JP3412018B2 - Power converter - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention
Regarding the improvement of a power conversion device that converts a current into a direct current,
The present invention relates to control of an output voltage of a power converter. 2. Description of the Related Art A three-level inverter uses a DC power supply voltage.
(Overhead wire voltage) with two capacitors connected in series
By dividing the voltage into high and low potentials,
The main circuit switching element
Of these three levels by turning on and off the
It is selectively derived to the output terminal of the inverter.
It has such features. That is, the output voltage pulse
The apparent number of steps can be increased by increasing the number of steps.
The frequency is increased and an output with less distortion can be obtained. element
Since the voltage applied to is approximately halved compared to the two levels,
A switching element having a relatively low withstand voltage can be used. Element applied voltage
As the pressure decreases, the loss generated around the element can be reduced, etc.
It is. By the way, the output voltage of the above three-level inverter
There are the following methods for controlling the generation of pulses. (1) New Developments of 3 Level
Pied Bruem Strategies "New Developmen
ts of 3-Level PWM Strategies "(EPE'89 Recor
d, 1989), page 412, and FIG. 1 shows dipolar modulation.
(Pulse within the half cycle of the output voltage
The output voltage is expressed by alternating output)
Modulation method, unipolar modulation (single in half cycle of output voltage
The output voltage is expressed by outputting a pulse of polarity) and
Modulation system called and the above-mentioned dipolar modulation and unipolar
A modulation method in which modulation is mixed in one cycle (hereinafter, this specification
Is called partial dipolar modulation).
You. (2) Peda Bruem System in Power Co
Inverters: An Extension Of The
Subharmonic method “PWM Systems inPower Co
nverters: An Extension of the “Subharmonic” metho
d "(IEEETrasaction on Industrial Electoronics and
Control Instrumentation, Vol.IECI-28, No.
4, November 1981), page 316, and FIG.
The half cycle of the output voltage consists of multiple unipolar pulses.
To fill the slit between the pulses from this central part
The output voltage by reducing the number of pulses
Modulation system (hereinafter referred to as overmodulation) is proposed
Have been. (3) Study of 2 and 3 level pre
Calculated Modulations "Study of
2 and 3-Level Precalculated Modulations ”(EP
E'91 Record, 1991), page 411, FIG.
Is the output to cover the output voltage from 0 to 100%.
A control method of force voltage pulse generation has been proposed. [0003] By the way, the railway car
When using a three-level inverter for such applications,
Voltage control from zero voltage to achieve speed control over
Maximum voltage at which the usage rate reaches 100% (within a half cycle of the output voltage)
Is a voltage region where only a single pulse exists.
Up to one pulse)
Continuous and smooth harmonics of inverter output voltage
Is required to be controlled. By the way, the conventional technology
Surgery (1) uses a dipole that can control minute voltages including zero.
Modulation, unipolar modulation covering the medium speed region (medium voltage)
Control means, switch up to 1 pulse covering the maximum voltage
Output the maximum voltage from zero voltage.
Can maintain the continuity of the fundamental wave,
When switching to a pulse, the output voltage harmonics become discontinuous.
When noise is generated due to sudden and large changes in frequency
There was a problem. Also, as shown in the prior art (2),
Technology can express the maximum voltage from zero voltage.
There was a problem that can not be. Incidentally, the above-mentioned prior art (1) discloses an output power
In order to continuously control the fundamental wave of pressure,
The pulse data corresponding to the pressure is stored in the memory,
Output pulse trains corresponding to each modulation based on data
Therefore, the control is complicated. Further, the above prior art
(3) is a half-period of the fundamental wave in unipolar modulation.
Since the modulation method switches the number of existing pulses,
There is a problem that it complicates control. In addition,
The technology is uncomfortable when switching modulation schemes and pulse numbers.
There is a problem that a continuous sound is generated. It is an object of the present invention to provide a three-level inverter output.
Inverter output can be controlled from zero to maximum
3-level pulse generation for continuous and smooth voltage
Control, realizing an electric vehicle equipped with a three-level inverter
To prevent discontinuous sound. [0006] In order to achieve the above object,
In addition, in the half cycle of the fundamental wave of the output phase voltage of the power converter,
Power converter output from multi-pulse mode with pulses
A single pulse of the same polarity is provided in the half cycle of the fundamental wave of the phase voltage.
And the maximum of the fundamental wave of the output phase voltage of the power converter is 1
Output phase voltage of power converter
A single pulse of the same polarity in the half cycle of the fundamental
The pulse width of the single pulse is defined as the output phase voltage of the power converter.
