GB2627238A

GB2627238A - Motor control

Info

Publication number: GB2627238A
Application number: GB2302211.4A
Authority: GB
Inventors: Tatlow Mark; Dehghan-Azad Ehsan; Xu Jin
Original assignee: McLaren Applied Ltd
Current assignee: McLaren Applied Ltd
Priority date: 2023-02-16
Filing date: 2023-02-16
Publication date: 2024-08-21
Also published as: WO2024170869A1; GB202302211D0

Abstract

A control circuit for generating a duty cycle for a plurality of pulse-width modulated signals comprises an input to receive an indication of a voltage required of the pulse-width modulated signals S701. A transform unit calculates, in dependence on the indication, a respective duty cycle for each of the plurality of pulse-width modulated signals S702. A gain unit selects a gain that will result in a linear relationship between the voltage required of the pulse-width modulated signals and a voltage that will be achieved by the plurality of pulse-width modulated signals, even if the voltage required of the pulse-width modulated signals is comprised in an overmodulation zone. The gain is applied to respective duty cycles to generate adjusted duty cycles S703. An output outputs the adjusted duty cycles for generating the plurality of pulse-width modulated signals S704. The control circuit may comprise an adjustment unit to add a zero-sequence component to the duty cycles calculated by the transform unit. A flag unit may determine whether the voltage required of the pulse-width modulated signals is comprised in a zero-sequence injection zone. The gain unit may select the gain by accessing a look-up table, where the look-up table may be precalculated.

Description

MOTOR CONTROL

The invention relates to control apparatus and methods.

A schematic diagram of a motor control system is shown in Figure 1. The system comprises a pulse-width modulator 101, an inverter 102 and a motor 103. The voltage demands calculated by the controllers (not shown) are modulated into pulses by the pulse-width modulator. This waveform drives gate drivers in the inverter.

Pulse-width modulated (PWM) waveforms are particularly suited to running inertial loads like motors. Their inertia causes them to react slowly to any changes in an input signal, so they are not readily affected by the rapidly-switched discrete pulses of a PWM signal. A PWM signal is digital, but when input into an appropriate load it can effectively appear analogue. Two examples of pulse-width modulated (PWM) waveforms are shown in Figure 2 (201, 202). They both comprise a series of rectangular, on-off pulses. The pulses are generated at a switching frequency. The average voltage of the waveform is dependent on the fraction of time that the pulse is on versus off. The duty cycle of the PWM waveform is the time that the waveform is "high" divided by the switching period. Waveforms 201 and 202 have the same switching frequency but different duty cycles, with the result that waveform 202 has a higher average voltage than waveform 201.

One straightforward way to generate a pulse-width modulated wave is to use a comparator configured to receive a sine wave and a sawtooth wave as inputs (or any other suitable comparison circuit), and to output either a maximum circuit voltage or a minimum circuit voltage depending on which of its two inputs has the higher value at a particular clocking instant. An example of this is shown in Figure 3. The respective amplitudes of the carrier triangle wave 301 and modulated sine wave 302 are compared. PWM signal 303 is set "low" when the amplitude of the sine wave is less than that of the triangle wave, and "high" when the amplitude of the sine wave is higher than that of the triangle wave. When the sinewave reaches the peak of the triangle, the PWM pulses obtain their maximum width. This is shown in Figure 4, where modulated sine wave 401 exceeds the peak of carrier triangle wave 402. The modulation has entered a state of saturation.

Three-phase motors require three PWM waveforms -one for each coil -that are 120° out of phase with each other. The three waveforms can be generated using the same technique shown in Figure 2. The maximum line-to-line voltage you can get from the PM waveforms in this scenario, without creating distorted waveforms, is./3Vdc, or 86.6% of the DC link voltage. The DC link voltage is under-utilised. The motor coils, however, see a potential difference rather than a point voltage. The average voltage of each waveform can be increased, thereby increasing utilisation of the DC link, while keeping the effective voltage between any two waveforms unchanged, by appropriately adding/subtracting the three PWM waveforms.

One mechanism for doing this is Space Vector Pulse Width Modulation (SVPWM). The principle behind this technique is to analyse the state of the motor by considering per-phase voltages together as a group. The PWM waveforms are switching waveforms intended for six transistors that deliver voltage to the motor coils. Each coil is connected to two transistors, and each grouping of two transistors can be in one of two states (top transistor closed and bottom transistor open, or vice-versa). One state can be notionally assigned a "1" and the other a "0". In essence the switching circuit can be represented by three binary bits, which have eight possible output states.

