RU2560090C1 - Electric motor control unit with permanent magnets without position sensors - Google Patents

Abstract

FIELD: electricity.

SUBSTANCE: control unit without position sensor of the electric motor with permanent magnets uses the induced voltage evaluator is used which uses the current value measured in γ-δ coordinates during the period of zero vector of voltage and the information on electric current derivative, and calculates the electric motor electromotive forces eγ, eδ. The calculated electric motor electromotive forces eγ, eδ are used in the rotation speed evaluator, the estimated angular rotation speed ω^∧ is determined and by time integration of estimated angular rotation speed the estimated phase θ^∧ is calculated.

EFFECT: improvement of accuracy of determination of angular position at low angular rotation speeds of the rotor of electric motor with permanent magnets.

13 cl, 6 dwg

Description

Technical field

[0001] The present invention relates to a permanent magnet motor control device without using position sensors, wherein it relates to a sensorless control device capable of estimating an induced voltage based on information about an electric current derivative, and also estimating an angular rotation speed and a position of a magnetic poles based on the estimated voltage during operation of the said permanent magnet motor at low speed.

BACKGROUND OF THE INVENTION

[0002] Owing to the increased efficiency of the magnets, it has become possible to create synchronous electric motors incorporating permanent magnets as a source of a magnetic field without damper winding (hereinafter referred to as permanent magnet electric motors). When using such motors as servos to control the speed of rotation, information is needed on the position of the magnetic pole of the rotor. However, in many cases, the sensor for measuring position incorporates electrical parts such as semiconductor elements. Moreover, if the sensor is small, its mechanical strength is low. Thus, in a permanent magnet motor there are problems associated with reliability and resistance to environmental factors. Accordingly, in applications that do not require high accuracy and sensitivity of position control, there may be a need to increase reliability and controllability by applying a control method without using position sensors, in which the position of the magnetic pole is calculated based on information about the voltage and current of the inverter without using position sensor, and therefore a lot of research has been done.

[0003] A control method without using position sensors is based on one of two principles described below. The first method is a method for calculating a component of a motor electromotive force generated by a magnetic flux of an exciting electromagnet. This method is used to calculate the motor electromotive force arising from the flux of the magnetic field of the motor, based on the component of the fundamental harmonic voltage or electric current at the inverter output. Permanent magnet motors can be classified as having an implicit pole property, which is that the induction component along an axis perpendicular to the d axis (q axis) is equal to the induction component along an axis perpendicular to the d axis (q axis), or an explicit pole property that lies in the fact that the induction component along the d axis is not equal to the induction component along the q axis. The present method is applicable to both of these types, and therefore is applicable to most permanent magnet motors. However, in situations where the angular velocity of the engine is in the region of low speeds, the motor electromotive force is small. As a rule, the inverter is configured to provide electric power to the motor, and voltage control is based on a pulse-width modulated (PWM) system. When measuring the inverter output voltage, a problem arises, which is that accurate and quick measurement of the voltage of the fundamental component of the PWM signal containing high-frequency components is a difficult task.

[0004] There is a method in which, instead of measuring voltage information, a voltage command in the state before applying PWM modulation is used. This method is influenced by a voltage error resulting from the delay time or the delay of the switching element. When the angular speed of rotation of the electric motor with permanent magnets decreases, when the motor electromotive force becomes small, the mentioned voltage error does not decrease. Accordingly, the voltage error becomes relatively more significant, making it difficult to implement precise control without the use of position sensors.

[0005] There is another way of estimating the position of the axis of an explicit magnetic pole when measuring induction by superimposing a high-frequency component on a voltage component or an electric current component at the inverter output. This method allows you to calculate the phase of the axis of the magnetic field pole by sequentially measuring the induction of each axial component, if the permanent magnet motor has the property that the induction along the d axis is different from induction along the q axis. However, this method is not applicable to permanent magnet motors having an implicit pole property. Moreover, since the axis of the magnetic pole of the excitation (axis d) has two polarities, i.e. the north pole and the south pole, then an additional method of identifying the magnetic pole must be introduced into the control system to recognize its type based on magnetic saturation.

[0006] That is, the first method in which the electromotive force is calculated cannot be accurately performed in the region of low angular rotational speeds. The second method, in which a high-frequency wave is superimposed, can be performed in the region of low angular rotational speeds, but is limited to specific types of motors and cannot be used for a permanent magnet motor that has an explicit pole property or does not have sufficient magnetic saturation.

Documents Describing the Prior Art

Non Patent Documents

[0007] Non-Patent Document No. 1: M. Janson, L. Harnefors, O. Wallmark, and M. Leksell: “Synchronization at Startup and Stable Rotation Reversal of Sensorless Nonsalient PMSM Drives” rotation of implicit PMSM drives without the use of sensors ”), IEEE Trans. IE, Volume 53, No. 2, pp. 379-387 (2006),

Non-Patent Document # 2: L. Harnefors, M. Janson, R. Ottersten, and K. Pietilatinen: "Unified Sensorless Vector Control of Synchronous and Induction Motors" ("Unified Sensor Vector Control of Synchronous and Induction Motors without the Use of Sensors") , IEEE Trans. Ind. Electron., Volume 50, No. 1, pp. 153-160, February 2003.