Pulse controlled to spread as the fundamental wave grows
The width control mode is interposed. [0007] The three-level power converter (inverter)
Output voltage such as inverter frequency command and output voltage command.
The fundamental wave of the output phase voltage of the inverter according to the voltage-related information
From the multi-pulse mode having a plurality of pulses in a half cycle of
One pulse at which the fundamental wave of the output phase voltage of the inverter is maximized
When switching to the mode, the output phase voltage
In the half cycle of the main wave, there is a single pulse of the same polarity,
The pulse width of one pulse is based on the output phase voltage of the power converter.
Pulse width control to control the wave to spread as it grows
Over-modulation and one-pulse control
Control at a predetermined timing,
A continuous transition of the voltage is realized, and the output voltage changes smoothly.
Can be done. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the outline of the present invention is shown in Table 1 and FIG.
After description with reference to FIGS. 1 to 3, one embodiment is shown in FIGS.
This will be described with reference to FIGS. 3-level inverter
(Also called NPC inverter) is a DC power supply voltage (electrical
In the case of a car, the overhead line voltage) is
Divided into two DC voltages, high potential, intermediate potential
And three voltage levels of low potential and main circuit switch
These three levels of power are turned on and off by the switching element.
Voltage is selectively derived to the inverter output terminal.
You. As an example of this main circuit configuration, applied to railway electric vehicles
FIG. 1 shows a basic configuration (in the case of three phases) in the case where the above is performed. FIG.
, 4 is a DC overhead line (train line) which is a DC voltage source,
50 is a direct current reactor, 51 and 52 are direct current
Creates intermediate potential point O (hereinafter called neutral point) from voltage
It is a clamp capacitor divided and arranged for the sake of convenience. 7a,
7b and 7c are composed of self-extinguishing switching elements.
In response to the gate signal given to this switching element.
High point voltage (P point voltage), neutral point voltage (O point voltage)
Voltage) and low potential point voltage (N point voltage)
It is a switching unit. In this example, the switch
The switching unit 7a has 70 to 73 self-extinguishing switches.
A switching element (here, IGBT, but GTO,
Rectifier for reflux of 74 to 77
, And 78 and 79 auxiliary rectifying elements. Also,
The load is the case of the induction motor 6. Switching Uni
The units 7b and 7c have the same configuration as the unit 7a. Here, first, the U-phase switching unit is used.
The basic operation will be described with reference to Table 1 by taking the example 7a as an example.
I will tell. [Table 1] In the following, the clamp capacitors 51 and 52
Voltages vcp and vcn are DC which is completely smooth and divided to Ed / 2
Voltage, and the neutral point (point 0) is virtually grounded.
And Unless otherwise specified, the output voltage is
It refers to the inverter output phase voltage. Switching Uni
The switching elements 70 to 73 constituting the unit 7a are:
As shown in Table 1, ON / OFF is performed according to the three conduction patterns.
Work. That is, an output that outputs the P-point potential on the DC side
In the force mode P, 70 and 71 are on, and 72 and 73 are off.
Then, the output voltage becomes Ed / 2, and a neutral point potential is output.
In output mode O, 71 and 72 are on and 70 and 73 are off.
The zero potential is output as the output voltage, and the potential at the N point is
In the output mode N for output, 70, 71 are off, 72,
73 is on, and the output voltage is -Ed / 2. In Table 1
One phase of the main circuit in each output mode (switching unit
The equivalent circuit of the unit and the clamp capacitor is shown. Sui
The switching unit is equivalent to a three-way switch
Can be considered. Here, the conduction state of the element is represented by two values of 1, 0.
When the switching functions Sp and Sn are used, it can be expressed as Sp = 1, Sn = 0 in the output mode P, Sp = 0, Sn = 0 in the output mode O, Sp = 0, Sn = 1 in the output mode N. At this time, the switching functions Sp, Sn
To the switching elements 70, 71, 72, 73
Gate signals Gpu, Gpx, Gnx, Gnu (off signal 0, off
The relationship of (1) is expressed by the following equation. (Equation 1) Therefore, two switching functions Sp and Sn are provided for each phase.
The conduction state of the switching element is determined by preparing
Can be specified. This switching function Sp, Sn
Is output voltage eu by pulse width modulation (PWM) control.
Is determined to be sinusoidal. In addition, three levels
For details of the main circuit of the inverter, see JP-A-51-47848,
It is described in JP-A-56-74088. By the way, a limited power supply like an electric car is used.
Variable from the variable voltage variable frequency (VVVF) range
Performs a wide range of speed control over the frequency (CVVF) domain
In this case, an output voltage characteristic as shown by a solid line in FIG. 2 is required.
