JPH07194130A - Controller for pulse-width modulation inverter - Google Patents

Abstract

PURPOSE:To smoothly perform switching between asynchronous modulation and synchronous modulation and continuously control the output voltage of a pulse-width modulation inverter in a whole area so that a comfortable ride can be obtained by adjusting the phase angle of a basic modulated wave at the time of switching the control between multiple-pulse control and one-pulse control. CONSTITUTION:A three-level inverter is provided with a means which outputs a pulse- width-modulated signal synchronized with the modulated-wave signal of the inverter is provided for one-pulse control under which the inverter outputs the highest voltage and a means which outputs a pulse-width-modulated signal asynchronous to the modulated-wave signal of the inverter for multiple-pulse control other than the one- pulse control. The pulse-width modulated signal synchronous with the modulated-wave signal of the inverter under the one-pulse control is formed so that the turn-on or turn-off phase angle of the signal can become coincident with that of a basic modulated wave. The pulse-width-modulated signal asynchronous with the modulated-wave signal of the inverter under the multiple-pulse control is formed so that the turn-on or turn-off phase of the signal can become coincident with that of the commanding value of an instantaneous voltage to be outputted.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a controller for a three-level inverter, and more particularly to a controller for a pulse width modulation type inverter.

[0002]

2. Description of the Related Art A so-called inverter train, in which an induction motor is driven by a variable voltage / variable frequency (VVVF) inverter control based on a pulse width modulation system, has been put into practical use, but recently, an inverter is used not only for voltage but also for locomotives. Drive systems are becoming mainstream.

In the current inverter system for driving a vehicle, there is a so-called pulse width modulation (PWM) control system in which a pulse signal is generated by comparing a modulated wave signal which is a basic component of an output voltage with a carrier wave signal such as a triangular wave. It has been adopted from the past.

In the PWM control method, there are a synchronous modulation method in which the zero-cross points of the modulated wave signal and the carrier wave signal are always controlled to coincide with each other, and an asynchronous modulation method in which the carrier wave is constant regardless of the modulated wave.

In the present vehicles, most of the vehicles are two-level inverters that control two-level voltage using a GTO thyristor. However, in the two-level GTO inverter, there is a restriction on the switching frequency, and asynchronous modulation is performed in the minute voltage range including zero voltage, and the synchronous modulation system is switched in the other voltage ranges. There is a big problem, and the change in sound quality due to the switching of pulse modes is also a problem.

On the other hand, the three-level inverter has a feature that the number of steps of the output voltage is higher than that of the two-level inverter and the apparent switching frequency is higher, so that problems in the two-level inverter such as reduction of electromagnetic noise can be solved. Therefore, a three-level inverter system for a vehicle, which uses a high voltage IGBT as a main circuit element, has begun to be developed.

PWM control of a three-level IGBT inverter is shown in FIG. In other words, the dipolar modulation that outputs positive and negative pulses alternately via zero voltage is introduced to control the minute voltage, and the unipolar modulation that outputs the pulse train of the same polarity every half cycle, which is the most common PWM control of a three-level inverter. In order to achieve smooth transition by suppressing fluctuations in the inverter output current, partial dipolar control in which dipolar modulation and unipolar modulation coexist is introduced as a transition to control. Furthermore, overmodulation is used to control the output voltage continuously up to one pulse region.

Thus, (a) dipolar modulation, (b)
By performing partial dipolar modulation, (c) unipolar modulation, (d) overmodulation, and (e) single pulse control, the voltage can be continuously controlled over the entire area. In this case, along with the increase in the switching frequency, (a) to (d) are asynchronous modulation systems, and (e) only one pulse is a synchronous modulation system.
The control of (a) to (d) is called multi-pulse control in contrast to the one-pulse control.

[0009]

As described above, in the three-level IBGT inverter, since the asynchronous modulation is performed by the multi-pulse control and the synchronous modulation is performed by the one-pulse control, the switching frequency between the asynchronous modulation and the synchronous modulation is high and the accuracy is high. Switching control is required.