Instead of viewing the problem as one of generating three independent switching waveforms, the three-phase motor inverter can be viewed as one unit that can generate eight different switching states: a three-dimensional Cartesian product of the two states on phase A, the two states on phase B, and the two states on phase C. A two-dimensional projection of this arrangement is illustrated in Figure 5. Two of the states, (A,B,C) = (0,0,0) and (1,1,1), represent zero instantaneous line-to-line voltages, and are often called the "zero" or "null" vectors (502). The remaining six states represent nonzero vector voltages applied across the motor terminals (503). The states are referred to as base vectors. The goal of SVPWM is to produce a "mean vector" during the PWM period (TRANI) that is equal to the desired voltage vector 501 (Vout)-SVPWM can be successfully used in the saturated part of the modulation zone (also known as the overmodulation zone). It can increase utilisation of the DC link voltage. However, it is an independent calculation from that which would normally be applied in the non-saturated part of the modulation zone (also known as the linear zone) meaning that the overall PWM method would require two parts. SVPWM can also require angle calculation to compute the "mean vector" as a weighted average of the base vectors. This complicates the duty cycle calculation logic in the motor controller.

Therefore, there is a need for an improved mechanism for controlling an electric motor.

According to a first aspect of the invention, there is provided a control circuit for generating a duty cycle for a plurality of pulse-width modulated signals, the circuit comprising an input configured to receive an indication of a voltage required of the pulse-width modulated signals, a transform unit configured to calculate, in dependence on the indication, a respective duty cycle for each of the plurality of pulse-width modulated signals, a gain unit configured to select a gain that will result in a linear relationship between the voltage required of the pulse-width modulated signals and a voltage that will be achieved by the plurality of pulse-width modulated signals, even if the voltage required of the pulse-width modulated signals is comprised in an overmodulation zone, and apply that gain to the respective duty cycles to generate adjusted duty cycle and an output configured to output the adjusted duty cycles for generating the plurality of pulse-width modulated signals.

The gain unit may be configured to determine if the voltage required of the pulse-width modulated signals is comprised in a linear modulation zone and if so, apply a gain of one to the respective duty cycles to generate the adjusted duty cycles.

The gain unit may be configured to determine if the voltage required of the pulse-width modulated signals comprised in the overmodulation zone and if so, apply a gain that is not one to the respective duty cycles to generate the adjusted duty cycles.

The control circuit may comprise an adjustment unit configured to add a zero-sequence component to the duty cycles calculated by the transform unit.

The control circuit may comprise a flag unit configured to determine whether the voltage required of the pulse-width modulated signals is comprised in a zero-sequence injection zone.

The flag unit may be configured to determine, if the voltage required of the pulse-width modulated signals is comprised in a linear zone, that the voltage required of the pulse-width modulated signals is also comprised in the zero-sequence injection zone.

The flag unit may be configured to determine, if the required voltage is comprised in the overmodulation zone: that the voltage required of the pulse-width modulated signals is comprised in the zero-sequence injection zone if that voltage is below a predetermined threshold; and that the voltage required of the pulse-width modulated signals is comprised in the zero-sequence injection zone if that voltage is above the predetermined threshold.

The flag unit may be configured to: if it determines that the voltage required of the pulse-width modulated signals is comprised in the zero-sequence injection zone, cause the zero-sequence component to be added to the duty cycles calculated by the transform unit; and if it determines that the voltage required of the pulse-width modulated signals is not comprised in the zero-sequence injection zone, cause the zero-sequence component not to be added to the duty cycles calculated by the transform unit.

The adjustment unit may be configured to add a zero-sequence component by increasing each duty cycle calculated by the transform unit by an identical amount.

The control circuit may be configured to receive the indication of the required voltage and to output the adjusted duty cycles irrespective of whether the voltage required of the pulse-width modulated signals is comprised in the overmodulation zone or in a linear modulation zone.

The indication of the voltage required of the pulse-width modulated signals may comprise a number of elements that is fewer than the number of the plurality of pulse-width modulated signals.

The indication of the voltage required of the pulse-width modulated signals may represent a coordinate in a two-dimensional system.

The indication of the voltage required of the pulse-width modulated signals may be an output from an inverse Park transform.