You. That is, in the low speed region, the inverter frequency
Adjust the output voltage in proportion to this area (the VVVF control area
To make the magnetic flux in the motor almost constant.
To maintain the specified torque, and at high speeds
Inverter while maintaining the maximum output voltage of the inverter
Data frequency (this area is called the CVVF control area).
The maximum voltage utilization with a limited voltage.
As a rule, high speed operation is realized. However
According to the conventional unipolar modulation method,
Low inverter frequency requires control of minute output voltage
(In the vicinity of the starting point of the VVVF control area)
Minimum output pulse determined by the minimum on-time of the switching element
Voltage pulse smaller than the pulse width cannot be realized,
As shown by the broken line in FIG.
Would be. For example, the inverter output voltage
All pressure pulses are at the minimum ON time Ton of the switching element
Considering the case where the minimum pulse width is determined more,
Output voltage effective value E is Here, Fc is given by the carrier frequency, and a voltage smaller than this cannot be controlled. This
Here, Emax is the effective value of the 180 ° conduction square wave voltage.
, [Equation 3] And the maximum output voltage of the three-level inverter is almost
It matches this Emax. According to the above (Equation 2), Fc =
When 500 kHz and Ton = 100 μs, E = 0.1 Em
ax, in this case, 10% or less of the maximum output voltage Emax
Cannot be controlled. Therefore, Unipo
Modulation alone limits the lower limit of the controllable output voltage.
This makes continuous voltage control difficult.
Was. To solve this, dipolar modulation is used.
(Dipolar mode) is effective, but in the prior art,
From this dipolar modulation, unipolar modulation (unipolar
Caution was required when moving to Meanwhile, Uni
The maximum voltage E that can be output by polar modulation is an ideal sine wave
At the limit point of modulation (modulation rate A = 1), And considers the minimum off time Toff of the switching element
If so, Here, Fc is the carrier frequency. For example, Fc = 500 Hz, Toff = 200 μ
In the case of s, E = 0.707Emax, and in this case,
It can cover only about 70% of the maximum output voltage Emax
Will be. At this time, adjust the pulse width of 1 pulse mode
If not, the fundamental wave will be discontinuous and
If the pulse width of the pulse mode is adjustable, the pulse width
In order to maintain continuity by reducing
Wave continuity is lost. Cover this voltage range
There are various modulation methods that can be used, but pulse generation control is easy.
The consistency with the unipolar modulation and the high
Overmodulation (overmodulation mode) is the best in terms of harmonic continuity, etc.
Can also be said to be effective. In the overmodulation region, the output voltage is half
Central part of periodic voltage pulse train (peak of fundamental wave instantaneous value)
Gradually narrow the narrow slit between pulses
By expanding the output voltage to around one pulse
Is possible. The limit of overmodulation control, that is, the modulation rate is
In the largest region, one pulse is output for one half cycle of the output voltage.
Shift to the so-called one-pulse mode
In this case, the output voltage almost reaches Emax. However
If this is the case, one pulse or one pulse from overmodulation
The transition timing from source to overmodulation depends on the modulation rate and carrier
This timing can be set arbitrarily because it depends on the frequency.
If hysteresis is provided during this period, the fundamental voltage
Continuity is lost. So, from overmodulation control,
Pulse width control that is not an extension of
Voltage control by 1-pulse mode)
Shift to possible one-pulse control. This allows overmodulation
Transition at predetermined timing is possible between and pulse control
And a continuous transition of the fundamental voltage is realized. [0013] Continuously performing these series of transition control
Pulse mode corresponding to the required output voltage
While selecting, only continuously from zero voltage to maximum voltage
Also obtains a highly accurate and stable output voltage. That is, in FIG.
As shown, V / F = constant as shown in FIG.
Control, use dipolar modulation from startup to F1
When the inverter frequency reaches F1, unipolar
Shift to the modulation area, overmodulation area at F2, and 1 at F3
Shift to the pulse area sequentially. Based on the unified voltage command,
FIG. 3 shows an example of a modulated wave that can be realized by the above. Output power
The fundamental modulation wave a proportional to the fundamental wave component of the pressure
Inverter frequency command Fi * from control means and output voltage
It is created from the following equation based on the command E *. (Equation 6) Where A: modulation rate, t = time, θ: phase (= 2πFi)
* t) Here, the modulation factor A (0 ≦ A ≦ 1) in the sine wave modulation region.
Is given by the following equation. (Equation 7) This basic modulated wave a is a dipolar modulation, a unipolar modulation
Are completely the same, and in overmodulation,
The same is true except that the calculation method of the adjustment factor A is different. A link between dipolar modulation and unipolar modulation
In order to enable continuous migration, here
The positive and negative bias modulation waves abp and abn are provided. (Equation 8) In the dipolar modulation control, the above abp and abn are directly correct.
