again. 3d illustrates the time profile of the voltage acting on the load. It can be clearly seen how the distribution of the switching work on both partial inverters is achieved by exchanging the tolerance bands.

The manner in which a device for multiphase inverters described above can be expanded is illustrated in FIG. 2, for example for a three-phase inverter. The same parts are provided with the same reference numerals as in FIG. 1. For standardization, all parts which are each assigned to one of the three phases R, S, T are also provided with the index R, S or T.

It can be seen that the modules inverter power section, tolerance band control device and tolerance band replacement are present for each phase; they are designated WLr, WLs, WLt, TBr, TBs, TBt or TUr, TUs or TUt. The circuit arrangement of the setting of the switching level NI acts in parallel on all three tolerance band exchange devices 25 TUr, TUs, TUt. The same applies to the proportional / integral controller PI, which also only exists once and acts in parallel on all three tolerance band control devices TBr, TBs, TBt, in the case of the hysteresis control. The same applies to the hysteresis setting via the direct current source 31.

30th

An additional component is a mean value or also a maximum value generator MG, which serves to form the mean or maximum value of the output voltages of the three frequency-voltage converters in the three tolerance band control devices TBr, TBs and TBt, which average or maximum value is fed to the proportional / integral controller PI.

Due to the special design of the tolerance band control, it is necessary that the multi-phase load, e.g. ei-40 ne electrical machine, galvanically isolated load circuits - in the case of an asynchronous machine, so galvanically isolated stator windings.

v

3 sheets of drawings

Claims (7)

662 019 1. Inverter arrangement with a two-pulse inverter with a device for tolerance band control with two partial inverters (2, 3), between which a load (4) is connected, and with comparison of the setpoint current (is) and the actual value current (ii) to achieve it a predetermined current profile through the load (4), characterized in that that each of the two partial inverters (2, 3) is connected to its own comparator (6, 7) and that the switching levels of the individual comparators (6, 7) with the associated switching arrangements for hysteresis setting and hysteresis formation (31, 34, 35) for both partial inverters (2, 3) are mutually displaced. 2. Inverter arrangement according to claim 1, characterized in that a common DC voltage source (16) is provided for the common displacement of the switching levels of the comparators (6, 7) and that the DC voltage source is in operative connection with a changeover switch (17). 2nd PATENT CLAIMS 3. Inverter arrangement according to claim 1 or 2, characterized in that means (18, 19, 20, 21, 22) for mutual exchange of the two tolerance bands assigned to the partial inverters (2, 3) after a predetermined number of switching operations of the partial inverters (2, 3 ) are provided. 4. Inverter arrangement according to claim 2 and 3, characterized in that the changeover switch (17) is connected to a bistable multivibrator (18), the preparation inputs of which partly directly, partly via inverters (19, 20) and whose clock input via an exclusive-or- Circuit (22) are connected to the outputs of the comparators (6, 7). 5. Inverter arrangement according to one of claims 1 to 4, characterized in that a potentiometer (31) is provided for hysteresis setting. 6. Inverter arrangement according to one of claims 1 to 4, characterized in that a controller (23) is provided for hysteresis setting, the inputs of which are on the one hand with a direct current source (33) as setpoint generator for the switching frequency and on the other hand with a frequency-voltage converter (36) are connected as actual value transmitters of the switching frequency. 7. Inverter arrangement according to one of claims 1 to 6, characterized in that in multi-phase operation with electrically isolated load circuits, each individual load is assigned a two-pulse inverter (WLr, WLs, WLt), each with two partial inverters, that each inverter has means for exchanging tolerance bands ( TUr, TUs, TUt) and comparators cooperating with them (TBr, TBs, TBt) has that the DC source for shifting the switching level (16) and for hysteresis setting (31) or the controller (23) together for all phases ( R, S, T), and that to form the actual frequency value, an average or maximum value generator (MG) for the output voltages of all frequency-voltage converters of the individual phase inverters (WLr, WLs, WLt ) is provided (Fig. 2).