The transform unit may be configured to convert the indication of the voltage required of the pulse-width modulated signals into the respective duty cycles using the inverse Clarke transform.

The duty cycles calculated by the transform unit may represent coordinates in a three-dimensional system.

The number of the plurality of pulse-width modulated signals may be three.

The gain unit may be configured to select the gain by accessing a look-up table.

The look-up table may be precalculated.

According to a first aspect of the invention, there is provided a method for generating a duty cycle for a plurality of pulse-width modulated signals, the method comprising receiving an indication of a voltage required of the pulse-width modulated signals, calculating, in dependence on the indication, a respective duty cycle for each of the plurality of pulse-width modulated signals, selecting a gain that will result in a linear relationship between the voltage required of the pulse-width modulated signals and a voltage that will be achieved by the plurality of pulse-width modulated signals, even if the voltage required of the pulse-width modulated signals is comprised in an overmodulation zone, and apply that gain to the respective duty cycles to generate adjusted duty cycles and outputting the adjusted duty cycles for generating the plurality of pulse-width modulated signals.

The present invention will now be described by way of example with reference to the drawings. In the drawings: Figure 1 shows an example of a motor control system; Figure 2 shows examples of low-average voltage and high-average voltage pulse-width modulated signals; Figure 3 shows an example of a pulse-width modulated signal being generated; Figure 4 shows an example of saturation in the generation of a pulse-width modulated signal; Figure 5 shows the eight different switching states of a Space Vector Pulse Width Modulation representation of a three-phase motor inverter; Figure 6 shows an example of a control circuit; Figure 7 shows an example of a control method; Figure 8 shows an overview of an electric motor arrangement; Figure 9 shows an example of a motor controller; Figure 10 shows an example of a three-dimensional coordinate system; Figure 11 shows an example of a two-dimensional coordinate system; Figure 12 shows an example of a method for controlling the PWM signals for a three-phase electric motor; and Figure 13 shows illustrates a linear relationship between a required and achieved modulation index.

An example of a control circuit is shown in Figure 6. The circuit comprises an input 601, a transform unit 602, a gain unit 603 and an output 604. The circuit is capable of generating a duty cycle for a plurality of PWM signals. The input is configured to receive an indication of a voltage that is required of the pulse-width modulated signals. The indication may be an average voltage to be met by a combination of the PWM signals.

The transform unit 602 may be configured to calculate, in dependence on the indication, a respective duty cycle for each of the plurality of pulse-width modulated signals. The gain unit 603 is preferably configured to select a gain that will result in a linear relationship between the required voltage and a voltage that will be achieved by the PWM signals. The gain is selected so that this linear relationship is achieved even when the required voltage is comprised in an overmodulation zone. The term "overmodulation zone" may refer to required voltages that require the PWM generation process to enter a state of saturation. For example, the required voltage may be greater than that which is achievable solely by increasing the width of the PWM pulses. The gain unit may apply the selected gain to the duty cycles output by the transform unit to generate adjusted duty cycles. These adjusted duty cycles are then output by output 604, suitably to some form of signal generator that is capable of generating the PWM signals.

The structures shown in Figure 6 (and indeed all block apparatus diagrams included herein) are intended to correspond to a number of functional blocks in an apparatus. This is for illustrative purposes only. Figures 6 is not intended to define a strict division between different pads of hardware on a chip or between different programs, procedures or functions in software. In some embodiments, some or all of the procedures described herein may be performed wholly or partly in hardware. In some implementations, the transform unit 602 and the gain unit 603 (for example) may be implemented by a processor acting under software control. Any such software is preferably stored on a non-transient computer readable medium, such as a memory (RAM, cache, FLASH, ROM, hard disk etc.) or other storage means (USB stick, FLASH, ROM, CD, disk etc).

An overview of an appropriate method for calculating the duty cycles is shown in Figure 7. In step S701 an indication of a voltage that is required of the pulse-width modulated signals is received. This is converted into respective duty cycles (step S702). An appropriate gain is selected and applied to the duty cycles to generated adjusted duty cycles to linearise the relationship between the required voltage and a voltage that will be achieved by the PWM signals (step S703). The final step is to output the adjusted duty cycles (step S704).