A side modulated wave ap and a negative side modulated wave an are obtained. (Equation 9) Here, the creation of the switching functions Sp and Sn is simplified.
For convenience, set both ap and an to be positive.
I have. Finally, the pulse width of the output voltage is
It is set in proportion to the size, and in the case of dipolar modulation
Control the positive and negative pulses by shifting them by approximately 180 °. In the unipolar modulation, positive and negative modulated waves ap, a
n is: (Equation 11) Given by Minimum off time of switching element is ignored
If it is as small as possible, the instantaneous values of ap and an
When it is above, the maximum pulse is output (overmodulation described later). This
Here, setting of bias B is extremely important in transition control
It can be seen that it is. Dipolar modulation region depending on the value of B
And (a) when A / 2 ≦ B <0.5, dipolar modulation (b) when B = 0, and unipolar modulation is performed. On the other hand, in overmodulation control, the modulation rate A is set to 1 or more.
Increase the slip between the pulses in the middle part of the half cycle of the output voltage.
Output voltage (zero voltage output period) to improve output voltage.
Let If the voltage command is further increased, overmodulation mode
To 1-pulse mode. For this behavior,
This will be described in the following embodiments. Thus, dipolar
Modulation, unipolar modulation and overmodulation with unified voltage command
Continuous transfer up to the maximum output 1 pulse
Line control becomes possible. Hereinafter, the structure of an embodiment for realizing the above concept will be described.
The configuration will be described. FIG. 1 shows the switching unit described above.
Control to output an AC voltage having three levels of potential
It is an example of a pulse width modulation device. In FIG. 1, 1 is an output
Dipolar modulation according to voltage related information and transition control information
Waveform or unipolar modulated waveform or overmodulated waveform
Multi-pulse generating means to output shape, 2 is output voltage related information
1 pulse waveform (1 pulse mode)
Loose generation means 3 continuously shifts each PWM mode
Transfer control means. The output of the transition control means 3
The start signal of each phase is passed through a gate amplifier (not shown).
Given to the switching element in the switching unit,
ON / OFF controlled. These multi-pulse generating means 1, 1
Consisting of pulse generation means 2 and transition control means 3
The pulse width modulation means is a feature of the present invention. In addition, this
In the example, the output voltage relation taken into the pulse width modulation means is
The link information is provided from the higher-level current control unit 8. this
The current control means 8 controls the current control means 81 from the current command.
To create the slip frequency command Fs * of the induction motor 6
Flow command value and the actual motor current)
Rotation frequency detecting means 61 attached to the machine 6
The detected rotation frequency Fr of the induction motor and the Fs *
To generate an inverter frequency command Fi *. Further
And the DC voltage Ed of the three-level inverter
(In the PN voltage, the sum of the clamp capacitor voltage vcp + v
cn), the output voltage setting means 82 outputs
Create voltage command E *. This output voltage setting means 82
Is that the slope is large when Ed is low (Ed = Ed1),
If Ed is high (Ed = Ed3), set a small slope
And always make the output voltage as required,
This realizes the output voltage characteristics shown. These currents
The control means outputs an instantaneous value of the output voltage.
Is also good. The configuration and operation of the pulse width modulation means will be described.
This will be described in detail with reference to FIGS. FIG.
2 shows an example of the overall configuration of a loose width modulation unit. Where multi-pulse
The generation means 1 includes a basic modulation wave generation means 11, a bias superposition
Means 12, positive / negative distribution means 13, reference signal generation means 14,
And pulse generating means 15. Basic modulation wave generation
The generation means 11 receives the input received as the output voltage related information.
The barter frequency command Fi * is calculated by the phase calculating means 112.
The phase θ is obtained by time integration, and the phase θ
Find the sine value sinθ. On the other hand, one of the output voltage related information
From the voltage command E * which is
The amplitude A (modulation rate) of the modulated wave was calculated and output, and was halved.
Instantaneous basic modulation with half amplitude multiplied by sinθ
A wave a / 2 is created and output. Bias superimposing means 12
Is the multi-pulse transition control procedure of the transition control means 3
Add and subtract the bias B from stage 31 to obtain two positive and negative
Generate and output bias modulation waves abp and abn. Here, dipolar modulation and unipolar modulation
The continuous transition between and depends on the setting of the bias B. FIG.
