JP4962766B2 - AC / AC direct converter controller - Google Patents

Description

  The present invention provides an AC / AC direct converter that directly converts an AC input voltage into an AC voltage of an arbitrary magnitude and frequency by turning on / off a bidirectional switch without using a large energy buffer, and an output voltage of the direct converter The present invention also relates to a control device that can reduce distortion of input current.

Conventionally, a matrix converter is known as an example of this type of direct converter.
FIG. 8 is a block diagram showing the configuration of a general matrix converter control device.
First, in the main circuit of FIG. 8, 1 is a three-phase AC power source, 2 is a matrix converter, 3 is a load such as an AC motor, and the control device detects the input voltage of the matrix converter 2 and performs various calculations. The on / off signal for the semiconductor switching elements constituting the bidirectional switch in the matrix converter 2 is generated by PWM control.

Hereinafter, a specific configuration and operation of the control device will be described.
First, the input voltage of the matrix converter 2 detected by the input voltage detection means 4 is sampled at a predetermined period (for example, a PWM control carrier period), converted into a digital signal by the AD conversion means 5 and input to the microprocessor 11. . Here, the input voltage detection means 4 is comprised, for example by the instrument transformer, the analog filter, the sample holder, etc.

The input current control signal calculation means 8 in the processor 11 generates an input current control signal based on the detected input voltage value from the AD conversion means 5, and this control signal is turned on / off time ratio calculation means 9 together with the output voltage command value. To enter.

  The on / off time ratio calculation means 9 synthesizes the input current control signal and the output voltage command value to calculate the on / off time ratio of the switching elements in the matrix converter 2 and outputs this as a command value to the PWM generation means 10. The PWM generation means 10 compares the on / off time ratio with the carrier to generate an on / off signal for each switching element in the matrix converter 2. By controlling the switching element using this on / off signal, the matrix converter 2 directly cuts the AC input voltage to generate an AC voltage having a predetermined magnitude and frequency, and outputs it to the load 3.

  Here, as the control device, a digital arithmetic unit such as a microprocessor 11 or a DSP (Digital Signal Processor) is generally used as shown in the figure, and the on / off time ratio is set using the sampling value of the input voltage detection value. I'm calculating.

  In general, digital control uses a fixed value for each sampling period, so there is a time delay between the actually detected voltage value and the sampled value after AD conversion. In addition, in order to prevent malfunction due to the influence of noise, a low-pass filter or the like is used in the analog signal detection unit, so that there is also a time delay of input / output signals due to this filter. Furthermore, calculation delay due to a digital arithmetic unit such as a microprocessor cannot be ignored.

If the sampling value used for control has a time delay, an error occurs between the actual detection value and the control performance deteriorates. As a result, for example, in the control of the matrix converter, an error appears in the output voltage and the input current.
Here, the error in the output voltage causes torque ripple or rotation unevenness when the AC motor is driven as the load 3 and causes the loss of the motor to increase. In addition, an error in the input current appears as distortion of the input current, which is not preferable because it causes overheating of the transformer connected to the input side of the matrix converter and malfunction of other power supply side devices.

In order to correct the time delay of the sampling value as described above, for example, the phase angle of the input voltage is calculated by PLL (Phase Locked Loop) or arc tangent, and the AC input voltage detection value is coordinated based on the obtained phase angle. In addition to conversion to a DC amount, as described in Equation 1 of Patent Document 1, the correction angle is added to the obtained phase angle to advance the phase angle, and the DC angle is calculated using this phase angle. It is conceivable to restore the AC input voltage detection value to an AC signal again.
The correction angle may be set in consideration of sampling delay, time delay due to a filter, and the like.

Japanese Patent Laying-Open No. 2005-168197 (paragraphs [0027] to [0029], FIG. 2, FIG. 4 etc.)

However, the above-described method requires a complicated calculation such as PLL and arc tangent to detect the phase angle of the input voltage, which increases the load on the calculation device and increases the cost.
Furthermore, even if the arctangent calculation is tabulated to reduce the calculation amount, a large-capacity storage means is required. In addition, since there is a discontinuous point in the arctangent value, there is a problem that high accuracy cannot be expected.