SUMMARY OF THE INVENTION

[0008] Non-patent document 1 describes a position control circuit at a low angular rotation speed without using sensors, which controls without using sensors based on a control unit in accordance with the illustration of FIG. 6. The control circuit described in document No. 1, which is not a patent, is, in essence, one of the types of schemes for calculating the electromotive force. The principle of its operation is described in document No. 2, which is not a patent. The calculation is carried out by the rotation speed estimation section circled by the dashed line in FIG. 6.

[0009] Namely, the induced voltage e is calculated by subtracting the component of the resistance of the winding and the component of the voltage drop from the voltage control signal Vs, wherein said component of the resistance of the winding is the product of the signal i _ref controlling the electric current and the constant Rs + jw ₁ Ls of the motor, and said voltage drop component is a component of the armature reaction. When controlling the position without using sensors, the estimated phase of the magnetic field pole is used as the control reference axis (d axis), while with this reference axis they work like a real axis. The imaginary component of the induced voltage e is calculated as the induced voltage eq along the q axis, while the real part of the induced voltage e is calculated as the induced voltage ed along the e axis. For the components in the reference control axes, the following corrections are performed: the voltage component eq is corrected by multiplying by the gain coefficient λs, corresponding to the calculated angular velocity ω ₁ and ed, this induced voltage is divided by the magnetic flux Ψm, and then used to estimate the rotation speed. In fact, since there is a feedback component in the voltage component eq arising from the perturbation and position estimation, as well as due to the transformation of rotational coordinates, the angular velocity ω _{1 is} calculated after applying a low-pass filter that limits the passband to prevent generation due to this feedback loops.

[0010] One of the distinguishing features of the circuit illustrated in FIG. 6, lies in the fact that even if the direction of rotation is incorrectly defined as normal rotation or reverse rotation in the region of zero angular velocity of rotation, the estimated angular velocity of rotation is automatically corrected to the correct direction of rotation. However, for example, in a circuit in which the control voltage Vs is applied to the input of a traditional PWM modulation, if the countermeasure is not enough to correct the voltage error resulting from the delay time or switching delay of the switching element, then even if the rotation direction is determined erroneously, this error it is impossible to detect instantly until the angular velocity of the reverse rotation reaches a certain value, and therefore, it will not be possible to return to the correct direction of time scheniya until the reverse rotation angular speed reaches a certain value.

[0011] The purpose of the present invention is to further reduce the lower limit of the angular velocity of rotation, from which you can return to the correct direction of rotation, even if there is a voltage error between the voltage command and the actual voltage, and accordingly, to propose a permanent magnet motor control device without using sensors a position that is capable of performing accurate calculations even at low angular rotational speeds.

[0012] In accordance with one aspect of the present invention, there is provided a control device without using position sensors for a permanent magnet motor, in which: an electric current control signal is generated based on a speed control signal and an estimated speed; a voltage control signal is generated based on said electric current control signal and a measured value of electric current in the γ-δ coordinate system, wherein said measured electric current value is measured by a rotational coordinate conversion section; and said permanent magnet motor is controlled by an inverter, while the inverter is controlled by a section for the inverse transformation of rotational coordinates and PWM modulation; characterized in that the induced voltage calculation section calculates the electromotive force (e _γ , e _δ ) based on the measured electric current value (i _γ , i _δ ) and information (pi _γ , pi _δ ) on the electric current derivative γ-δ coordinate system during the period of the zero voltage vector; the angular rotational speed estimation section determines the estimated angular velocity (ω ^∧ ) based on the electromotive force (e _γ , e _δ ) calculated by the induced voltage calculation section; the estimated phase (θ ^∧ ) is calculated by integrating said estimated angular velocity of rotation over time; and said evaluation phase is outputted to said rotational coordinate transformation section and to a rotational coordinate inverse transformation section, and is used as a reference phase of a rotating coordinate system.

[0013] In accordance with one aspect of the present invention, it is characterized in that said induced voltage calculation section calculates a motor electromotive force (e _γ , e _δ ) using the following equation:

[0014]

[0015] where R is the resistance of the winding, L is the inductance of the winding, φ _d is the magnetic flux of coupling, and p is the differentiation operator (d / dt). In accordance with one aspect of the present invention, it is characterized in that said rotation speed estimating section calculates said estimated angular velocity (ω ^∧ ) using the following equation:

[0016]

[0017] where α = α ₀ + λ ₂ | ω ^∧ |, α ₀ is a constant, λ ₁ , λ ₂ are positive gains, and z ^-1 is the last read value. In accordance with another aspect of the present invention, it is characterized in that on the input side of the measured electric current into said rotational coordinate conversion section there is an electric current derivative determination section, and said rotational coordinate conversion section receives three-phase electric current measurement input data during a period of zero vector voltage of said PWM modulation and derivative of this electric current; and said induced voltage calculation section calculates a motor electromotive force (e _γ , e _δ ) based on a measured current value (i _γ , i _δ ) of electric current supplied to its input and information (pi _γ ′, pi _δ ′) about the derivative of electric current, obtained by converting rotational coordinates to said rotational coordinate transforming section.

[0018] In accordance with another aspect of the present invention, it is characterized in that said induced voltage calculation section calculates said electromotive force (e _γ , e _δ ) using the following equation:

[0019]

[0020] In accordance with another aspect of the present invention, it is characterized in that said information (pi _γ ′, pi _δ ′) about the electric current derivative is obtained by calculating the derivative after converting said section to transform rotational coordinates into components (i _α , i _β ) electric current along two orthogonal axes (α, β) of a fixed coordinate system.

