RU2414800C1 - Method of scalar control of (3×3)-phase matrix frequency converter - Google Patents

Abstract

FIELD: electricity.

SUBSTANCE: scalar control method consists in pulse-duration control of three-phase voltage by means of fourfold alternating switching of valves at the period of carrier frequency only in two load phases, while in the third phase the valve remains ON throughout the repetition interval equal to 1/6 of period of output voltage. Switchings are performed at moments of equality of reference and three-phase modulating signals. The latter is obtained by multiplying the three-phase demand signal of harmonic shape by switching function, each phase of which is formed with alternation of various clock pulses of carrier frequency of unit amplitude, which divide the period of demand signal into six equal time intervals.

EFFECT: reducing commutation power losses and increasing maximum value of output voltage of amplitude of line-to-earth voltage of the network.

6 dwg

Description

The invention relates to a conversion technique, one of the applications of which is a variable frequency drive. In this area, matrix converters of frequency running on nine transistor switches with two-sided current conductivity are beginning to get widespread use. By analogy with autonomous voltage inverters (AIN), the 3-phase voltage and frequency output at the MFC output can be controlled using the scalar method at a high carrier frequency using pulse width modulation (PWM) methods (see Sidorov S.N. Matrix frequency converter is a scalar control object . - “Power Electronics”, No. 3, 2009, p.31).

In order to reduce switching power losses in transistor switches, the number of switching over the period of the carrier frequency in the AIN circuits is reduced from six to four. This problem can be solved by using the so-called 2-phase PWM, according to which the transistors switch alternately in only two phases of the load, while in the third phase the transistor remains in the entire repeatability interval equal to 1/6 of the output voltage period in the on state. Despite switching in only two phases, the output voltage, due to the connectivity property of the 3-phase load, maintains a symmetrical shape. Another advantage of the 2-phase version compared to the classic 3-phase version of the sinusoidal PWM is the better use of the mains supply voltage.

который получают потактными изменениями величины и формы гармонического 3-фазного задающего сигнала

той же частоты, согласно зависимости

с помощью 3-фазной переключательной функции

каждая фаза которой образована чередованием тех или иных тактовых импульсов z_i=1, 2, … 6 единичной амплитуды длительностью π/3, делящих период задающего сигнала X_A(t) с момента прохождения через нуль в положительную сторону на шесть равных промежутков времени.The closest solution is contained in the work of N.V. Donskoy and K.A. Chubukov “Comparison of Autonomous Voltage Inverters with Two-Phase Vector and Pulse-Width Modulation” - Dynamics of nonlinear discrete electrical and electronic systems: Materials of the 8th All-Russian nt.k. ., Cheboksary. Ed. CTU. 2009, pp. 246-252. The 2-phase PWM method presented in this work is carried out on the basis of the scalar principle of control at a carrier frequency that exceeds the frequency of the supply voltage of the network with a 4-fold switching of power transistors at each carrier frequency period occurring when the 1-phase reference signal generated in the modulator is equal sawtooth and 3-phase baseband signal

which is obtained by captive changes in the magnitude and shape of a harmonic 3-phase reference signal

the same frequency according to

using 3-phase switching function

each phase of which is formed by the alternation of certain clock pulses z _i = 1, 2, ... 6 of unit amplitude of duration π / 3, dividing the period of the driving signal X _A (t) from the moment it passes through zero to the positive side by six equal time intervals.

The listed advantages of a 2-phase PWM are achieved by giving modulating signals a special shape. However, as noted in the description of the prototype, these signals, providing a quasi-sinusoidal form of linear voltages, can lead to distortions in the shape of the phase voltage of the load.

The present invention allows to transfer the idea of a 2-phase PWM from the field of control of the AIN to the field of control of the MUF. The achieved result is presented in the form of reducing the number of switching power transistors on the period of the carrier frequency and increasing the conversion coefficient of the mains voltage. For this, the phase values of the modulating signals are proposed to be formed according to the dependencies:

x _a (t) = z _ai (t) · X _A (t) + z ₂ (t) -z ₅ (t);

x _b (t) = z _bi (t) · X _B (t) -z ₁ (t) + z ₄ (t);

x _c (t) = z _ci (t) · X _C (t) -z ₃ (t) + z ₆ (t).