Each duty cycle may be considered to represent a voltage that is required of the PWM signals. They will therefore also be termed "voltage requirements" in the description that follows, to aid understanding of the conversion from one quantity to another. For example, in a further example, the control circuit may additionally comprise an adjustment unit. The transform unit may be configured to convert the received indication into respective duty cycles (or a "first voltage requirement"). The first voltage requirement may comprise a plurality of elements -one for each of the plurality of pulse width modulated signals. Each element could, for example, be a scalar or vector quantity. The adjustment unit may be configured to add a zero-sequence component to the first voltage requirement to generate a second voltage requirement. The second voltage requirement is then converted into a third voltage requirement by the gain unit. This is achieved by multiplying it by an appropriate gain to linearise the relationship between the required voltage and a voltage that will be achieved by the PWM signals. The third voltage requirement is then output as the duty cycles required of the PWM signals.

The control circuit may be used in any implementation in which PWM signals are generated. One example is the field of electric motor control. An overview is shown in Figure 8. In this example the electric motor 805 is driven by the electric grid 801. The main electrical components are the motor-side converter or motor drive 804, a DC bus 803, and a grid-side converter 802. Grid-side converters are required to regulate the DC bus voltage, active and reactive power, as well as comply with grid code regulations. The control circuit shown in Figure 6 might be implemented as part of motor-side control 806 or as part of grid-side control 807. The description below will focus on motor-side control, but the invention is not limited to this implementation.

The behaviour of electric motors can be modelled mathematically by considering their voltages and currents. This modelling tends to be complex as the variables that affect the behaviour of the system, such as induced voltages, currents and flux linkages, change continuously as the electric circuit is in relative motion. For such complicated analysis, mathematical transformations are often used to decouple variables and refer time-varying quantities to a common frame of reference. Clarke and Park transforms are two commonly used transformations. For example, motor drives and grid-side converters are often controlled using field-oriented control, where the three-phase voltages and currents are converted into the d-q reference frame using Clarke and Park transforms.

An example of a motor-side controller is shown in Figure 9. The electric motor is shown at 907. The remaining components 901 to 906, 908 and 909 are part of the system that controls the motor. The system is configured to use field-oriented control to independently control the magnetizing and torque-producing components of the stator current. This requires a series of transform units 904, 905, 908 and 909, which in this example are configured to perform Clarke and Park Transforms.

The motor 907 is a three-phase induction motor that is driven by three PWM signals. The PWM signals are output by PWM generator 906. The motor phase currents and iw are fed back via transform units 908 and 909. The first feedback transform 908 uses the Clarke transform to convert the three phase currents (m it, and iw into real and imaginary currents in a two-dimensional orthogonal system ia, iR. The second feedback transform 909 uses the Park transform to convert the two-axis stationary system into a two-axis rotating system id, ig (i.e. it converts from a stationary to a moving reference frame). The d-axis current may be aligned to the rotor flux and the q-axis current (the torque producing component) may be orthogonal to the rotor flux.

Speed controller 901 may compare a measured speed of motor 907 with a desired speed to generate control currents idref, iqref * These are compared with the feedback currents id, iq to generate difference signals for inputting into d-axis controller 902 and q-axis controller 903. The d-axis and q-axis controllers may be implemented by proportional integral (PI) controllers. The stator current flux and torque are typically controlled independently. The controllers determine voltages vd, vq. These voltages have a moving reference frame. An inverse Park transform 904 converts them to a stationary orthogonal reference frame and outputs va, vs. The description above refers to n-dimensional coordinate systems. An example of a three-dimensional coordinate system is shown in Figure 10. It comprises three orthogonal axes u, v and w (1001, 1002, 1003) that together define a cube. If the three-phase voltages in the controller system are normalised, then the "coordinates" are three numbers between 0 and 1 and the coordinate system can be represented by a unit cube 1004. In the example of Figure 10, the axes of the cube are stationary, i.e. it is a three-axis stationary system. Figure 11 shows an example of a two-dimensional coordinate system. It has two orthogonal axes a and f3 (1101, 1102). It is a two-dimensional stationary axis system. Superimposed upon the axes of Figure 11 is a hexagon 1103 (the same hexagon is shown in Figure 5). This represents the unit cube of Figure 10 viewed down one of its diagonals. It is an isometric projection, where the perpendicular axis is the long diagonal of the cube. The Clarke transform essentially replicates the transform between the unit cube of Figure 10 and the hexagon of Figure 11. The Clarke transform maps per phase components u,v and w (a three-dimensional system) onto a two-dimensional plane with components a and /1.