[0021] In accordance with another aspect of the present invention, it is characterized in that there is a voltage drop correction section on the input side of said rotational coordinate conversion section for correcting a voltage drop on the switching element, said rotational coordinate conversion section receiving input data of three-phase electrical measurement the current performed during the period of the zero voltage vector of said PWM modulation, and the input voltage drop correction data from voltage drop correction section; and said induced voltage calculation section calculates said motive electromotive force based on an input value of the measured electric current, said information about the derivative of the electric current, and said voltage drop correction.

[0022] In accordance with another aspect of the present invention, it is characterized in that said derivative of an electric current is based on a rotating coordinate system or on a fixed coordinate system.

[0023] In accordance with another aspect of the present invention, it is characterized in that said induced voltage calculation section calculates a motor electromotive force (e _γ , e _δ ) using the following equation:

[0024]

[0025] where ν _ceγ , ν _ceδ are voltage corrections. In accordance with another aspect of the present invention, it is characterized in that said induced voltage calculation section calculates said electromotive force (e _γ , e _δ ) using the following equation:

[0026]

[0027] In accordance with another aspect of the present invention, it is characterized in that said information about the derivative of electric current is introduced into said section of conversion of rotational coordinates in the form of a calculated value of the difference between the values of electric current measured at two points between which there is a carrier peak, during the period of the zero voltage vector.

[0028] In accordance with another aspect of the present invention, it is characterized in that information about the electric current derivative is inputted into said rotational coordinate transformation section as a value obtained by differentiating a moving average value for electric current values measured between the peak and the lower part carrier during the period of the zero voltage vector.

[0029] In accordance with another aspect of the present invention, it is characterized in that said voltage drop correction is determined using tabular data elements corresponding to the measured values of the three-phase electric current, while for both the positive and negative polarity of each of the three phases an individual element from said tabular data elements.

Result of invention

[0030] As described above, in accordance with the present invention, the estimated phase θ ^{∧ is} calculated based on the signals i _γ , i _{δ of the} electric current in the coordinate system γ-δ during the period of the zero voltage vector and based on the information pi _γ , pi _δ о derivative of electric current. This allows you to calculate the induced voltage with the elimination of the influence of the delay time, and therefore, to obtain normal control without the use of position sensors in the range of lower speeds.

Brief Description of the Drawings

[0031] FIG. 1 is a block diagram of a configuration of a control device in accordance with one embodiment of the present invention.

FIG. 2 is a block diagram of a configuration of a control device in accordance with another embodiment of the present invention.

FIG. 3 is a block diagram of a configuration of a control device in accordance with yet another embodiment of the present invention.

FIG. 4 is a block diagram illustrating electric current measurement in accordance with the present invention.

FIG. 5 is a timing chart illustrating a simulation result in accordance with the present invention.

FIG. 6 is a block diagram of a configuration of a known control device without using sensors for a permanent magnet motor.

Preferred Embodiments of the Present Invention

[0032] The present invention is intended to calculate a motor electromotive force based on measured electric current signal information and electric current derivative information inputted to the induced voltage calculation section, for calculating an estimated phase of a magnetic pole based on said motor electromotive force. Below the present invention is described in more detail by the example of options for its implementation.

Option No. 1 of the invention

[0033] FIG. 1 is a block diagram of a control device without using position sensors in accordance with a first embodiment of the present invention. Position 1 denotes an inverter with PWM control. Position 2 denotes a permanent magnet motor. In the present description, the magnetic axis of the north pole excitation of a real permanent magnet motor is called the d axis, and the phase shifted 90 degrees in the direction of normal rotation relative to the d axis is called the q axis. However, the d axis and q axis cannot be measured directly since the position sensor is not used. Accordingly, the north pole obtained by estimating the position of the magnetic pole is called the γ axis, and the phase shifted 90 degrees in the direction of rotation relative to this γ axis is called the δ axis.

[0034] Reference numeral 3 denotes a rotational coordinate transformation section that performs coordinate transformations of the input three-phase electric current i _u , i _v , i _w measured by the current sensor to the values of i _γ , i _δ in the evaluation axes. The signals i _γ , i _{δ of the} electric current obtained by the coordinate transformation are input to the derivative of the electric current determination section 4, to the induced voltage calculation section 5 and to the electric current control section 9. Section 4 for determining the derivative of electric current determines the rate of change of electric currents in the γ-δ axes during the period of the zero voltage vector, pi _γ , pi _δ , and then enters these values into section 5 for calculating the induced voltage. The induced voltage calculation section 5 calculates the electromotive forces e _γ , e _δ based on the input signals i _γ , i _{δ of the} electric current and information pi _γ , pi _δ of the derivative of the electric current in accordance with the following description. Position 6 denotes the rotation speed estimation section, and position 7 indicates the integrator. The integrator 7 integrates the estimated speed obtained by the rotational speed estimation section 6 in order to calculate the estimated position θ ^{∧ of the} magnetic pole and then outputs it to the rotational coordinate transformation section 3 and the rotational coordinate inverse transformation section 10. Position 8 denotes a section for controlling the angular speed of rotation, and position 9 denotes a section for controlling electric current.

