Estimation of the amplitude of a periodic component in a measured signal by a delta-sigma modulator

DE602019061448T2Active Publication Date: 2024-11-06SCHNEIDER TOSHIBA INVERTER EUROPE SAS
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
DE602019061448
Authority / Receiving Office
DE · DE
Patent Type
Patents
Current Assignee / Owner
Filing Date
2019-03-12
Publication Date
2024-11-06
Estimated Expiration
2039-03-12

AI Technical Summary

Technical Problem

Existing sensorless control methods for electrical motors face limitations in precision and cost due to the need for expensive fast Analog to Digital Converters (ADCs) when estimating amplitudes of periodic signals, particularly when the parameter ε is small, as they require measuring at least two points per period according to the Nyquist-Shannon theorem.

Method used

The use of delta-sigma modulators without preprocessing the analog signal, combined with Finite Impulse Response (FIR) filters, to retrieve accurate representations of periodic signals, which are orthogonal and calculated based on a clock signal or received from an external device, improving estimation accuracy and reducing costs by eliminating the need for expensive ADCs.

Benefits of technology

This approach enhances the accuracy of amplitude estimation, reduces costs, and allows for flexible signal processing, enabling better control of motor speed and position, even at low speeds, while detecting abnormal situations through monitoring values like RMS and THD.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader
Need to check novelty before this filing date? Find Prior Art

Description

TECHNICAL DOMAIN

[0001] The invention concerns signal processing, in particular while not exclusively, in the context of control of devices. More particularly, the invention may be used in the context of sensorless control of electrical motors.BACKGROUND ART

[0002] Electrical devices such as motors are generally powered via a variable speed drive connected to a main. Classic voltage / frequency control laws are more and more replaced by sensorless control laws that are able to control both the speed and the torque of the electrical motor, without a mechanical speed or position sensor.

[0003] In the context of the present invention, "sensorless" does not refer to the complete absence of sensors but to the absence of some sensors, such as rotor speed or position sensors. It however generally relies on measurements of motor currents (or potentially of motor voltages).

[0004] Sensorless control of devices, in particular electric motors, may rely on extraction of information through signal injection, based on a measured variable (such as the current values). High-frequency signal injection consists in superimposing a high-frequency signal to the control signals of electric motors. The measured current response of the motor to this supplementary excitation is then extracted from the current measurement and additional signal processing allows to retrieve the speed or the position of the rotor at low or even zero speed.

[0005] The measured signal y(t) can be assumed to be of the form: y t = ∑ i y i t s i t / ϵ + O ϵ p where: y i are unknown signals (also called amplitudes hereafter) to be estimated; s i are independent known periodic functions of period 1 (normalized time), of class I i (piecewise) and globally of class I i-1 ; ε « 1 is a known positive parameter; and O(ε p< ) is a small unknown signal which amplitude is of order ε p< (for some known p).

[0006] The unknown signals y i can then be used to estimate the position of the rotor or its speed, so as to determine the control signal to be applied to the motor.

[0007] Some known techniques may be used to retrieve the unknown signals y i : Band-pass filtering y(t) to isolate the desired frequency and extract the amplitude of the resulting signal using amplitude calculators such as described in US patent 7,045,988 B2; Using a procedure similar to heterodyning and sliding averages filters, which are basic Finite Impulse Response filters, to obtain the respective amplitudes of the components. This technique is for example described in A. K. Jebai, F. Malrait, P. Martin, and P. Rouchon, "Sensorless position estimation and control of permanent-magnet synchronous motors using a saturation model" in International Journal of Control, vol. 89, no. 3, pp. 535-549, 2016, and in P. Combes, A. K. Jebai, F. Malrait, P. Martin, and P. Rouchon, "Adding virtual measurements by signal injection" in American Control Conference, 2016, pp. 999-1005.

[0008] When using High-Frequency signal injection for sensorless control of electric motors, for example, the desired frequency can be obtained by band-pass filtering of the signal y(t) and by using an arctangent function to retrieve the position of the rotor, as detailed in the document Yoon Y.-D. and Sul S.-K., "Sensorless Control for Induction Machines Based on Square-Wave Voltage Injection" in IEEE Transactions on Power Electronics, vol. 29, pp. 3637-3645, 2014.