Returning to the control system of Figure 9, va, vfl generated by transform unit 904 can be considered to represent a coordinate in a two-dimensional system with orthogonal axes a and /3. In an SVPWM system, the time-invariant coordinates va, vp would be used directly to generate the three-phase voltages for the motor. The number of elements here is one fewer than the number of duty cycles required. The third element may be introduced by also considering angle. For example, by calculating a "mean vector" as a combination of a plurality of base vectors, as illustrated in Figure 5. In the control system of Figure 9, the transform unit 905 instead converts va,vfi to per phase voltage components /Au v, and va,. These per-phase voltage components can be considered as representing a coordinate in a three-dimensional system such as that shown in Figure 10. Since there are three elements, there is one element for each phase of the motor. The voltages vu, v, and vw are output to a PWM generator 906, and the resulting PWM signals are provided to the electric motor 907.

A detailed breakdown of processes that might be carried out by the transform unit 905 are shown in Figure 12. This figure relates to one specific implementation and is provided for the purposes of example only. It includes steps that are not essential to the successful operation of the transform unit, and which may be omitted or replaced by alternatives in other implementations.

The transform unit 905 is configured to receive values andd Vadem that are output from inverse Park transform. These values are an indication of an eventual voltage required of the PWM signals. In this example the motor is three-phase, so the number of PWM signals is also three. The received indication comprises two elements, Vpde,, and Va_dem, so one fewer than the number of duty cycles required.

dem and 17", dem can also be thought of as representing a coordinate in a two-dimensional system (for example, that shown in Figure 11).

The voltage demands 17,q dept and Vadeni are first scaled to per-unit values by dividing by the dc-link voltage Vd, (step S1201). This results in two values between 0 and 1, V fl_dent_pu and Vadem_p". These values are converted into voltages having a three-dimensional representation using the inverse Clarke transform (step S1203). This calculation may be expressed as follows: VU_dent_pzt = Va_dem_pu (1) 1%/3 Vv d dnt -encpu = v 2 Vi. e_pu + 2 v _ d ent_pu 1ua'13 VW_dent_pu = v dein-Pu vfl dein 2 2 u -Pu Vv dem_p" VV_dem_pu and Vw_dem_p, together can be thought of as a first voltage requirement of this specific implementation. There is one element for each PWM signal. Each element is then multiplied by itself and summed to generate a squared scaled voltage amplitude (step S1202). This branch of the process then proceeds to deal with challenges that are presented if the required voltages are in the overmodulation zone.

The control circuit is configured to add a zero-sequence component to the voltages output by the inverse Clarke transform (step S1206). Figure 11 shows a flat image of the cube in Figure 10 -it is a two-dimensional projection of a three-dimensional system. Flat images of three-dimensional systems give up one degree of freedom, perpendicular to the plane of view. In this case, if a zero-sequence component is added to any point in the cube, it does not affect where that point shows up in the projection of Figure 11. The plane of view in Figure 11 represents the voltages "seen" by the motor. The zero-sequence component does not affect the voltage between the phases, but it does increase the combined three phase voltage, and thus increase utilisation of the dc link. Step S1206 adds the same voltage identically to each phase. The resulting values can be considered to form a second voltage requirement.

The zero-sequence component is used in both the linear modulation zone and the overmodulation zone. In the linear zone, the zero-sequence component is used to achieve the same maximum (linear) voltage that would be achieved if SVPWM modulation were being used. In the overmodulation zone, the zero-sequence component is injected in dependence on the required modulation index. (2) (3)

It can be seen from Figure 12 that Vamp dem_pu_sqr = 0.333 is equivalent to Mdem = 1 because Kimpdern_pu_sqr = /19dem2/ 3. When Vampdeni_pu_sw, < 0.333, the technique of Figure 12 operates in the linear modulation zone. Pow?, = 1 and the zero-sequence is included in duty cycle calculation. When Vampdem_pusqr > 0.333, the technique operates in the overmodulation zone, but below the threshold at which zero-sequence injection loses efficacy. Pgam is selected in dependence on Vamp dencpu_sqr * The zero-sequence is injected. The zero-sequence might not be used when liamp_dem_puscir > 0.371. When M = 1.055, the waveform of the duty cycle may become a "trapezoidal" shape. M = 1.055 may be the boundary between two different sections of the overmodulation zone. Above the M = 1.055 threshold, injecting a zero-sequence will not benefit the trapezoidal shape. This creates an overmodulation zone between M = 1.055 and M = 1.1 in which Pgain is selected from the look-up table and the zero sequence is not injected. Above M = 1.1 is typically the six-step range.