An object of the present invention is to provide a controller for a pulse width modulation type inverter that smoothly switches between asynchronous modulation and synchronous modulation, continuously controls the output voltage in all control areas, and obtains a smooth riding comfort. To provide.

[0011]

In order to achieve the above object, the present invention is such that when the phase angle of the fundamental modulated wave of the inverter is switched from multi-pulse control to single-pulse control, it is 1 / th of the sampling period of the phase angle calculation. When switching from single-pulse control to multi-pulse control by subtracting the equivalent of two cycles and halving the sampling cycle of the phase angle calculation,
The phase equivalent is advanced by adding the equivalent of the cycle.

[0012]

The time delay in the sampling calculation can be corrected by adjusting the phase angle of the basic modulated wave when switching between the multi-pulse control and the single-pulse control, so that smooth control can be realized.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A control device for a pulse width modulation type inverter according to the present invention will be described in detail below with reference to embodiments shown in the drawings.

FIG. 1 shows a main circuit configuration (in the case of three phases) of a three-level inverter device for driving a vehicle to which the control device for a pulse width modulation type inverter of the present invention is applied. In FIG. 1, reference numeral 60 is a train line that is a DC voltage source, 61 and 62 are capacitors that are divided (divided) to create an intermediate point N (hereinafter referred to as a neutral point) from the voltage of the DC voltage source 60, 70
˜73, 80 to 83, 90 to 93 are self-extinguishing switching elements (in this example, I
Although it is a GBT, it may be a GTO thyristor, a transistor, etc.), 74, 75, 84, 85, 94 and 95 are auxiliary rectifying elements for deriving the neutral point potential of the capacitors.
The load is shown for the induction motor 10.

The operation of the switching arms 7 to 9 that can operate independently for each phase is performed by the switching arm 7
The basic operation will be described by taking as an example.

Voltages ed ₁ and ed _{2 of} capacitors 61 and 62
As a completely smooth DC voltage source, ed ₁ = ed ₂ = Ed /
2 (Ed: total DC voltage).

At this time, the switching elements 70 to 73 are on / off controlled as shown in Table 1 to obtain the three-level output voltage e of Ed / 2, 0, -Ed / 2 at the AC output terminal U. .

[0018]

[Table 1]

S _p , S _o , S _n and S are switching functions expressing the conduction states of the switching elements 70 to 73 by 1, 0 and -1, and the output voltage e is

[0020]

[Equation 1]

It is represented by

Although e has a waveform obtained by combining pulse-like voltages having a size of Ed / 2, 0, -Ed / 2, in general,
e is pulse width modulation (PWM) of S so that it approaches a sine wave
Control.

Here, the PWM control will be described before the description of the present embodiment.

The PWM control of the three-level inverter is, as described above, (a) dipolar modulation, (b) partial dipolar modulation, (c) unipolar modulation, (d) overmodulation, (e).
One pulse control is performed, and the output voltage is controlled almost continuously from 0 to 100%.

Generally, the inverter output voltage command E * is
It is set according to the inverter frequency command Fi * as shown by the solid line in FIG. This inverter output voltage command E *
From the DC voltage Ed and the fundamental wave amplitude command (modulation rate) A (0 ≦ A ≦ 1) in the sine wave modulation region,

[0026]

[Equation 2]

Is given by From this, the basic modulated wave command a
From modulation factor A and phase θ

[0028]

[Formula 3] a = Asin θ (Formula 3) θ = 2πFi * t Fi *: inverter frequency command, t: time.

In order to enable continuous transition between the dipolar modulation of (a) and the unipolar modulation of (c), the basic modulation wave command a of Equation 3 is divided into two, and the positive bias modulation wave command a _bp , The negative bias modulated wave command a _bn is created as in the following equation.

[0030]

[Equation 4]

By setting the bias amount B, (a) dipolar modulation to (c) unipolar modulation are realized.
Next, each modulation will be described.

The dipolar modulation will be described with reference to FIG.