Here:

z _ai (t) = (z ₁ (t) + z ₃ (t) + z ₄ (t) + z ₆ (t)); X _A (t) = µSin (kω ₀ t);

z _bi (t) = (z ₂ (t) + z ₃ (t) + z ₅ (t) + z ₆ (t)); X _B (t) = µSin (kω ₀ t-2π / 3);

z _ci (t) = (z ₁ (t) + z ₂ (t) + z ₄ (t) + z ₅ (t)); X _C (t) = µSin (kω ₀ t + 2π / 3),

where ω ₀ is the circular frequency of the network; k = ω / ω ₀ - the ratio of the frequency of the master signal to the frequency of the network; µ is the depth of voltage modulation.

The proposed control method can be implemented using MUF, a generalized diagram of the power and control parts of which are shown in figure 1. Figure 2 in an open form shows a block diagram of the formation of modulating signals (BF), the operation of which is illustrated by the diagrams in figure 3. In Fig.4, 5 presents the results of the verification of the proposed solutions by computer simulation. The diagrams in Fig.6 serve for a comparative assessment of the voltage conversion coefficient with one of the existing and new control methods.

The power part of the device is made in the form of a 3 × 3 matrix (1) on transistor switches V1-V9 with two-sided current conductivity, the horizontal buses of which are connected to the phases e _A , e _B , e _{C of the} supply network, and the vertical buses to circuits 3- phase load 2. The control part (3) contains a BF unit (4), the inputs of which serve to supply a 3-phase drive signal X _A (t), X _B (t), X _{C of a} harmonic shape, and the outputs to obtain a 3-phase modulating signal x _a (t), x _b (t), x _c (t) of a non-sinusoidal shape. There is also a modulator-distributor block (5), at the inputs of which a comparison of the modulating and reference x _on (t) signals coming from block (6), which is common for scalar control systems, results in the receipt of nine sequences of control pulses at the outputs of block (5) for keys in the matrix (1). According to FIG. 2, the BF unit contains a clock generator (7) z ₁ -z ₆ , a 3-channel analog multiplier (8), each channel of which is made on two-input multiplication circuits (9-14), with a 3-channel analog connected to the output an adder (15), each channel of which is a 6-input summing amplifier (16). At the outputs of the analogue adder, modulating signals x _a (t), x _b (t), x _c (t) are generated, which then go to the inputs of the modulator-distributor (5).

чередующиеся изменения которого с периодом π/3 происходят в каждый момент времени лишь в двух фазах. Как отмечалось, эти изменения осуществляются по гармоническому закону и происходят только в двух фазах нагрузки с амплитудой, зависящей от глубины модуляции µ=0-1. В то время как в третьей фазе модулирующий сигнал в течение каждого интервала повторяемости π/3 остается на постоянном максимальном уровне. Кроме фазных сигналов на фиг.3 представлена весьма близкая к синусоиде форма линейного модулирующего сигнала x_a(t)-x_b(t). Данные диаграммы свидетельствуют о том, что при µ=1 предлагаемый способ не ведет к существенным искажениям управляющих воздействий, в связи с чем снижения качества выходного напряжения МПЧ ожидать не следует.We consider the operation of the device using an example of one of the BF channels with a driving signal X _A (t) = µSin (kω ₀ t) at the input and a modulating signal x _a (t) at the output. From the diagram of figure 2 it is seen that the specified driving signal is fed to the first inputs of the analog multipliers and at the same time to the triggering input of the shaper (7). It is believed that each passage of X _A (t) through zero to the positive side leads to the appearance at the outputs of this shaper of a periodic sequence of clock pulses z ₁ -z _{6 of} unit amplitude dividing the period of the specified driving signal into six equal time intervals of duration π / 3. The arrival of clock pulses at the second inputs of the multipliers allows the master signal to pass through the BF at appropriate time intervals without distorting the shape of this signal. In this case, such an arrival of pulses z ₁ , z ₃ , z ₄ , z ₆ occurs at the second inputs of the multipliers (9, 11, 12, 14), which will be accompanied by a periodic appearance on the 2π / 3 intervals of the signal X _A (t) at the outputs these multipliers, and then at the output of the adder (16). In contrast to the indicated pulses from the remaining outputs of the shaper z ₂ , (-z ₅ ) are fed to both inputs of the multipliers (10, 13) and therefore will be repeated in shape taking into account the sign first at the outputs of these multipliers, and then at the output of the adder 16 in in the form of pulses of constant amplitude x _a (t) = ± 1 and duration π / 3. As a result, the signal x _a (t) takes on the period 2π the form of a piecewise function, which is formed by alternating sections 2π / 3 with a harmonic law of change and sections of duration π / 3 corresponding to the operation of the BP in the saturation mode. Similarly, the work of other channels of the block (4). The result is a 3-phase baseband signal.

whose alternating changes with a period of π / 3 occur at each moment of time in only two phases. As noted, these changes are carried out according to a harmonic law and occur only in two phases of the load with an amplitude depending on the modulation depth µ = 0-1. While in the third phase, the modulating signal remains at a constant maximum level during each repeat interval π / 3. In addition to phase signals, Fig. 3 shows a linear modulating signal x _a (t) -x _b (t) that is very close to a sinusoid. These diagrams indicate that at µ = 1, the proposed method does not lead to significant distortions of the control actions, and therefore, one should not expect a decrease in the quality of the output voltage of the MUF.