In addition, the dipole performed by setting the bias B
4 shows a configuration example of a la / unipolar transfer control unit 311. Da
The bipolar / unipolar transition control means 311 outputs the output voltage
The modulation rate A is obtained by multiplying the command E * by 4 / π by 311a.
Calculated by the bias generation means 311b.
The determined bias B is determined. That is, the modulation rate A is small.
Where a very small output voltage is required, B = Bo (
However, set Bo ≧ A / 2) and A = A1
And B = 0. The output voltage when A = A1 is given by the formula
A1 is increased so that it becomes higher than the voltage shown in (2).
If determined in advance, the voltage from the minute voltage including zero
Control becomes possible. Further, the positive and negative bias modulation waves a
bp and abn are converted by the positive / negative distribution means 13 into abp and abn.
The positive part is ap, and the negative part of abp and abn is a
n and distributes and synthesizes them to reduce
The continuity of the output voltage fundamental wave component over Nipolar modulation
The maintained positive / negative modulated waves ap, an are created. This positive and negative
Based on the modulated waves ap and an, the pulse generation means 15
Switching functions Sp and Sn with a pulse generation period of 2To
Generate The reference signal generating means 14
The pulse generation period To is determined according to the wave number command Fsw *.
Here, the relationship between Fsw * and To can be expressed by the following equation. (Equation 12) The pulse generation operation of the pulse generation means 15 is illustrated.
6 will be described. In FIG. 6, the pulse timing
The setting means 151 includes ap, an, aoff, To (an,
aoff will be described later) based on
Timing Tpup and falling timing Tn of Sn
dn is obtained from the following equation (Process 1). (Equation 13) [Equation 14] In the next cycle, Sp falling timing Tpdn and
The rising timing Tnup of Sn is obtained in the same manner as in processing 1.
(Process 2). (Equation 15) (Equation 16) By performing the above processing 1 and processing 2 alternately, the switch
Ching functions Sp and Sn are created. Where aon, a
off is the minimum ON time Ton and minimum
It is a value determined from the off time Toff. Given by That is, as shown in FIG. 7 (example of Sp)
The on-pulse width Twon and the off-pulse width Twoff are given by: And has the characteristic shown by the broken line in FIG. Where
The loose width Twon is the maximum determined by the switching element.
In order not to be less than the small on-time Ton,
The width Twoff is the maximum determined by the switching element.
The solid line in FIG.
The characteristics shown below. To achieve this, the pulse
The function of the imaging setting means 151 has been added. By this
Output voltage fundamental wave component discontinuity is extremely small
Therefore, they can be ignored. Note that aoff is a discontinuity of the output voltage fundamental wave component.
It is variable within the range where the connection can be neglected.
The timing of transition from polar modulation to overmodulation
Polar / overmodulation transition control means 312 provides this. Also
If aoff is set to be constant,
Can be simplified. That is, pulse timing setting means
151 automatically transitions from unipolar to overmodulation
Output control means 3 to output aoff
12 does not need to be provided. Switching function generating means 1
52 generates a reference signal having a period To, and in synchronization with the reference signal
Based on Tpup, Tndn or Tpdn, Tnup, Sp, S
Set n. FIG. 9 shows an example of a switching function during overmodulation.
Shown in Switch when instantaneous value Ap of ap exceeds aoff
Between the pulses of the switching function Sp (see the hatching in FIG. 9C).
Filling part). The width of this filled slit is
Smaller than the minimum off time Toff of the switching element, 1
The output voltage fundamental wave
Has little effect. The pulse timing setting means 151 is implemented by software
FIG. 10 shows a flowchart in the case of realizing by hardware.
You. By the way, in overmodulation control, the center of the output voltage half cycle is
Maximum voltage by filling the slit between partial pulses
Maintain the state, PWM control only near the zero cross of the modulated wave
It is carried out. Therefore, in this region, the modulation rate A and the actual
Is non-linear, the modulation rate A is linear.
Output voltage linearly follows this
Does not increase. Therefore, it is necessary to make the setting of the modulation rate A nonlinear.
Thus, the output voltage at the time of overmodulation is linearized. Sand
That is, the switching frequency in the PWM control section is sufficiently high.
In this case, the fundamental value R of the output voltage and the modulation factor A
Can be expressed by the following equation. [Equation 19] Therefore, the relationship between E * and A is calculated in advance from the relationship in the above equation.
The amplitude setting means 111 shown in FIG.
By adjusting the output voltage linearly with respect to E *
Wear. As a result, the voltage in a high voltage region, particularly close to one pulse,
Controllability can be improved. If the voltage command is further increased,
The operation of the changeover switch 32 of the control means 3 causes the overmodulation mode.