When the bias amount B is in the range of A / 2≤B <0.5, positive and negative pulses are generated by the two layers of modulated wave commands of positive and negative bias modulated wave commands a _bp and a _bn. Thus, the inverter output voltage can be controlled to zero by ensuring a predetermined minimum on-pulse width.

Partial dipolar modulation will be described with reference to FIG.

When the bias amount B is in the range of 0 <B <A / 2, partial dipolar modulation is performed. Depending on the magnitude of the bias amount B, the bias modulation wave commands a _bp and a _bn will cross zero, but during this period, a negative (or positive) voltage is realized by the positive (or negative) side pulse. Since this is not possible, the output voltage decreases as compared with the fundamental modulation wave. Therefore, in this period, the positive and negative modulated wave commands a _p and a _n are set as follows to compensate for the insufficient voltage.

[0036]

[Equation 5]

[0037]

[Equation 6]

However, in FIG. 5, for easy understanding, a _p and a _n are represented by positive signals. Where period I
Is dipolar modulation and period II is unipolar modulation. With this correction, an output voltage equal to the fundamental modulated wave is obtained,
Partial dipolar modulation in which dipolar modulation and unipolar modulation are mixed can be realized.

Unipolar modulation will be described with reference to FIG.

The bias amount B = 0 is in the unipolar modulation region, and the positive and negative bias modulation wave commands a _bp and a _bn are the same. Positive and negative modulated wave commands a _p and a _n are given by the following equations.

[0041]

[Equation 7]

[0042]

[Equation 8]

However, in FIG. 6, both a _p and a _n are displayed so as to be positive signals. Overmodulation control also belongs to this area.

The overmodulation control will be described with reference to FIG.

The overmodulation control is performed by the positive and negative modulation wave commands a _p ,
in which the amplitude of a _n exists one or more regions, in this one or more regions, eliminating the drop to zero duration of the pulse, perform the PWM control only near the zero cross of the modulation wave.
However, the actual output voltage becomes smaller than that of the fundamental modulation wave, and the output voltage does not increase linearly with respect to the linear increase of the modulation factor A.

Therefore, the setting of the modulation factor A is made non-linear to correct the output voltage. If the switching frequency in the PWM control portion of overmodulation is sufficiently high, the relationship between the fundamental wave effective value E of the output voltage and the modulation rate A can be expressed by the following equation.

[0047]

[Equation 9]

If the required modulation factor A is calculated in advance from Equation 9 and the modulation factor A is changed non-linearly with respect to the output voltage command E * of the inverter as shown in FIG. The output voltage command E * can be linearly controlled.

From the above (a) dipolar modulation to (d) overmodulation, changes in the bias amount B and correction of the modulated wave command,
Although there are factors such as non-linearity of the modulation rate, the method of creating the PWM signal is basically the same.

The one-pulse control will be described with reference to FIG.

In the three-level PWM control, the width of the zero voltage period can be set arbitrarily, so that the output voltage can be adjusted during the one-pulse control. Assuming that the timing angle of the output voltage rise is α, the fundamental wave effective value E of the inverter output voltage is
Can be expressed by the following equation using α.

[0052]

[Equation 10] E = E _max · cos α (Equation 10) The coverage range of the voltage when α ≦ 30 ° is 86.6% to 100
%, The output voltage can be continuously controlled from the above-mentioned overmodulation in the range of α ≦ 30 °.

As shown in FIG. 9, the setting of the pulse width is such that the intersection of the inverter output voltage command E * and cos θ is the output voltage rise timing angle α, and the intersection of −cos θ is
It is the fall timing angle β. As described above, the method of creating the PWM signal of the single pulse control is completely different from (a) dipolar modulation to (d) overmodulation.