[0009] Wolfgang hammel et al:: "High-resolution sensorless position estimation using delta-sigma-modulated current measurement", ENERGY CONVERSION CONGRESS AND EXPOSITION (ECCE), 2011 IEEE, IEEE, 17 September 2011 (2011-09-17), pages 2717-2724, XP032067527, DOl: 10.1109 / ECCE.2011.6064133 ISBN: 978-1-4577-0542-7 proposes a technique for injecting an alternating carrier voltage for sensorless position estimation. The high frequency current components are measured using delta-sigma-modulators.

[0010] GB 2 314 707 A discloses a monitoring circuit to detect a predetermined, periodic pattern in the output bit stream if a sigma-delta modulator.

[0011] US 2004 / 225468 A1 discloses a method involving fitting components of a motor control signal to a corresponding number of orthogonal pairs in a set of weighted orthogonal pairs. Frequency pairs providing a reduction in the mean squared error between current wave and the weighted orthogonal pairs that satisfy criteria are identified. Desired frequencies from the identified frequency pairs are compared with a harmonic model to determine a motor speed.

[0012] EP 3 447 923 A1 provides a sigma-delta modulator (SDM) and associated system improving spectrum efficiency of wired interconnection. However, the known techniques suffer from a limited precision and require expensive fast Analog to Digital Converters ADCs, when ε is small, since it is imposed to measure at least two points per period ε, according to the Nyquist-Shannon theorem.

[0013] This lack of precision and cost problems apply, more generally, to the estimation of amplitudes of periodic signals in a measured signal.

[0014] There are therefore needs to improve the estimation of amplitudes of periodic components of a measured analog signal.SUMMARY OF THE INVENTION

[0015] The invention is set out in the appended set of claims.

[0016] The invention proposes to use delta-sigma modulators without preprocessing the analog signal, to retrieve an accurate representation of the analog input while being less expensive than other ADCs of similar quality. In addition, galvanic insulation can easily be embedded.

[0017] According to the invention as claimed the control device comprises at least two filters, including a first filter and a second filter, the first filter is applied to the digital signal and to a first periodic signal, the second filter is applied to the digital signal and to a second periodic signal, and the first and second periodic signals are independent functions.

[0018] Therefore, several information can be extracted from the measured analog signal.

[0019] The first and second periodic signals are orthogonal.

[0020] This improves the accuracy of the estimation of the amplitudes.

[0021] According to some embodiments, the periodic signal or periodic signals is calculated by a computing unit of the control device based on a clock signal.

[0022] This enables to predefine the periodic signals and to increase the responsiveness of the method.

[0023] Alternatively, the periodic signal or periodic signals is / are received by the control device from an external device.

[0024] This enables to improve the flexibility of the method as different periodic signals can be used over time.

[0025] According to some embodiments, the control device is configured to control a variable speed drive in charge of an electric motor, the method comprises, before receiving the measured analog signal, controlling the variable speed drive to inject a motor voltage comprising a control law voltage and an additional voltage component. In that case, the external entity is the variable speed drive.

[0026] Therefore, signal injection can be performed to obtain additional information about the electric motor and to enhance the control law of the electric motor.

[0027] In complement, the additional voltage component can be a high-frequency component.

[0028] This enables to facilitate obtaining the additional information even at low-speed.

[0029] According to some embodiments, adapting the control law comprises estimating a rotor position and / or speed based on at least the amplitude signal.

[0030] This enables to improve the control of the motor position / speed by the variable speed drive.

[0031] According to some embodiments, the monitoring value calculated based on at least the amplitude signal can be: a Root Mean Square, RMS, value; and / or a Total Harmonic Distortion, THDI, value.

[0032] According to some embodiments, the method may further comprise comparing the calculated monitoring value with a preset threshold to decide to issue or not an alert

[0033] Therefore, an abnormal situation can be automatically detected and signaled.

[0034] According to the invention as claimed, the first and second filters are Finite Impulse Response Filter, being a linear combination of basic blocks M, basic blocks M being iteratively defined by: M 0 f t = f t M k + 1 f t = 1 T ∫ t − T t M k f τ dτ where T is an integer multiple of the period of the periodic signals; and where f is a dummy function variable.