[0035] The operation of the device is described below. The speed control section 8 outputs an electric current control signal i _δ * of the component along the δ axis corresponding to the torque control signal based on the information entered therein about the angular speed control signal ω * and the estimated angular speed of rotation ω ^∧ . This electric current control signal i _δ * and an arbitrary electric current control signal i _γ * along the γ axis are input to the electric current control section 9. The electric current control section 9 receives the input data of the electric current signals i _γ , i _δ obtained by conversion to a rotating coordinate system using the rotational coordinate conversion section 3. The electric current control section 9 performs feedback calculation by comparing the electric current control signals i _γ *, i _δ * and the electric current signals i _γ , i _δ and generates voltage control signals ν _γ *, ν _δ * in the rotating coordinate system (coordinate γ-δ system) based on the calculated axis of the magnetic pole. The voltage control signals ν _γ *, ν _δ * are supplied with the inverse transformation of the rotational coordinates and with the conversion from two-phase to three-phase voltage in the section 10 of the inverse transformation of the rotational coordinates, which functions in the opposite way compared to the section 3 of the conversion of rotational coordinates, and then introduced as phase control signal ν _u, ν _γ, ν _w AC inverter 1. The inverter 1 increases three-phase power signal ν _u, ν _γ, ν _w AC control by PWM modulation and thereby produces an almost equivalent voltage. The control unit is described above, which is common for the case in which there is a position sensor, and for the case in which there is no position sensor.

[0036] The following describes the control subsystem without the use of position sensors. Section 4 determining the derivative of the electric current measures the rate of change of currents pi _γ , pi _δ in the γ-δ axes during the period of the zero voltage vector. The currents in the γ-δ axes are obtained by converting rotational coordinates from three-phase electric currents i _u , i _v , i _w . The said transformation of rotational coordinates is realized using a continuous signal method, using an analog multiplier or an analog adder-subtractor, or using a discrete signal method to convert three-phase electric currents i _u , i _v , i _w using an analog-to-digital converter (ADC) into digital signals, and then perform the calculation in the signal conversion section using a digital converter, such as a central processing unit (CPU).

[0037] In this embodiment of the invention, it is sufficient to detect information pi _γ , pi _δ about the derivative of the electric current in the form of a rate of change, therefore, the conversion of rotational coordinates can be implemented by either of the two mentioned methods. For example, in the case of analog conversion, in order to obtain the rate of change of pi _γ , pi _δ , the signal obtained after the conversion of rotational coordinates is processed using an analog differentiator and then the differential component of the current is measured during the period of the zero voltage vector and this value is stored. On the other hand, in the case of using the ADC, the electric current is measured and converted in the section for converting the signal to a multitude of times during the period of the zero voltage vector, and the differential component is obtained by approximating the derivative based on discrete values of the electric current at the multitude of times. Signals i _γ , i _{δ of} electric current and information about the derivative of electric current (rate of change) pi _γ , pi _δ obtained using any of the above methods are used to control without using position sensors.

[0038] An equation expressing the basic characteristic of a permanent magnet motor is defined herein. In the orthogonal coordinate system (d-q coordinate system), which is represented by the electric angle relative to the north pole axis of the real device, designated as the d axis, the equation for the voltage of the permanent magnet motor is as follows (1):

[0039]

[0040] where ν _d , ν _q are stresses along the d, q axes; i _d , i _q represent electric currents along the axes d, q; R is the resistance of the winding, L is the inductance of the winding, φ _d is the magnetic flux of communication with the stator winding, which is generated by the magnetic flux of said magnet; ω is the angular velocity of the rotor (expressed by the electric angle ω = dθ / dt), θ is the position of the magnetic pole of the rotor (in the form of an electric angle), and p is the differentiation operator (d / dt). In the present invention, since the control object is a permanent magnet motor having an implicit property, it is assumed that the induction component along the d axis and the induction component along the q axis are equal to each other, and accordingly, they are expressed using a common coefficient.

[0041] Further, in the equation (2), the phase error (axis error) is taken into account with respect to the actual phase θ, θ _e = θ-θ ^∧ , with respect to the estimated phase θ ^∧ in the γ-δ coordinate system. The same variables are used as in equation (1), with the assumption that the true values of the motor impedance and magnetic flux of coupling of the permanent magnet are available. Moreover, the estimated angular velocity ω ^∧ and the estimated phase θ ^∧ are related as ω ^∧ = θ ^∧ / dt.

[0042]

[0043] In equation (2), it is assumed that the rate of change of error of the θ _e axis with time in equation (2) is small, therefore, the term of the differential of the axis error can be omitted. Moreover, in equation (2), the information pi _γ , pi _δ about the derivative of the electric current in the γ-δ coordinate system during the period of the zero voltage vector is obtained by successively applying the transformation of rotational coordinates to the measured values of i _γ , i _{δ of the} electric current using the estimated phase θ ^∧ and differentiation.

[0044] The inverter output voltages during the period of the zero voltage vector are ν _γ = 0, ν _δ = 0. Accordingly, the left-hand side of equation (2) becomes zero, and the equation turns into equation (3) to calculate the electromotive forces e _γ , e _δ generated by the magnetic flux φ _{d of the} permanent magnet coupling.

[0045]

[0046] The position of the magnetic pole is calculated based on equation (3) expressing motor electromotive forces e _γ , e _δ during the period of the zero voltage vector. Equation (4) is obtained by applying the first-order delay line of a discrete system to the equation for estimating speed. The bandwidth of the first-order delay line changes as ω ^∧ = θ ^∧ / dt in accordance with equation (5).