Next, FIG. 2 shows an embodiment in which the present invention is applied to control of a three-level inverter device for driving a vehicle.
In FIG. 2, 1 indicates from the inverter frequency command Fi *,
It is a phase calculation unit that calculates the fundamental wave phase of the output voltage of the inverter. The inverter frequency command Fi * is obtained by calculating the rotation frequency of the electric motor from the output of a rotation frequency detector attached to a wheel (not shown) and adding it to the slip frequency command. The output 101 of the phase calculator 1 is input to a phase adjuster 17, which will be described later, and the output 117 is converted into a sine wave signal 102 by the sine wave converter 2. Reference numeral 3 is a calculation unit of the modulation factor A for obtaining the required instantaneous output voltage of the inverter from the output voltage command E * of the inverter. The output voltage command E * is a command value of the inverter output voltage obtained from the above-mentioned inverter frequency command Fi * and the filter capacitor voltage and the motor current (not shown) so that the V / F characteristic becomes constant. is there. The modulation factor A has a non-linear relationship with the output voltage command E * in the overmodulation region, as shown in FIG. The output 103 of the modulation factor calculation unit 3,
The sine wave signal 102 is multiplied by the multiplier 4 to obtain the basic modulated wave command 104. Reference numeral 5 denotes a bias setting unit that adds or subtracts the basic modulation wave command 104 to / from the modulation factor A output 103 to set a bias amount B for continuously realizing dipolar modulation, partial dipolar modulation, and unipolar modulation.
The basic modulated wave command 104 and the bias amount 105 are added to the adder 6
To obtain a positive bias modulated wave command 106 and a subtractor 7 to subtract a negative bias modulated wave command 107, respectively. In order to correct the output voltage described above, the positive bias modulated wave command 106 is divided into positive and negative signals by the distributor 601 that obtains the positive signal 1061 and the distributor 602 that obtains the negative signal 1062, and the negative bias modulated wave command 107 is positive. 701 for obtaining the signal 1071 of
The positive and negative signals are sorted by 2 and the positive signals 1061 and 1 by the adder 8
071 is added, and the positive side modulation wave command a _p 108 and the negative signals 1062 and 1072 are added by the adder 9 to create the negative side modulation wave command a _n 109. A pulse timing calculation unit 10 calculates pulse timings of the switching functions S _p and S _n based on the positive and negative modulation wave commands 108 and 109.

The timer output section 11 described later is a timer section for pulse generation. In general pulse generation, a means for setting the rising or falling of the pulse for each sampling cycle T _s is generally used. The T _S generator 18 outputs the sampling period T _S 118 from the clock CLK of the controller. This T _S is also a factor that determines the switching frequency of the inverter, and the switching frequency F of the inverter is
_sw is the following formula.

[0056]

F _sw = 1 / (2 × T _s ) (Equation 11) FIG. 10 shows a method of generating timings of switching functions S _p and S _n in dipolar modulation (region I) and unipolar modulation (region II). Shows. The rising and falling _edges of the pulse timing are set for each sampling cycle T _S. First, at time T ₁ , the rising timings T _pup and S _{n of} S _p are set.
_Fall timing T _ndn of positive and negative side modulated wave command 10
It is calculated by the following equation using 8,109.

[0057]

[Equation 12]

[0058]

[Equation 13]

[0059] In the next time T _2, obtains the falling timing T _pdn of S _p, the rising timing T _Nup of S _n by the following equation.

[0060]

[Equation 14]

[0061]

[Equation 15]

The processing of T ₁ and T ₂ is performed by sampling period T
S _p and S _n can be created by alternately performing every _s .

However, T _pup , T _ndn , T _pdn , and T _nup correspond to the output 100 of the pulse timing calculator 10 in FIG. 2, and the actual operation is the pulse mode changeover switch 15.
Is sent to the timer output unit 11 via. Timer output section 1
1 also operates in synchronization with the sampling cycle T _s , the calculation results T _pup , T _ndn , T _pdn , and T _nup are the next time, for example, the calculation result at the time T _{1 is} at the timing of the time T ₂ . It will be output by the timer. Therefore, the output voltage is delayed by the sampling period T _s as shown in FIG. This is a delay that naturally occurs when the timer operates in synchronization with the sampling cycle.