  Therefore, the problem to be solved by the present invention is to reduce the calculation load by eliminating the calculation of the phase angle of the input voltage, to avoid the increase in storage capacity due to the tabulation, and to predict the input voltage with high accuracy by simple calculation. Thus, it is to provide a control device with improved control performance of output voltage and input current at low cost.

In order to solve the above-mentioned problem, the invention described in claim 1 includes a bidirectional switch directly connected between an AC power source and a load, and the bidirectional switch has an on / off time ratio according to an output voltage command value. In the AC / AC direct converter that converts the AC input voltage directly into an AC voltage of any magnitude and frequency by switching
Means for detecting an input voltage of the converter;
Means for sampling the detected input voltage at a predetermined period;
Storage means for storing sampling values from the current sampling value of the input voltage to N (N is an integer of 1 or more) times before;
A predicting means for predicting a future sampling value using the sampling value of N times before and the current and previous sampling values;
Means for calculating the on / off time ratio using a future sampling value predicted by the prediction means ,
The prediction means includes
Means for calculating a difference value between the sampling value N times before and the current sampling value;
Means for outputting the sum or difference of the difference value and the current sampling value as a future sampling value in accordance with the magnitude relationship between the current sampling value and the previous sampling value .

According to a second aspect of the present invention, in the control apparatus for an AC / AC direct converter according to the first aspect, the prediction unit includes a multiplication unit that multiplies the future sampling value or the difference value by a predetermined gain . Is.

According to a third aspect of the present invention, in the control device for an AC / AC direct converter according to the first or second aspect , a band-pass filter is provided before or after the prediction means.

  According to the present invention, the time delay of the sampling value of the input voltage for directly controlling the converter is corrected by simple four arithmetic operations and a small amount of stored data without using the phase angle calculation of the input voltage or a huge table. Therefore, it is possible to provide a control device that can improve control performance and reduce costs.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
First, FIG. 1 is a block diagram showing a first embodiment of the present invention. The same reference numerals are given to the same components as those in FIG. Below, it demonstrates centering on a different part from FIG.

In FIG. 1, for example, an input voltage detection value sampled at a PWM control carrier period and converted into a digital signal by the AD conversion means 5 is input to the storage means 6 in the microprocessor 11A. The storage means 6 is configured to store the current sampling value of the input signal and the sampling value of N (N is an integer of 1 or more) times before. Note that the previous sampling value (N = 1) is included in the sampling value N times before.
Here, if the set value of N is set to a relatively small value, a storage unit 6 having a small capacity can be used.

Next, the predicting means 7 predicts a future sampling value (several times later) using the storage data of the storage means 6, and the input current control signal creating means 8 uses the predicted value to input the input current. Create a control signal. Although there are various methods for calculating the input current control signal, the present invention is not particularly limited.
As for the input current control signal generating means 8 and subsequent ones, the on / off time ratio calculating means 9 calculates the on / off time ratio and outputs it to the PWM generating means 10 as in FIG.

FIG. 2 is a block diagram of the prediction means 7.
In FIG. 2, the addition / subtraction means 701 calculates the difference value between the sampling value N times before (referred to as the value before N times) stored in the storage means 6 and the current sampling value (referred to as the current value) to obtain an absolute value. The calculating means 702 calculates the absolute value of the difference value. This absolute value is input to the subsequent addition / subtraction means 703 and 704, and the difference between the current value and the current value is calculated, and these sum and difference are given to the switching terminal side of the switching means 706, respectively. It is done.

On the other hand, the current value and the previous value stored in the storage unit 6 are input to the comparison unit 705, and the comparison unit 705 determines the magnitude relationship between the current value and the previous value and sends a switching signal to the switching unit 706. Is output.
The operation of the switching means 706 by the switching signal is as shown in parentheses in FIG. 2. When the current value> the previous value, the output of the addition / subtraction means 703 (the sum of the absolute value of the difference value and the current value) is obtained. If the current value is less than the previous value, the output of the addition / subtraction means 704 (the difference between the current value and the absolute value of the difference value) is selected as the predicted value. 1 to the input current control signal generating means 8.