[0047]

[0048]

[0049] where α ₀ is a constant value; λ ₁ , λ ₂ are positive gain, az ^-1 is the last measurement value. The rotational speed estimation section 6 illustrated in FIG. 1, calculates the estimated angular velocity ω ^{∧ of} rotation based on equations (4), (5) and outputs the estimated phase θ ^∧ through the integrator 7. Namely, the induced voltage calculation section 5 outputs the electromotive forces e _γ , e _δ during the period of zero the voltage vector based on the signals i _γ , i _{δ of the} electric current in the γ-δ coordinate system and information pi _γ , pi _δ about the derivative of the electric current during the period of the zero voltage vector using equation (3). Section 6 estimates the speed of rotation performs calculations according to equations (4), (5) using the applied electromotive forces e _γ , e _δ and then gives the estimated angular velocity ω ^∧ . The integrator 7 calculates the estimated phase θ ^∧ by time integration of the estimated angular velocity ω ^∧ and then outputs it to the rotational coordinate transformation section 3 and the rotational coordinate transformation section 10 for use as a reference phase of the rotational coordinate system.

[0050] According to this embodiment of the present invention, in the case where the permanent magnet motor having the non-polar property is controlled by an inverter with PWM modulation, the estimated phase θ ^{∧ is} calculated based on the signals i _γ , i _{δ of the} electric current in coordinate system γ-δ and information pi _γ , pi _δ about the derivative of the electric current during the period of the zero voltage vector. This allows you to calculate the induced voltage without affecting the delay time, and therefore, to implement control without the use of position sensors, which normally works in the region to a lower speed limit.

Option No. 2 of the invention

[0051] FIG. 2 illustrates a second embodiment of the present invention, which is different from the first embodiment of the present invention illustrated in FIG. 1, in that the electric current derivative determination section 11 operates in a fixed coordinate system to obtain information pi ′ _γ , pi ′ _δ . In the first embodiment of the present invention, in order to correctly obtain information about the derivative of the electric current, the estimated phase θ ^∧ , which is used to convert the rotational coordinates, must be continuous or similar continuous. Accordingly, in the case where the signals of a three-phase electric current are converted by an analog-to-digital converter into digital values, the estimated phase θ ^∧ used to convert the rotational coordinates must also be updated sequentially, which increases the computational load. This is considered in the second embodiment of the present invention where information pi _'γ, pi' _δ derivative of the electric current is obtained in the fixed coordinate system, therefore, the estimated phase θ ^∧ not need to successively update, which allows for a simple implementation of the calculations.

[0052] In particular, it is sufficient that the signals i _γ , i _{δ of the} electric current and information pi ′ _γ , pi ′ _δ about the derivative of the electric current are measured during the period of the zero voltage vector only at that time during the period of the zero voltage vector, in which calculates motor electromotive forces e _γ , e _δ . It is also sufficient that the coordinate transformation is performed only during the period of the zero voltage vector. This allows you to reduce the calculation time in the CPU and in the digital circuit that performs the calculation, and therefore, this method has the advantage of being able to be applied even in low-speed circuits.

[0053] Although in FIG. 2 illustrates an example in which the measured three-phase electric currents i _u , i _v , i _w differentiate directly, the calculation of the derivative can be implemented using either an analog signal or a digital signal. Namely, it is sufficient that the information is equivalent to the information pi ′ _γ , pi ′ _δ on the derivative of the electric current necessary for the resulting calculation of the electromotive forces e _γ , e _δ during the period of the zero voltage vector. For example, this can be implemented using another sequence of operations, in which the conversion from three phases to two phases is first performed to obtain current signals in the orthogonal axes of the fixed coordinate system (coordinate system α-β), i _α , i _β , and then perform calculating the differential, and then perform the conversion of rotational coordinates.

[0054] The following describes the calculation of the angular velocity in the case when the information pi ′ _γ , pi ′ _δ on the derivative of the electric current in accordance with FIG. 2. Equation (6) is an equation for the voltage of a permanent magnet motor in a coordinate system that is stationary relative to the stator winding.

[0055]

[0056] where ν _α , ν _β are the stresses along the axes α, β; i _α , i _β represent electric currents along the axes α, β; R represents the resistance of the winding; L represents the inductance of the winding; φ _d represents the magnetic flux of the magnet coupling; ω represents the angular velocity of the rotor (in electric angle), θ represents the position of the magnetic pole of the rotor; and p is a differentiation operator. Since equation (1) is processed in a rotating coordinate system, the impedance matrix of the first term on the right-hand side of equation (1) contains the component ω ^∧ L, which corresponds to the electromotive force due to the magnetic flux generated by the winding current. On the contrary, they work with equation (6) in a fixed coordinate system, such a component is absent, therefore, the equation is even simpler.

[0057] In the case where the differential component in the fixed coordinate system is directly converted into a rotational coordinate system rotating with the angular velocity ω of the rotor, it is necessary to use the transformation formula below in the vector calculus in which the orthogonal angular velocity vectors are specified in the α- coordinate system β. Namely, it is necessary to take into account the term of the electromotive force of the magnetic flux of the armature reaction in a strict system of differential equations that takes into account the phase change used to convert rotational coordinates.

pi _{(d, q)} + ω × i _{(α, β)} = pi _{(α, β)}

However, information pi ′ _γ , pi ′ _δ on the derivative of the electric current in the rotational coordinate system in FIG. 2 are obtained by differentiating in a fixed coordinate system and transforming rotational coordinates using the estimated phase θ ^∧ at that moment in time of transforming rotational coordinates in which there is no component of phase changes.