Returning to FIG. 2 again. As described above, the one-pulse control sets the timing angles of the rising and falling edges of the switching functions S _p and S _{n in} a half cycle in synchronization with the fundamental modulated wave. In 12 of cos ^-1 E * operation part, output voltage command E
To find the intersection of * and cos θ, find cos ^-1 E *.
In step 12, calculation may be performed, or a table may be prepared in advance. The output 112 of 12 is the switching function S _p , S _n
Is the rising angle α and the falling angle β of each. This is shown in FIG. The output 112 is used as the phase timing calculation unit 1
At 3, convert to hourly data. In this case, in the timing generation method such as the pulse timing calculation unit 10, for example, a delay of the sampling period T _s occurs after the determination of α and before the output from the timer output unit 11, so that the synchronous modulation is performed. Cannot be realized. Therefore, if the phase timing calculation unit 13 monitors the fundamental wave phase and determines that α or β is reached within the next sampling period, the amount of time from the past increase rate of the fundamental wave phase to α is determined. Calculate the data. That is, the fundamental wave phase at the current sampling is θ, the increment of the fundamental phase in the sampling period from the previous time is Δθ12, and the rising angle of the pulse S _p in the cos ⁻¹ E * computing unit is α _p and the falling angle. _Is β _p , the rising timing T _pup and the falling timing T _pdn of the switching function S _p are as follows.

[0065]

[Equation 16]

On the other hand, also in the case of the switching function S _n , similarly, when the rising angle is α _n and the falling angle is β _n , it can be expressed by the following equation.

[0067]

[Equation 17]

The above output 113 is sent to the timer output section 11 via the changeover switch 15. The pulse mode switching unit 14 determines whether to switch between (a) dipolar modulation to (d) overmodulation and one-pulse control according to the magnitude of the output voltage command E *.
5 to switch the control mode. In addition, the switching signal 114 is sent to the phase correction unit 16, and the correction amount and direction of the fundamental wave phase are calculated when the pulse mode is switched.
6 is input to the phase adjusting unit 17 to correct the fundamental wave phase 101. The part 100 surrounded by the dotted line is a part common to the three phases, and the other parts are required for each phase.

Next, the operation of the phase corrector 16 will be described with reference to FIG. FIG. 11 shows a case of switching from overmodulation to one pulse at a phase of 180 °.
In the figure, the diagram of the basic modulation wave command of (a) is shown by overlapping with the modulation wave commands a _p and a _n on the positive and negative sides. In the one-pulse control, the startup timing α _n is calculated and the negative side modulated wave command a
_In contrast to _n , the negative side pulse can be generated without a time delay, whereas in the multi-pulse control overmodulation, there is a time delay in creating the positive side pulse as shown in (c), so it is actually output. The fundamental wave is as shown by the dotted line a _p ′. Therefore, at the time of switching, the fundamental wave shifts from a _p ′ to a _n , so that the fundamental wave becomes discontinuous and an overcurrent occurs.

Therefore, in the phase correction section 16, in order to make the phase command of the fundamental wave continuous, in the case of FIG. 10, the negative side modulated wave command a _n is delayed and the transition from a _p ′ to a _n is made continuous.
When switching from one pulse to overmodulation, there is a time delay in pulse creation on the overmodulation side, so the phase command of the fundamental wave on the overmodulation side is advanced. Table 2 shows the relationship of the correction directions arranged in order of phases.

[0071]

[Table 2]

The correction amount will be described. In the case of sample value control in which control is performed for each sampling time as in the present embodiment, the frequency response is 1/2 of the sampling period T _s .
Becomes Therefore, the phase output corresponding to ½ × T _s may be corrected for the phase output of the basic modulated wave calculated for each sampling period.

The phase of the basic modulated wave is obtained by integrating the increment Δθ for each sampling period T _s of the inverter frequency command Fi *. If the current phase is θ _n and the previous sampling phase is θ _n-1 ,

[0074]

[Equation 18] θ _n = θ _n-1 + Δθ (Equation 18) Δθ = k × T _s × Fi * Here, since k is a conversion coefficient, 1 / 2Δ when switching the pulse mode
The correction amount of θ is calculated and correction is performed.