When the above operation is expressed by a mathematical expression, the mathematical expression 1 is obtained.
[Formula 1]
V _det = V _det0 + ABS (V _detN −V _det0 )
(When V _det0 > V _det1 ),
V _det = V _det0 −ABS (V _detN −V _det0 )
(When V _det0 <V _det1 )
However, V _{det is a} predicted value, V _{det 0 is a} current value, V _{det 1 is a} previous value, V _det N is a value N times before, and ABS is an absolute value operator.

FIG. 3 shows the relationship between the sampling values in Equation 1 on the voltage waveform. FIG. 3A shows the case of V _det0 > V _det1 , and FIG. 3B shows the case of V _det0 <V _det1 . It is. Here, the predicted value V _det is set as a sampling value after two times, and the current value V _det0 , the previous value V _det1 , and the sampling value V _det2 before two times are used for the _prediction .
That is, in this embodiment, depending on the tendency of increase or decrease the current value _{V DET0} from the previous value _{V det1,} a difference value between N (Fig. 3, N = 2) times the previous value _{V DETN} and the present value _{V DET0} added to the current value _{V DET0,} or by subtracting from the current value _{V DET0,} is predictive sampling value _{V det} after N times.

FIG. 4 is a waveform diagram showing a comparison result between a predicted value (predicted value after two times) and a theoretical value (theoretical value after two times) of the sampling value after two times of the input voltage obtained by this embodiment. The current value and the sampling value two times before are also shown.
In FIG. 4, the input voltage waveform is displayed in a unit method using a cosine wave, the sampling period is 500 μs, and the cosine wave period is 20 ms. As described above, in order to predict the sampling value after the second time, N = 2, and the sampling value before the second time is used together with the current value and the previous value for the prediction.

According to FIG. 4, the predicted value obtained by Equation 1 is almost identical to the theoretical value especially near zero of the cosine wave. In addition, although there is a slight error between the predicted value and the theoretical value near the top of the cosine wave, a predicted value almost equal to the theoretical value is obtained.
Here, the sampling value after the second sampling is predicted in consideration of the sampling delay. However, depending on the control device, the sampling delay time varies depending on the sampling time, such as three samplings or one sampling. Is predicted, the value of N in Equation 1 may be changed according to the number of times.

FIG. 5 is a waveform diagram showing simulation results when the present invention (this embodiment) is applied to the control of the matrix converter and when it is not applied. From the top, the input voltage, input current, and output voltage waveforms of the matrix converter are shown. Is shown.
When the present invention is not applied, since the sampling value is not predicted and the sampling delay directly affects the control, distortion appears particularly in the input current. This is because a numerical error caused by a time delay is large between the input voltage detection value used for the input voltage and the waveform of the input current is deteriorated particularly near the switching of the input voltage.
On the other hand, by applying the present invention to predict the sampling value, it can be seen that a good waveform without distortion is obtained particularly for the input current.

Next, a second embodiment of the present invention will be described.
FIG. 6 is a block diagram of the prediction means 7 in the second embodiment. The output of the switching means 706 shown in FIG. 2 is input to the gain multiplication means 707 and multiplied by the gain k, and the output is used as a predicted value. This is different from FIG.
In this embodiment, the amplitude of the predicted value is corrected to obtain a waveform closer to the theoretical value in order to reduce the error between the predicted value and the theoretical value near the vertex of the cosine wave in FIG. is there.

The correction gain k may be set to a ratio between the maximum amplitude value of the input signal and the maximum amplitude value of the predicted value as shown in Equation 2. As the maximum amplitude value, a point at which the absolute value of the amplitude becomes maximum in at least one cycle may be stored.
[Formula 2]
k = maximum amplitude value of input signal / maximum amplitude value of predicted value

Although not shown, a correction gain multiplication unit 707 is arranged on the output side of the absolute value calculation unit 702 (the difference value (V _detN −V _det0 ) is multiplied by the correction gain k) and the correction gain k is _reduced to a decimal number. By setting and obtaining the predicted value by the following Equation 3, the time delay in the elements other than the digital arithmetic unit can be corrected, such as the time delay in the analog circuit such as the input voltage detecting means 4 in FIG. it can.
[Formula 3]
V _det = V _det0 + k × ABS (V _detN −V _det0 )
(When V _det0 > V _det1 ),
V _det = V _det0 −k × ABS (V _detN −V _det0 )
(When V _det0 <V _det1 )

  For example, when the time delay in the analog circuit is 50 μs and the sampling period is 100 μs, and the delay is one sampling period, the delay of the analog circuit is also considered by setting the correction gain k to 1.5. Therefore, it is possible to reduce the error from the actual detection value and to improve the control performance.