For example, the zero-sequence component may be injected as follows: When: Vmnp_dent_pu_sq? < 0.333 (0 < M < 1), linear modulation range Gain = 1 zero -sequence injected (Flaa zero = 1) (4) When: 0.333 < Vamp_dein_pu_sqr < 0.371 (1 < M < 1.055), overmodulation range, zone 1 Gain = selected from look -up table zero -sequence injected (Flayzevo = 1) (5) When: 0.371 < Vamp dein_pu_sqr < 0.405 (1.055 < M < 1.1), overmodulation range, zone 2 Gain = selected from look -up table zero -sequence not injected (Fla gZeI0 = 0) (6) An example of a suitable technique for introducing a zero-sequence component is to inject harmonics of the operating frequency to increase the line-to-line voltage. For example, a triangle wave having a frequency that is 3 times the fundamental frequency may be injected as the zero-sequence. The injection affects the line-to-ground voltage but not the line-to-line voltage because the third harmonic on each phase lines up. The zero-sequence injection can be expressed as follows: Vzo = fmin(Vu_deni_pu,VV _dem_pu, VW _dent_pu) + max( VU _dem _ozoliV _dem_pu, VW _dent_pu)} (7) The control circuit includes a flag unit that determines whether the required voltage is comprised in a zero-sequence injection zone (step S1206). Setting a flag F",., to 0 or 1 controls whether a zero-sequence is added to the outputs of the inverse Clarke transform or not.

The control circuit can be considered as operating in an overmodulation zone when the controllers demand a combined three-phase voltage that is greater than that which is achievable solely by increasing the width of the PWM pulses (see e.g. Figure 4 and the associated explanation). Mck,", denotes the demanded modulation index and Al -achieved denotes the achieved modulation index. The modulation index is defined as: V (7) Where M is the modulation index, V is the maximum amplitude (required or achieved) of each per-phase voltage signal supplied to the motor and Vd, is the lineto-line voltage of the dc link.

When 0 < M < 1 there is a linear relationship between the voltage demanded by the controller and the modulated voltage of the PWM signals. Outside of this range is the overmodulation zone. In the overmodulation zone the linear relationship no longer exists because of voltage saturation. This results in control mismatch. To address this, the control circuit applies a gain to the voltages output by the inverse Clarke transform. This gain may be selected from a look-up table (step S1104). In one implementation the gain is selected from a 1-D look-up table based on the squared scaled voltage amplitude. The gain look-up table is suitably precalculated. The gains are suitably calculated to give a linear relationship between m -dent and M -achieved * For an Mdem that lies within the linear modulation zone, the look-up table may simply output a gain of "1" The same gain is applied to all three components, Vudem_pu, Vv_demcp,,, and VI,vdeni_p".

Finally, the control circuit checks and corrects for any saturation issues with the resulting voltages (step S1207). In this example the resulting values are normalised with respect to the do link voltage (see step S1201). They therefore directly represent the required duty cycles of the three PWM signals: r_ duty, r -vduty and rwd"ty. In other implementations the various "voltage requirements" may represent duty cycles indirectly. These can be used to generate the pulses that will drive the gate drivers of the motor controller. They can be considered as the third voltage requirements of this specific implementation.

The terms "first", "second" and "third" voltage requirements may be applied to different values in different implementations. Their use in describing the implementation above is solely illustrative, and they should not be taken to be limited to the specific values and steps with which they are associated above.

The appropriate gain will typically not vary across different motors and/or vehicles. Since there is no need to tune the control circuit for different motors or vehicles, it can be precalculated.

One example of how the look-up table may be calculated is described below. It should be understood that different approaches may be taken to calculating the lookup table and this explanation relates to one specific implementation. It is provided for the purposes of example only. Other techniques may be used and still generate a table that will acceptably linearise the relationship between Mde", and Ma -achieved in the overmodulation zone.

The look-up table may be divided into two segments. The first segment of the look-up table may provide Pgain for 0.333 < Vammdem_pu_sql. 5 0.371 (for Vamp_dem_pusqr 0.333, the modulation is in the linear zone, as explained above). The appropriate Pgai" values for this section of the table may be calculated using the following technique: 1) For the technique illustrated in Figure 12, set Pgai" = 1 2) For a given Vdc, inject the inputs Va_de.," and Vp dem with the following: Mdem Vdc I( = 1 + kC step (8) Where Cstep is a sweeping step value and k = 1,2,3....,0.154/Cs"p.