Mode to 1 pulse mode. Changeover switch 3
2 is one of the outputs of the multi-pulse shift control means 31
_PM Is S _PM When = 0, multiple pulse side S _PM When = 1, it is switched to one pulse side. FIG. 12 shows one-pulse / multi-pulse switching
One example of the control means 313 is shown. In this example, the voltage command E
When * exceeds E1P, one pulse mode from multi-pulse mode
Pal when E * becomes smaller than EMP
Hysteresis to shift from the pulse mode to the multi-pulse mode.
A lysis is provided. As a result, careless PWM mode
Output voltage and stable output voltage with less transient fluctuation
Is to be obtained. 1 pulse generating means 2
It comprises a phase calculation means 21 and a pulse generation means 22
You. The operation of the phase calculation means 21 is exactly the same as 111.
Alternatively, the output of 111 may be used by omitting 21. FIG. 13 shows an example of the structure of the pulse generating means 22.
You. Unlike the two-level PWM, the three-level PWM has one
The output voltage is adjusted by controlling the pulse width during pulse control.
I can. Then, from the voltage command E *, the rising of the pulse
Timing phase α and falling timing phase
β is given by Ask for. These α and β are set based on the phase θ,
By creating and outputting Sp and Sn, one pulse waveform
To achieve. As described above, the dipolar modulation, the unipolar
Modulation and overmodulation are realized based on a unified voltage command,
Continuous transition control up to the maximum output of one pulse is possible.
You. In this embodiment, the output voltage ranges from zero to the maximum voltage.
Can be adjusted continuously and smoothly.
In addition, it has the effect of providing a highly accurate and stable output voltage.
You. By the way, in the first embodiment shown in FIG.
Converts the output pulse train of the multi-pulse generation means into an inverter
It is generated asynchronously with the frequency, and the output
Luth is controlled in synchronization with the inverter frequency. This
The reason is that the synchronous method is adopted in the multi-pulse region
In the prior art described above, first, control for managing the phase is performed.
Is complicated, and secondly, the output voltage command is
When it is necessary to distort from a sine wave (in FIG. 1,
The inverter frequency Fi * and the output voltage command E * are used for controlling electric vehicles.
If the output voltage command is
There is a problem that it cannot be reproduced. That is, the first question
The problem is that the synchronous method uses pulses that are an integral multiple of the inverter frequency.
For output, the phase and generated pulse for each pulse mode
Table with the relationship of
Pulse generation phase from the phase obtained from the
And output it. It is necessary to manage this phase.
Calculation amount and memory for each pulse mode are huge.
Therefore, the control becomes complicated. Also, the second problem
In the synchronous type shown in the prior art, a pulse
Data, but the output voltage is a sine wave
The output voltage is exactly as instructed
There is a problem that cannot be expressed. Therefore, in this embodiment, the multi-pulse mode is set.
Pulse generation is asynchronous with the inverter frequency.
In order to solve these problems, That is, the first
For problems, the frequency of the inverter
Pulses can be generated independently of the number
Wear. That is, in FIG.
Fsw * is set independently of the inverter frequency command Fi *
(FIG. 4, reference generation 14 is the inverter frequency
Independent.) Because of this, complex
Control can be simplified without the need for simple control procedures.
You. Also, for the second problem, if it is asynchronous,
There is no need to have data for each phase, and instantaneous voltage commands
Can output a pulse equivalent to
So even a distorted sine wave can be faithfully represented
You. Further, as described above, control relating to phase calculation and the like is simplified.
A pulse corresponding to the voltage command is output successively
Calculation can be performed, and the calculation cycle can be shortened.
Can increase your fidelity.
Wear. In addition, the switching frequency becomes
The switching frequency is independent of the inverter frequency.
Changes in numbers can be minimized, seen synchronously
Changes in sound quality before and after pulse mode switching (abnormal noise,
There is also an effect that the pleasant sound can be minimized. In the above embodiment, a three-level inverter is used.
As explained for the example, two-level inverter and three-level inverter
The same applies to the above multi-level inverter. By the way, switching at a relatively low frequency
Switching devices such as GTO thyristors
Depends on the switching frequency among the output voltage harmonics.
Fundamental wave of sideband component and inverter frequency generated
Interference with components may occur. I avoided this
In the PWM mode of the multi-pulse generator,
Modulation mode and unipolar modulation mode with inverter frequency
And the overmodulation mode and 1-pulse mode
Synchronize (FIG. 14). With such a configuration,
More stable voltage can be supplied even during overmodulation.
Become. FIG. 15 shows another embodiment of the multi-pulse shift control.
Show. FIG. 15 shows only the multi-pulse transition control means 31.