[0058] The following is a description of an example in which the above is implemented by computing using a discrete system. In a discrete system, the calculation of the derivative of electric currents along the αβ axes can be approximated by the difference in the following equation.

[0059]

[0060] where t + ΔT represents two time points of electric current measurement having the interval ΔT during the period of the zero voltage vector, and Δi _α , Δi _β represent changes in electric current over the time interval ΔT.

[0061] In order to use the electric current differences Δi _α , Δi _β to convert the approximated differential component of the rotational coordinates, the estimated phase θ ^∧ for the intermediate time between the time t and the time t + ΔT is calculated and used to convert the rotational coordinates. Namely, since the differences Δi _α , Δi _{β of the} electric current already contain the term ω × i _{(α, β)} , it is necessary to take into account the rate of change of the estimated angle θ ^∧ when converting rotational coordinates.

[0062] The physical meaning of the differential component pi ′ _γ , pi ′ _δ , therefore, differs from equation (2), for which it is necessary to use equation (8). In the γ-δ coordinate system based on the estimated phase θ ^∧ , it is used in equation (8) to take into account the phase error with respect to the actual phase θ, θ _е = θ-θ ^∧ .

[0063]

[0064] The remainder is similar to the first embodiment of the present invention, wherein the change in phase error θ _{e is} assumed to be small, and therefore, its differential component is neglected. During the period of the zero voltage vector, the left side of equation (8) becomes zero, while the equation turns into equation (9) of the induced voltage.

[0065]

[0066] FIG. 2 illustrates a configuration in which the induced voltage is calculated using equation (9), and the angular velocity of rotation and the position of the magnetic pole are calculated similarly to the first embodiment of the present invention. The physical meaning of the mentioned differential component in a fixed coordinate system has been described on the example of a discrete system, since in a discrete system the absence of a phase change is more obvious compared to analog systems. Obviously, the differentiation of the electric current and the transformation of rotational coordinates can be realized both with the help of an analog computational circuit or with the help of a digital circuit. Accordingly, in the second embodiment of the present invention, the calculation of the derivative of the electric current is not limited to discrete systems.

[0067] According to a second embodiment of the present invention, it is sufficient that the rotation coordinate conversion is performed only in the calculation intervals of the induced voltage, since the calculation corresponding to the differentiation of the electric current is performed in a rotating coordinate system. Despite the doubling of the number of coordinate transformations due to the measurement of electric current and differentiation of electric current, the procedure is simpler compared to the case in which the rotational coordinates are sequentially converted.

Option No. 3 of the invention

[0068] In the first and second embodiments of the present invention, the induced voltage is calculated based on the derivative of the electric current during the period of the zero voltage vector. This eliminates the influence of the delay time and the operation delay of the switching element used in the inverter, since the change in electric current is measured over a period of time in which this switching element does not work. However, if we consider the situation more strictly, in a configuration in which the switching elements that form the inverter are semiconductor elements, for example, IGBT transistors or diodes, there is a voltage drop on each element, so the output voltage of the inverter is not zero even during the period of zero voltage vector.

[0069] To account for the above, a third embodiment of the present invention is configured in accordance with the illustration of FIG. 3. Since the component of the voltage drop across the switching element is connected, as a rule, with electric current, then this component of the voltage drop across the switching element is calculated and a correction is introduced based on the measured values of the components of the three-phase electric current. This embodiment of the present invention is applicable to any of the embodiments of the present invention illustrated in FIG. 1 and 2, however, is particularly well suited to apply to the embodiment of the present invention shown in FIG. 2, as illustrated in FIG. 3.

[0070] FIG. 3, 12 denotes a voltage drop correction section that receives an input signal i _u , i _v , i _{w of a} three-phase electric current and determines corrections ν _ceu (i _u ), ν _cev (i _u ), ν _cew (i _u ) values of the components of the electric current based on tabular data. Moreover, in equation (10), conversion to a rotating coordinate system is performed to obtain corrections ν _ceγ , ν _ceδ of the voltage drop.

[0071]

[0072] Then, the induced voltage calculation section 5 produces motor electromotive forces e _γ , e _δ , and equation (9) is transformed into equation (11) by introducing corrections ν _ceγ , ν _{ceδ of the} voltage.

[0073]

[0074] In the case where illustrated in FIG. 1, the first embodiment of the present invention is modified using voltage component corrections, the induced voltage calculation section 5 for determining the electromotive forces e _γ , e _δ performs the calculation according to equation (12), instead of equation (3).

[0075]

[0076] According to this embodiment of the present invention, the distinguishing feature is that the calculation of the electromotive force based on the voltage drop across the semiconductor switching element allows normal control in the region to lower rotation speeds. Moreover, it improves the accuracy of determining the estimated angular velocity and position of the magnetic pole, and also reduces the ripple component 6f of the estimated angular velocity and position of the magnetic pole resulting from a voltage drop on the said element.

Option No. 4 of the invention

[0077] In a third embodiment of the present invention, it is possible to stably operate in a region up to a lower speed limit by introducing a voltage drop correction on the semiconductor switching elements. However, if the elements of the inverter main circuit have a spread of parameters, then accurate voltage correction by voltage drop correction section 12 may not be possible. Moreover, with regard to the corrections ν _ceγ , ν _ceδ and the electric current component for calculating the voltage drop across the resistance R of the winding, they can become a factor that generates a voltage error if they are not consistent with the information about the time derivative of the electric current.