Then, for each Mdem, calculate the achieved modulation index Machieved from the duty cycles ru_duty, r v_duty and rw_duty output by the process. This enables a table comparing /V/den, and Machieved achieved to be established. The following table is given as an example. It was generated using Cstep = 0.001.

Mdem Machieved -again = Mdem/Machieved 1.001 1.001 1 1.002 1.002 1 1.154 1.055 1.094 Table 1: Gain calculation with Cstep = 0.001 A fitting equation Pgai" = J f (Machieved) achieved) with an acceptable R2 value, which reflects the fitting precision, can be achieved based on the values in Table 1.

lla2dem + VILdem 3) To linearise the overmodulation range, Pgain is used to correct the duty cycle calculation so that kl -dem = Machieved. Assum ing M -dem = M achieved, Vampdem_pu_sqr should be equal to Mdem2/3. Then, the relationship between Pgai" and Vompdem_pusqr can be established, based on the fitting equation. A 1-D Pgain look-up table against Vamp _dem_pzi_ser within (0.333, 0.371] with any number of breakpoints can be generated.

To calculate the second segment of the 1-D look-up table, 0.371 < Varnp_dern_puscir 0.405, the same approach applies, but the sweeping /lide," is changed from (1, 1.154] to (1.154, 1.256].

One aspect of the performance of the motor control circuit is illustrated in Figure 13. The graph plots A4 -achieved against /V/d,. The two are in a linear relationship, even in the overmodulation zone 1301. This linearisation reduces the mismatch between the demanded voltage and the achieved voltage, thereby improving the motor control accuracy. Being able to extend the linearisation into the over modulation zone retains these control benefits while increasing utilisation of the DC link. This enables larger three-phase voltage outputs to be achieved, which means a potential larger torque/power output with the same DC link voltage. An approximately 10% higher voltage output can be achieved compared with existing linearised schemes. This is especially useful for motor control in the automotive and motorsports industries. It reduces the dynamic response time while expanding the motor operating region. In addition, less flux weakening current is needed when the motor operates at high speeds. Overmodulation provides additional voltage output capability. The motor needs flux weakening current later than linear modulation to maintain the voltage within voltage output capability at high speeds.

The control circuit, by implementing the process shown in Figure 12, is thus able to calculate the duty cycle of three phases r" d",y,r,_duty. and rw_d"ty from the voltage demands Vpdern and Vadem. The control circuit performs this calculation irrespective of whether the voltage demands are in the linear modulation zone or the overmodulation zone. In fact, the key processes are identical for both modulation zones. Differences between the different zones are dealt with through gain selection (e.g. having a gain of "1" for the linear modulation zone) and in the setting of the zero-sequence flag. The linear modulation and the overmodulation are therefore extremely well integrated. This reduces complexity compared with solutions that require separate processing for the two zones.

A control circuit for calculating the duty cycles required of the three phase signals for driving an electric motor may be configured to apply a gain to linearise a relationship between a demanded voltage and a voltage that is achievable by the signals. This circuit is simpler and easier to implement than those based on existing arrangements. It offers good compatibility with different processors for implementation because all calculation is arithmetic. The processing can also be achieved at high frequency.

The applicant hereby discloses in isolation each individual feature described herein and any combination of two or more such features, to the extent that such features or combinations are capable of being carried out based on the present specification as a whole in the light of the common general knowledge of a person skilled in the art, irrespective of whether such features or combinations of features solve any problems disclosed herein, and without limitation to the scope of the claims. The applicant indicates that aspects of the present invention may consist of any such individual feature or combination of features. In view of the foregoing description it will be evident to a person skilled in the art that various modifications may be made within the scope of the invention.