Next, FIG. 7 is a block diagram showing a third embodiment of the present invention.
In Equation 1 and Equation 3 described above, since the seeking predicted value V _det in accordance with the magnitude relationship between the current value V _DET0 and the previous value V _det1, and distortion or ripple is superimposed on the input signal to the prediction means 7 If this is the case, an error will occur in the case of large and small relations, and as a result, the calculation accuracy of the predicted value may decrease.

Therefore, in the third embodiment, as shown in FIG. 7, for example, a bandpass filter 20 is provided in the input stage of the microprocessor 11B, and only the desired frequency component is extracted from the input signal by the bandpass filter 20. And ripples are removed, and the magnitude relationship between the current value V _det0 and the previous value V _det1 is accurately detected to reduce the error of the predicted value. If the band-pass filter 20 is provided on the output side of the prediction means 7, a smooth sine wave having a desired frequency component can be obtained as a predicted value.

  The pass band of the band pass filter 20 may be set according to the frequency of the input voltage. For example, if the frequency of the input voltage is 50 Hz, a band pass filter with 50 Hz as the center frequency may be used. Similar effects can be obtained by using a low-pass filter, a high-pass filter, or the like other than the band-pass filter.

1 is a block diagram showing a first embodiment of the present invention. It is a block diagram of the prediction means in FIG. It is a wave form chart for explaining Numerical formula 1 in a 1st embodiment. It is a wave form diagram which shows the predicted value, theoretical value, etc. which were obtained by 1st Embodiment. It is a wave form diagram which shows the simulation result when not applying the case where 1st Embodiment is applied to control of a matrix converter. It is a block diagram of the prediction means in 2nd Embodiment of this invention. It is a block diagram which shows 3rd Embodiment of this invention. It is a block diagram which shows a prior art.

Explanation of symbols

1: AC power supply 2: Matrix converter 3: Load 4: Input voltage detection means 5: AD conversion means 6: Storage means 7: Predictive means 701, 703, 704: Addition / subtraction means 702: Absolute value calculation means 705: Comparison means 706: Switching means 707: Gain multiplying means 8: Input current control signal calculating means 9: On / off time ratio calculating means 10: PWM generating means 11A, 11B: Microprocessor 20: Band pass filter

Claims (3)

A bidirectional switch directly connected between the AC power source and the load is provided, and the bidirectional switch is switched at an on / off time ratio according to the output voltage command value, whereby the AC input voltage is changed to an arbitrary magnitude and frequency. In an AC / AC direct converter that converts directly to AC voltage,
Means for detecting an input voltage of the converter;
Means for sampling the detected input voltage at a predetermined period;
Storage means for storing sampling values from the current sampling value of the input voltage to N (N is an integer of 1 or more) times before;
A predicting means for predicting a future sampling value using the sampling value of N times before and the current and previous sampling values;
Means for calculating the on / off time ratio using a future sampling value predicted by the prediction means;
Equipped with a,
The prediction means includes
Means for calculating a difference value between the sampling value N times before and the current sampling value;
According to the magnitude relationship between the current sampling value and the previous sampling value, a means for outputting the sum or difference of the difference value and the current sampling value as a future sampling value;
AC AC direct conversion device of the control device, characterized in that it comprises a. In the control apparatus for an AC / AC direct converter according to claim 1,
The prediction means includes
A control apparatus for an AC / AC direct converter, comprising a multiplying unit that multiplies the future sampling value or the difference value by a predetermined gain . In the control apparatus for an AC / AC direct converter according to claim 1 or 2,
A control apparatus for an AC / AC direct converter, characterized in that a band-pass filter is provided before or after the prediction means.

Images