[0035] Further objects, aspects, effects and details of the invention are described in the following detailed description of a number of exemplary embodiments, with reference to the drawings.BRIEF DESCRIPTION OF DRAWINGS

[0036] By way of example only, the embodiments of the present disclosure will be described with reference to the accompanying drawings, wherein: FIG. 1 represents a system according to the invention as claimed; FIG. 2 is a flowchart showing the steps of a method according to the invention as claimed. DETAILED DESCRIPTION

[0037] Figure 1 shows a system according to the invention as claimed.

[0038] On Figure 1, a control device 120 is included in a system comprising a variable speed drive, VSD, 110 and an electrical motor 100, According to some embodiments, the control device 120 may be included in the VSD 110. The control device 120 may also be used in any other context requiring signal processing of a measured signal to estimate amplitudes of periodic signals within the measured signal.

[0039] In Figure 1 the amplitudes are used to adapt the control law of the electrical motor 100. In particular, the unknown signals y i can be alternatively used to extract harmonics at known frequencies (for example in the context of Root Mean Square RMS value calculation, or Total Harmonic Distortion THDI calculation is sensors) so as to monitor these values.

[0040] No restriction is attached to the type of electrical motor, which can be for example a three-phase AC motor, such as a Synchronous Reluctance Motor, SynRM, a Permanent Magnet Synchronous Motor, PMSM, or an Induction machine also known as asynchronous motor.

[0041] The VSD 110 can be powered by a power source 111.

[0042] The control device 120 may comprise a measuring unit 121, such as current sensors, configured for measuring currents running through the electric motor 100. On Figure 1, the VSD 100 is powered by a three-phase power source, for illustrative purposes only. In that context, the measuring unit 121 may measure currents on the three phases (or only two of them as the third one can be deduced from the two measured ones).

[0043] Alternatively, the measuring unit 121 may be an interface communicating with a measuring unit of the VSD 110, when the VSD comprises such a measuring unit (current sensor) configured for measuring stator current(s). This enables to reduce the cost of the control device 120, and to take advantage of the fact that VSDs are generally equipped with current measurement units.

[0044] The signal obtained by the measuring unit 121 is an analog signal.

[0045] The control device 120 further comprises a delta-sigma modulation unit 122 that is configured to apply a delta-sigma modulation to the analog signal obtained by the measuring unit 121.

[0046] Regular Analog to Digital Converters, ADCs, other than delta-sigma modulators, sample a received analog signal at given instants and output representations of the instantaneous value of the signal over N bits. To sample a signal of a given frequency, the Nyquist-Shannon theorem imposes to have two samples per period, which is called the Nyquist rate.

[0047] Delta-sigma modulators are ADCs, which operate at higher frequencies but with a lower precision.

[0048] To obtain an accurate measurement from the output signal of the Delta-Sigma modulator (whose frequency is for example 15 MHz and whose resolution is low, for example 1 bit), the measurement can be decimated by 3750 for instance (oversampling ratio), to obtain a signal sampled at lower frequencies (4 kHz in the example).

[0049] Delta-sigma modulators yield a high-frequency 1-bit signal (called bitstream), which is proportional in average to the analog input. To retrieve an accurate representation of the analog input, the average over N (oversampling ratio) samples has to be taken.

[0050] Delta-sigma modulators have the advantage to be less expensive than other ADCs of similar quality, and galvanic insulation can easily be embedded.

[0051] The principle of a delta-sigma modulation is well known and is not further described in the present application.

[0052] The control device 120 further comprises at least one estimator 123 connected at the output of the delta-sigma modulation unit 122. On Figure 1, the control device 120 comprises n estimation units 123.1 to 123.n, n being an integer equal to or greater than 1.

[0053] The invention therefore proposes to directly connect the delta-sigma modulation unit 122 with the one or several estimation units 123 without preprocessing.

[0054] The estimation units 123 comprise computing capabilities or an electronic circuit that is / are configured to determine, based on a respective periodic signals s 1 (t), s 2 (t) ... s n (t), and based on the digital signal y ΔΣ output by the delta-sigma modulation unit 122, amplitude signals y i corresponding to the respective periodic signals s i (t). For example, the estimators can be implemented either on an Application Specific Integrated Circuit, ASIC, a Field-Programmable Gate Array, FPGA, or a Digital Signal Processor, DSP.