Confirmation are presented in figures 4, 5, the results of computer simulation. The diagrams in figure 4 illustrate the proposed method in the next stages of the formation of control signals and output voltages. It can be seen that a comparison of the modulating x _a (t), x _b (t), x _c (t) and the reference x _on (t) signals allows the modulator to obtain three (by the number of phases) sequences of control pulses for transistor switches V1,4, 7; V2.5.8 and V3.6.9 respectively. The subsequent distribution and supply of these pulses from block (5) to power transistors leads to the appearance of a 3-phase voltage e _a , e _b , e _c and current i _a , i _b , i _c in the load circuits. For the purpose of comparison, Fig. 5 shows similar diagrams of a 3-phase sinusoidal PWM. A comparison shows that the proposed control method reduces the number of switching operations in each period from six to four, thereby reducing the switching power losses in power transistors by an appropriate number of times. We show that at the same time, an increase in the maximum voltage occurs, which, as is known, is limited in frequency converters by sinusoidal PWM processes. A comparative analysis of the maximum voltages in the conditions of a 3- and 2-phase PWM is feasible by calculating the limiting value of the average cycle voltage in one of the load phases. The diagrams in Fig. 6 show that the condition for obtaining such a voltage is a symmetrical arrangement of the clock cycle with respect to the peak of the cosine wave of the modulating signal of maximum amplitude. A preliminary search of the switching times t ₁ , t ₂ , t ₃ at the equal points of the modulating x _a (t), x _b (t), x _c (t) and the reference x _on (t) signals allows us to calculate the indicated voltage by the formula:

and then compare the obtained value with the amplitude of the phase voltage of the network. Calculations show that for variations in the carrier frequency in the range of 1-10 kHz, the conversion coefficient, as the ratio of the indicated voltages, in the conditions of a 3-phase PWM is in the range u _dm / E _fm = 0.83-0.86. We notice that in the conditions of a 2-phase PWM, a decrease in the number of switching is accompanied by the same decrease in the fraction of zero pauses in the composition of the output voltage pulses. In this regard, the voltage conversion coefficient in the 2-phase PWM mode rises, as the calculations showed, to the level of 0.87 and higher. Confirmation of this property is the increased amplitude of the load current in the diagrams of figure 4 in comparison with similar diagrams in figure 5.

Claims (1)

A control method for a 3 × 3-phase matrix frequency converter, carried out by pulse-width modulation of voltage in 2 phases of the load at the carrier frequency, exceeding the frequency of the supply voltage of the network with 4-fold switching of power transistors for each period that occurs at equal times generated in the modulator 1- phase reference signal of an equilateral sawtooth shape and a 3-phase modulating signal

, which is obtained by captive changes in the magnitude and shape of a harmonic 3-phase reference signal

the same frequency, according to

using 3-phase switching function

, each phase of which is formed by the alternation of certain clock pulses z _i = 1, 2, ... 6 of unit amplitude of duration π / 3, dividing the period of the driving signal X _A (t) from the moment of passing through zero in the positive direction, into six equal time intervals characterized in that the phase values of the modulating signal are obtained according to the algorithm:
x _a (t) = z _ai (t) · X _A (t) + z ₂ (t) -z ₅ (t);
x _b (t) = z _bi (t) · X _B (t) -z ₁ (t) + z ₄ (t);
x _c (t) = z _ci (t) X _C (t) -z ₃ (t) + z ₆ (t),
here
z _ai (t) = (z ₁ (t) + z ₃ (t) + z ₄ (t) + z ₆ (t)); X _A (t) = µsin (kω ₀ t);
z _bi (t) = (z ₂ (t) + z ₃ (t) + z ₅ (t) + z ₆ (t)); X _B (t) = µsin (kω ₀ t-2π / 3);
z _ci (t) = (z ₁ (t) + z ₂ (t) + z ₄ (t) + z ₅ (t)); X _C (t) = µsin (kω ₀ t + 2π / 3),
where ω ₀ is the circular frequency of the network; k = ω / ω ₀ - the ratio of the frequency of the master signal to the frequency of the network; µ is the depth of voltage modulation.

Images