[0078] Accordingly, a fourth embodiment of the present invention is designed to reduce factors causing voltage errors by applying error reduction methods based on statistical processing of data and by selecting time points for measuring electric current. This embodiment of the present invention is limited to discrete systems. In FIG. 4 S0-S7 are the moments of measurement of electric current, and TS is the reference period of the electric current.

[0079] The following is a case in which electric current measurements are performed at eight times S0-S7 in one carrier period. At a low angular rotation speed, the amplitude of the control signal ν _u , ν _v , ν _{w of the} voltage is small, therefore, the operation of the switching element of the three-phase inverter is concentrated in the time period from S1 to S3 and in the period from S5 to S7 near the midpoint of the carrier wave. In accordance with this, there are two types of the period of the zero voltage vector, in which all three phases are in the ON state (on) in the period from S3 to S5, and in which all three phases are in the OFF state (off) in the period from S7 to S9. Accordingly, during these two types of the period of the zero voltage vector, electric current measurement and analog-to-digital conversion are performed twice at time points S3 and S5 or S7 and S9.

[0080] For example, if the electric currents measured at times S3 and S5 are denoted as I3 and I5, then the calculated value of the difference corresponding to the derivative of the electric current for this period is expressed by equation (13).

[0081]

[0082] Such a calculation of the derivative of the electric current is realized by calculating the difference after converting the rotational coordinates in the first embodiment of the present invention or is implemented by converting the rotational coordinates after calculating the difference in the fixed coordinate system in the second embodiment of the present invention. Thus, in the fourth embodiment of the present invention, the measurement of electric current is carried out in advance at many points in time synchronously with the carrier, and if necessary, the calculation of the difference in electric current is performed on the basis of data between samples at two points between which there is a peak of a carrier wave. In order to control the electric current and the value of the electric current to calculate the voltage drop across the resistance R, the values of the samples of the electric current are used in synchronization with the PWM modulation carrier at time instants S0 and S4 at the vertices of the carrier wave. Namely, to measure the electric current and its derivative, the necessary parts are selected and used to measure the electric current at a plurality of points synchronized with the PWM modulation carrier. A typical three-phase inverter is configured so that the switching elements, for example, IGBTs, in which six reverse conductivity diodes are installed, form three branches. During periods of the zero voltage vector of the two types of paths described above, the currents circulating through the switching elements are different, so the variation in the parameters of the switching elements causes voltage errors. Accordingly, in order to average these two types, to calculate equations (11), (12), we use the moving average of two measurements of the electric current at the peak part and at the bottom of the carrier wave. This allows, using statistical methods, to suppress the component of the voltage error.

[0083] For the voltage drop determined by the voltage drop correction section 12 and the measured electric current value used to determine the voltage drop at the resistance R, the electric current value measured at time moments S0, S8 near the upper peak, or the electric current value can be used measured at time S4 near the lower peak at the bottom of the carrier. However, instead of measuring the electric current at time S4, the average of the measured current values at moments S3, S5 can be used, or the average value of the electric current measurements at moments S7, S9 can be used instead of the value of the electric current measured at time S8.

[0084] FIG. 5 illustrates a simulation result in the case of a fourth embodiment of the present invention, wherein FIG. 5A illustrates angular velocity, FIG. 5B illustrates electric current components along two axes; FIG. 5C illustrates the torque, and FIG. 5D illustrates the difference between the actual position of the magnetic pole and the calculated position of the magnetic pole. FIG. 5 illustrates a case in which the load torque changes at time t2 and the angular velocity control signal changes at time t6. Despite the fact that it is usually possible to accurately calculate the position and angular velocity based on electromotive force only in the region up to about 5-10% of the nominal value, FIG. 5D shows that the error θ _e is very small, while it is possible to accurately calculate the position and angular velocity, even when the angular velocity is less than 5% or crosses 0%, moving from normal rotation to reverse rotation.

[0085] Accordingly, in this embodiment of the invention: (1) time matching is provided, so that differentiation is based on measurements of electric current at or near the carrier peak, used as a measured electric current value, or (2) use the average of the components of the electric current at two points before and after the peak for the difference approximating the derivative of the electric current, instead of the value of the electric current counted synchronously on the peak of the carrier. This allows you to match the differential component of the electric current with the electric current used to correct ν _ce voltage and voltage drop across the resistance R in time, and therefore, reduce the effect of noise. This also makes it possible to weaken the significance of the spread in the parameters of semiconductor elements using statistical methods, while taking a moving average of the values of the electric current for two types of the period of the zero voltage vector in the peak part and in the lower part.

Option No. 5 of the invention

[0086] The voltage drop correction section 12 in the third embodiment of the present invention shown in FIG. 3, uses tabular data for the voltage drop component corresponding to the measured values of the components of the three-phase electric current. Similarly, in the fifth embodiment of the present invention, the voltage drop correction section 12 uses tabular data to calculate the voltage drop component. However, the fifth embodiment of the present invention is configured in such a way that tabular data can be set individually for each phase, as well as for each of the polarities - positive and negative. This allows a more accurate correction, even when the aforementioned semiconductor elements have a spread of parameters.