Was. This means that four types of PWM modes are designated as inverter frequency
To shift depending on both the command Fi * and the voltage command E *
It is. That is, when Fi * <F1 and E * <E1,
Polar modulation, when Fi * ≧ F1 and E1 ≦ E * <E2
Unipolar modulation, overmodulation when E2 ≦ E * <E3, E * ≧
In the case of E3, one pulse is used. As a result, for example,
Output power in a high-frequency high-speed range, such as during operation or repowering.
Even when the pressure is soft-started,
Transition condition: key modulation → unipolar modulation → overmodulation → 1 pulse
Is satisfied, and a stable voltage rise is possible. Ma
Also, since dipolar modulation control is always performed in the low frequency region,
To a specific switching element such as unipolar modulation
Current concentration can be avoided. Next, a second embodiment will be described. First
As shown in FIG.
A part where both modulation waveforms coexist between modulation and unipolar modulation
By introducing minute dipolar modulation, the output voltage and
The smoothness of the switching frequency can be increased. FIG. 17 shows an example of the output voltage command waveform.
17 is exactly the same as FIG. 3 except for (b).
Hereinafter, the partial dipolar will be described. bias
Bias B is dipolar due to the effects of superposition and positive / negative distribution.
Range that is neither modulation nor unipolar modulation (0 <B <A /
Even if it is set to 2), the power as required by the basic modulated wave
It is possible to reproduce the pressure without excess or deficiency. in this case,
The unipolar modulation is applied near the peak of the output voltage, and the
It becomes partial dipolar modulation, which is polar modulation. This and
The positive-side modulated wave ap and the negative-side modulated wave an of (Equation 22) It becomes. (Ap-an) always matches the fundamental modulation wave a,
The continuity of the instantaneous value of the output voltage fundamental wave is also maintained.
Call Utilizing the above property, the buffer rate increases as the modulation rate A increases.
If the bias B is gradually reduced, the
Continuous through partial dipolar modulation up to nipolar modulation
Can be migrated. Of course, the reverse is also possible. One of the dipolar / unipolar transition control means
An example is shown in FIG. As shown by the solid line in FIG.
If B is set, then in the region of 0 ≦ A ≦ A1, dipolar
Modulation, partial dipolar modulation in the region of A1 <A <A2,
Unipolar modulation is performed in the region of A ≧ A2. in this case,
When switching between dipolar modulation and unipolar modulation, is it an electric motor?
Since these noises do not occur, it is effective for reducing the noise of the equipment.
You. Applying FIG. 18, as shown in FIG.
The PWM mode can be managed for each area. FIG.
Only the transfer control means 31 is shown. This is five kinds of PW
The M mode is controlled by the inverter frequency command Fi * and the voltage command E *.
The transition depends on both. That is, Fi
When * <Fo and E * <Eo, dipolar modulation, Fo ≦ F
Partial dipolar transformation when i * <F1 and Eo ≦ E * <E1
Key, Unipolar when Fi * ≧ F1 and E1 ≦ E * <E2
Modulation, overmodulation when E2 ≦ E * <E3, when E * ≧ E3
One pulse is assumed. This makes it possible, for example, to start
Output voltage in the high-speed and high-frequency range
Even when starting, dipolar modulation → partial
Polar modulation → Unipolar modulation → Overmodulation → One pulse
Transfer conditions are satisfied and stable voltage startup is possible.
Become. Also when idling and re-adhesion is the same as when regenerative startup
Effects. In addition, in any operating condition
Even when noise is generated from the motor when switching the pulse mode
Has the effect of being minimized. FIG. 20 shows inverter frequency and switching.
FIG. By the way, electric car system for railway vehicles
In the inverter used for the control device, the inverter frequency
The variable range of Fi * is about 0 to 300 Hz. Output power
The inverter frequency Fcv at which the pressure becomes maximum is
The upper limit of Fcv is about 10 at 1/5 to 1/3 of the wave number variable upper limit.
It is about 0 Hz. When generating pulses asynchronously,
Harmonics generated around the switching frequency and the inverter
Avoid fluctuations in output current due to interference with the fundamental frequency
Has a switching frequency about 10 times higher than Fcv,
A switching frequency of 1 kHz or more is required. Further
In order to reduce noise (abnormal noise, etc.)
It is effective to minimize fluctuations in numbers,
Switching frequency in the multi-pulse range
Can be controlled within 1-2 Fi. Of course
If a microprocessor or the like is used,
Program some or all of the
It is also possible to realize it as a software. FIG. 21 shows the pulse width modulation means of FIG.