Claims

CLAIMS1. A control circuit for generating a duty cycle for a plurality of pulse-width modulated signals, the circuit comprising: an input configured to receive an indication of a voltage required of the pulse-width modulated signals; a transform unit configured to calculate, in dependence on the indication, a respective duty cycle for each of the plurality of pulse-width modulated signals; a gain unit configured to select a gain that will result in a linear relationship between the voltage required of the pulse-width modulated signals and a voltage that will be achieved by the plurality of pulse-width modulated signals, even if the voltage required of the pulse-width modulated signals is comprised in an overmodulation zone, and apply that gain to the respective duty cycles to generate adjusted duty cycles; and an output configured to output the adjusted duty cycles for generating the plurality of pulse-width modulated signals.
2. A control circuit as claimed in claim 1, wherein the gain unit is configured to: determine if the voltage required of the pulse-width modulated signals is comprised in a linear modulation zone; and if so, apply a gain of one to the respective duty cycles to generate the adjusted duty cycles.
3. A control circuit as claimed in claim 1 or 2, wherein the gain unit is configured to: determine if the voltage required of the pulse-width modulated signals comprised in the overmodulation zone; and if so, apply a gain that is not one to the respective duty cycles to generate the adjusted duty cycles.
4. A control circuit as claimed in any preceding claim, wherein the control circuit comprises an adjustment unit configured to add a zero-sequence component to the duty cycles calculated by the transform unit.
5. A control circuit as claimed in any preceding claim, wherein the control circuit comprises a flag unit configured to determine whether the voltage required of the pulse-width modulated signals is comprised in a zero-sequence injection zone.
6. A control circuit as claimed in claim 5, wherein the flag unit is configured to determine, if the voltage required of the pulse-width modulated signals is comprised in a linear zone, that the voltage required of the pulse-width modulated signals is also comprised in the zero-sequence injection zone.
7. A control circuit as claimed in claim 5 or 6, wherein the flag unit is configured to determine, if the required voltage is comprised in the overmodulation zone: that the voltage required of the pulse-width modulated signals is comprised in the zero-sequence injection zone if that voltage is below a predetermined threshold; and that the voltage required of the pulse-width modulated signals is comprised in the zero-sequence injection zone if that voltage is above the predetermined threshold.
8. A control circuit as claimed in any of claims 5 to 7, wherein the flag unit is configured to: if it determines that the voltage required of the pulse-width modulated signals is comprised in the zero-sequence injection zone, cause the zero-sequence component to be added to the duty cycles calculated by the transform unit; and if it determines that the voltage required of the pulse-width modulated signals is not comprised in the zero-sequence injection zone, cause the zero-sequence component not to be added to the duty cycles calculated by the transform unit.
9. A control circuit as claimed in any preceding claim, wherein the adjustment unit is configured to add a zero-sequence component by increasing each duty cycle calculated by the transform unit by an identical amount.
10. A control circuit as claimed in any preceding claim, wherein the control circuit is configured to receive the indication of the required voltage and to output the adjusted duty cycles irrespective of whether the voltage required of the pulse-width modulated signals is comprised in the overmodulation zone or in a linear modulation zone.
11. A control circuit as claimed in any preceding claim, wherein the indication of the voltage required of the pulse-width modulated signals comprises a number of elements that is fewer than the number of the plurality of pulse-width modulated signals.
12. A control circuit as claimed in any preceding claim, wherein the indication of the voltage required of the pulse-width modulated signals represents a coordinate in a two-dimensional system.
13. A control circuit as claimed in any preceding claim, wherein the indication of the voltage required of the pulse-width modulated signals is an output from an inverse Park transform.
14. A control circuit as claimed in any preceding claim, wherein the transform unit is configured to convert the indication of the voltage required of the pulse-width modulated signals into the respective duty cycles using the inverse Clarke transform.
15. A control circuit as claimed in any preceding claim, wherein the duty cycles calculated by the transform unit represent coordinates in a three-dimensional system.
16. A control circuit as claimed in any preceding claim, wherein the number of the plurality of pulse-width modulated signals is three.
17. A control circuit as claimed in any preceding claim, wherein the gain unit is configured to select the gain by accessing a look-up table.
18. A control circuit as claimed in any preceding claim, wherein the look-up table is precalculated.
19. A method for generating a duty cycle for a plurality of pulse-width modulated signals, the method comprising: receiving an indication of a voltage required of the pulse-width modulated signals; calculating, in dependence on the indication, a respective duty cycle for each of the plurality of pulse-width modulated signals; selecting a gain that will result in a linear relationship between the voltage required of the pulse-width modulated signals and a voltage that will be achieved by the plurality of pulse-width modulated signals, even if the voltage required of the pulse-width modulated signals is comprised in an overmodulation zone, and apply that gain to the respective duty cycles to generate adjusted duty cycles; and outputting the adjusted duty cycles for generating the plurality of pulse-width modulated signals.