[0075]

According to the present invention, the output voltage can be continuously controlled by correcting the phase delay of the fundamental modulation waves of the asynchronous modulation and the synchronous modulation, so that the output current does not fluctuate due to the switching of the modulation mode. , Smooth control can be performed.

[Brief description of drawings]

FIG. 1 is a main circuit diagram of a three-level inverter device for driving a vehicle to which the present invention is applied.

FIG. 2 is a control block diagram showing an embodiment of the present invention.

FIG. 3 is a PWM control characteristic diagram of a three-level inverter.

FIG. 4 is a pulse waveform diagram of a basic modulation wave command for dipolar modulation and an output voltage.

FIG. 5 is a pulse waveform diagram of a basic modulation wave command of partial dipolar modulation and an output voltage.

FIG. 6 is a pulse waveform diagram of a basic modulation wave command of unipolar modulation and an output voltage.

FIG. 7 is a pulse waveform diagram of an overmodulation basic modulation wave command and an output voltage.

FIG. 8 is an explanatory diagram showing a relationship between an inverter output voltage command and a modulation rate.

FIG. 9 is an explanatory diagram showing a pulse waveform of an output voltage of one pulse control.

FIG. 10 is an explanatory diagram showing a pulse timing generation direction during multi-pulse control.

FIG. 11 is an explanatory diagram of a shift from overmodulation to one-pulse control.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Phase calculation part, 3 ... Modulation rate calculation part, 5 ... Bias setting part, 10 ... Pulse timing calculation part, 11 ... Timer output part, 12 ... Cos ^-1 E * calculation part, 13 ... Phase timing calculation part, 14 ... Pulse mode switching unit, 16 ... Phase correction unit, 1
7 ... Phase adjusting unit, 18 ... Sampling setting unit.

Claims (3)

[Claims] 1. A DC power supply voltage is divided to generate three voltage levels of a high voltage, an intermediate voltage and a low voltage, and the three voltage levels are turned on / off by a switching element of a main circuit to output the above three voltage levels to an output voltage of an inverter. In a three-level inverter that selectively derives in the one-pulse control that outputs the maximum voltage, in the multi-pulse control other than the one-pulse control and the means that outputs the pulse width modulation signal synchronized with the modulation wave signal of the inverter. And a means for outputting a pulse width modulation signal asynchronous with the modulation wave signal of the inverter, wherein the pulse width modulation signal synchronized with the modulation wave signal of the inverter in the one-pulse control is the instantaneous voltage to be output as the inverter device. An on or off phase angle of the pulse width modulation signal is calculated from a command value, and the phase angle and the inverter When the phase angle of this modulated wave is compared, and when the phase angle of ON or OFF is reached within the next sampling period, the difference between the phase angle of the basic modulated wave at the current sampling and the phase angle of ON or OFF is calculated. It is calculated by converting this phase difference into the amount of time for the sampling period, and turning it on or off after the amount of time has elapsed within the next sampling period, and is asynchronous with the modulation wave signal of the inverter in the multi-pulse control. The pulse width modulation signal is generated by converting the command value of the instantaneous voltage to be output as the inverter device into the amount of time for each of the constant sampling periods, and turning it on or off in the next sampling period after the amount of time has elapsed. A controller for a pulse width modulation type inverter, characterized by: 2. The method according to claim 1, wherein when switching between the one-pulse control and the multi-pulse control other than the one-pulse control,
A pulse width modulation type inverter control device for correcting the phase angle of the fundamental modulated wave of the inverter. 3. The correction amount of the phase angle of the basic modulated wave of the inverter according to claim 2, wherein the phase angle corresponding to a half cycle of the sampling cycle for calculating the phase angle of the basic modulated wave is set, When switching from pulse control to multi-pulse control, a correction amount is added to the phase angle of the basic modulated wave to advance the phase, and when switching from multi-pulse control to single-pulse control, the phase angle of the basic modulated wave. A controller of a pulse width modulation type inverter that delays the phase by subtracting the correction amount from.