Up to the calculation of pulse rise and fall timing
An example of a flowchart for implementing in software
Show. The above explanations are based on the case of induction motor load as an example.
However, the same applies to other AC motors
The effect can be expected. In addition, all of the above
It was explained that the output terminals of these inverters
Connected to an AC power supply via a reactance element,
To operate as a self-excited converter that converts
Both are possible. In this case as well, as with the inverter
The effect can be expected. The above description is for a three-level inverter.
As mentioned above, the concept of the present invention is more than three levels.
Bell inverters can also be used. According to the present invention, the inverter output voltage is
Continuous and smooth adjustment from zero voltage to maximum voltage
And simplifies the pulse generation control system.
Can be applied to electric vehicles, low noise electric vehicles
Can be provided.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a configuration diagram showing one embodiment of the present invention. FIG. 2 is a diagram illustrating a relationship between output voltage characteristics and a PWM mode. FIG. 3 is an explanatory diagram of a modulated wave for continuous transition to a PWM mode in a multi-pulse region. FIG. 4 is a detailed explanatory diagram of the configuration in FIG. 1; FIG. 5 is a diagram showing an example of dipolar / unipolar transition control means. FIG. 6 is a diagram showing an example of a pulse generating means in the multi-pulse generating means. FIG. 7 is a waveform diagram showing a relationship between on / off pulse widths. FIG. 8 is a view showing characteristics of an on / off pulse width. FIG. 9 is a diagram showing an example of an overmodulation waveform. FIG. 10 is a view showing a flowchart of pulse timing setting means by software. FIG. 11 is a diagram illustrating a configuration example of an amplitude setting unit. FIG. 12 is a diagram showing an example of a multi-pulse / 1-pulse switching control means. FIG. 13 is a diagram showing an example of one-pulse generation means. FIG. 14 is a diagram showing a configuration example of another embodiment. FIG. 15 is a diagram illustrating an example of a transition control unit. FIG. 16 is a relationship diagram between an output voltage characteristic and a PWM mode when another PWM mode is included. FIG. 17 is a diagram for explaining another PWM mode modulated wave. FIG. 18 is a configuration diagram of a transition control unit that realizes another PWM mode. FIG. 19 is a diagram illustrating an example of a transition control unit. FIG. 20 is a diagram illustrating a relationship between an inverter frequency and a switching frequency. FIG. 21 is a view showing a flowchart of pulse width modulation means by software. [Description of Signs] 1 ... Multi-pulse generating means, 2 ... 1 pulse generating means, 3 ... Transition control means, 4 ... DC overhead wire, 6 ... Induction motor, 7a, 7
b, 7c: switching unit, 8: current control means,
11: basic modulated wave generating means, 12: bias superimposing means,
13: positive / negative distribution means, 14: reference signal generation means, 15 ...
Pulse generating means, 21: phase calculating means, 22: pulse generating means, 31: multi-pulse shift control means, 32: changeover switch, 50: DC reactor, 51, 52: clamp capacitor, 61: rotational frequency detecting means

──────────────────────────────────────────────────続 き Continuing from the front page (72) Inventor Mutsumi Terunuma 4026 Kuji-cho, Hitachi City, Ibaraki Prefecture Hitachi, Ltd.Hitachi Laboratory (72) Inventor Yuto Suzuki 4026 Kuji-machi, Hitachi City, Ibaraki Prefecture Hitachi, Ltd. Hitachi, Ltd. In the laboratory (72) Inventor Yoshio Tsutsui 4026 Kuji-cho, Hitachi City, Ibaraki Prefecture Hitachi, Ltd.Hitachi Research Laboratory Co., Ltd. (72) Inventor Eiichi Toyoda 1070 Momo, Hitachinaka City, Ibaraki Prefecture Hitachi, Ltd.Mito Plant (56) References JP-A-3-159570 (JP, A) JP-A-4-112680 (JP, A) JP-A-62-296786 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name ) H02M 7/48 H02P 5/41

Claims (1)

(57) Claims 1. A power converter for converting a direct current to an alternating-phase voltage having a three-level potential by switching control of a plurality of switching elements, and controlling the switching elements of the converter. In the power converter including a control device, from a multi-pulse mode having a plurality of pulses in a half cycle of a fundamental wave of an output phase voltage of the power converter, to a half cycle of a fundamental wave of an output phase voltage of the power converter. It has a single pulse of the same polarity and shifts to the one-pulse mode in which the fundamental wave of the output phase voltage of the power converter is maximized. A pulse width control mode having a single pulse of the same polarity and controlling the pulse width of the single pulse so as to increase as the fundamental wave of the output phase voltage of the power converter becomes larger is provided. Power converter for.