[0087] Accordingly, this embodiment of the present invention allows correction to be performed individually in accordance with the variation in the parameters of the switching elements, and therefore, precisely adjust the voltage drop component.

[0088] According to the preceding description, in accordance with the present invention, the estimated phase θ ^{∧ is} calculated based on the signals i _γ , i _{δ of the} electric current and information pi _γ , pi _δ about the derivative of the electric current in the γ-δ coordinate system during the period of the zero voltage vector . This allows you to calculate the induced voltage with the elimination of the influence of the delay time, and therefore, allows normal control without the use of position sensors in the region to lower rotation speeds.

Claims (13)

1. A control device without the use of position sensors for a permanent magnet motor, in which:
an electric current control signal is generated based on a speed control signal and an estimated speed;
a voltage control signal is generated based on said electric current control signal and a measured value of electric current in the γ-δ coordinate system, wherein said measured electric current value is measured by a rotational coordinate conversion section; and
said permanent magnet motor is controlled by an inverter, while the inverter is controlled by a section for the inverse transformation of rotational coordinates and PWM modulation;
characterized in that
the induced voltage calculation section calculates the electromotive force (e _γ , e _δ ) based on the measured current value (i _γ , i _δ ) of the electric current and information (pi _γ , pi _δ ) on the derivative of the electric current in the γ- coordinate system δ during the period of the zero voltage vector;
the angular rotational speed estimation section determines the estimated angular velocity (ω ^∧ ) based on the electromotive force (e _γ , e _δ ) calculated by the induced voltage calculation section;
the estimated phase (θ ^∧ ) is calculated by integrating said estimated angular velocity of rotation over time; and
said evaluation phase is outputted to said rotation coordinate transformation section and to a rotation coordinate inverse transformation section, and is used as a reference phase of a rotating coordinate system. 2. The device according to p. 1, characterized in that the said section of the calculation of the induced voltage calculates the motor electromotive force (e _γ , e _δ ) using the following equation:

where R is the resistance of the winding, L is the inductance of the winding, φ _d is the magnetic flux of coupling, and p is the differentiation operator (d / dt). 3. The device according to claim 1 or 2, characterized in that said rotation speed estimation section calculates said estimated angular velocity (ω ^∧ ) using the following equation:

where α = α ₀ + λ ₂ | ω ^∧ |, α ₀ is a constant, λ ₁ , λ ₂ are positive gains, az ^-1 is the last read value. 4. The device according to p. 1, characterized in that:
on the input side of the measured electric current into said rotational coordinate conversion section, there is a section for determining a derivative of electric current, and said rotational coordinate conversion section receives input data of measuring a three-phase electric current during a period of zero voltage vector of said PWM modulation and a derivative of this electric current; and
said induced voltage calculation section calculates a motor electromotive force (e _γ , e _δ ) based on the measured current value (i _γ , i _δ ) of electric current and information ( $$p{i}_{\gamma }^{\prime }$$

, $$p{i}_{\delta }^{\prime }$$

) on the derivative of the electric current obtained by converting the rotational coordinates to the said rotational coordinate transformation section. 5. The device according to p. 4, characterized in that the said section of the calculation of the induced voltage calculates the mentioned motor electromotive force (e _γ , e _δ ) using the following equation:

6. The device according to p. 4 or 5, characterized in that the said information ( $$p{i}_{\gamma }^{\prime }$$

, $$p{i}_{\delta }^{\prime }$$

) about the derivative of the electric current is obtained by calculating the derivative after converting the said section of the conversion of rotational coordinates into components (i _α , i _β ) of the electric current along two orthogonal axes (α, β) of the fixed coordinate system. 7. The device according to p. 1 or 4, characterized in that
there is a voltage drop correction section on the input side of the said rotational coordinate conversion section for correcting the voltage drop on the switching element, said rotational coordinate conversion section receiving input data of three-phase electric current measurement performed during the period of the zero voltage vector of said PWM modulation, and input voltage drop correction data from the voltage drop correction section; and
said induced voltage calculation section calculates said electromotive force on the basis of the input value of the measured electric current, said information about the derivative of the electric current, and said voltage drop correction. 8. The device according to claim 1 or 4, characterized in that said information about the derivative of the electric current is based on a rotating coordinate system or on a fixed coordinate system. 9. The device according to p. 4, characterized in that the said induced voltage calculation section calculates a motor electromotive force (e _γ , e _δ ) using the following equation:

where ν _ceγ , ν _ceδ are voltage corrections. 10. The device according to p. 1, characterized in that the said section of the calculation of the induced voltage calculates the mentioned motor electromotive force (e _γ , e _δ ) using the following equation:

11. The device according to p. 9 or 10, characterized in that the said information about the derivative of the electric current is introduced into the said section of the transformation of rotational coordinates in the form of a calculated value of the difference between the values of the electric current measured at two points between which there is a peak of the carrier, during period of the zero voltage vector. 12. The device according to p. 9 or 10, characterized in that the said information about the derivative of the electric current is introduced into the said section of the transformation of rotational coordinates in the form of a value obtained by differentiating the moving average value for the values of the electric current measured between the peak and the lower part of the carrier in time period of the zero voltage vector. 13. The device according to p. 9 or 10, characterized in that the said voltage drop correction is determined using tabular data elements corresponding to the measured values of the three-phase electric current, while for both the positive and negative polarity of each of the three phases there is an individual element from the mentioned elements of tabular data.

Images