[0055] The periodic signals s i (t) are provided by the computing unit 125 or alternatively by an external entity. The computing unit 125 may compute the periodic signals s i (t) based on a clock signal. The computing unit 125 may be integrated in the ASIC, FPGA or DSP mentioned above. Estimation units 123 are described hereafter. The estimation units 123 may be identical (but their outputs differ as they are fed with different periodic signals s i (t)) or may be distinct.

[0056] The estimation units 123 comprise filters, such as Finite Impulse Response, FIR, filters, of order 1, 2, 3 or more. In what follows, there is detailed how to construct estimators of orders 1, 2 and 3, which basic building blocks are iterated sliding averages defined by recurrence as follows: M 0 f t = f t M k + 1 f t = 1 ∈ ′ ∫ t − ∈ ′ t M k f τ dτ where ε' is an integer multiple of ε and f is a placeholder (or dummy) function variable.

[0057] A family of FIR filters used in the estimation units 123 may be: F 1 f t = M 1 f t F 2 f t = M 2 f t + M 2 f t − M 2 f t − ∈ ′ F 3 f t = M 3 f t + 13 4 M 3 f t − 20 4 M 3 f t − ∈ ′ + 7 4 M 3 f t − 2 ∈ ′

[0058] More generally, the used FIR filter may be a linear combination of M 1< , ..., M k< for which the sum of the coefficients is 1. The coefficients do not depend on the periodic signals s i (t).

[0059] The s i (t), i varying between 1 and n, are orthogonal, for example orthogonal with respect to a scalar product over the set of 1-periodic functions, i.e.: < σ 1 | σ 2 > = ∫ 0 1 σ 1 τ σ 2 τ dτ = 0

[0060] The estimator ŷ i for the y i up to O(ε p< ) can be given by: y ^ i = E i p y = F p ys i .

[0061] The order p of the filter allows to adjust the accuracy of the estimators (more accuracy when p is high).

[0062] The E i p can be applied directly to the output y ΔΣ of the delta-sigma modulation unit 122.

[0063] An alternative to the above estimator ŷ i can be: y ^ i = E ι p ˜ y = F p y − ∑ j = 0 i − 1 y j ^ s j s i

[0064] The same estimators in the different estimation units 123.1-123.n can be used, provided that the s i (t) are made orthogonal. The computing unit 125 is configured to make the s i orthogonal, if they are originally not (for example, if they are received from an external entity and they are not orthogonal). Original signals s i' are therefore processed to obtain orthogonal signals s i and the above formula are applied to the orthogonal signals s i . Once the estimations ŷ i are obtained, the computing unit 125 may further be configured to calculate the estimations y' i in the original basis of signals s i '.

[0065] The filters, applied to the signal y ΔΣ instead of y, estimate ŷ i such that ŷ i = y i (t) + O(ε p< ) + o(N -k< ) + O(N -(l+1)< ), where k is the order of the delta-sigma modulation unit 122, N is the oversampling ratio of the delta-sigma modulation unit 122 and l = min i l i , l i being the class of the periodic signal s i (t). The amplitudes y i can therefore be estimated by the estimation units 123.

[0066] For example, in the context of signal injection for sensorless control of the motor 100 at low speed, the measured currents may be expressed as: y t = y 0 t ⋅ 1 + ϵ y 1 t ⋅ s 1 t ϵ + O ϵ 2 when considering a second order expansion, or y t = y 0 t ⋅ 1 + ϵ y 1 t ⋅ s 1 t ϵ + ϵ 2 y 2 t ⋅ s 2 t ϵ + O ϵ 3 when considering a third order expansion.

[0067] The knowledge of y 0 and y 1 (and optionally of y 2 ) may be used as parameters for the design of the VSD 110 control laws.

[0068] y 0 , y 1 (and y 2 ) can be determined with accuracy using the low-cost delta-sigma modulation unit 122, even if 1 / ε is larger than the effective Nyquist frequencies of the y i .

[0069] Once the amplitudes y i has or have been obtained, they are transferred to the control unit 124. The control unit 124 is configured to: adapt the control law of the VSD 110 based on the amplitude(s) y i , by optionally determining the speed and position of the rotor based on the amplitude(s) y i ; and / or the amplitude signals y i can be alternatively used to extract harmonics at known frequencies, for example in the context of the calculation of a monitoring value, such as Root Mean Square RMS value, or Total Harmonic Distortion THDI, so as to monitor this value. As an example, the control unit 124 may further calculate the THDI and compare it with a preset threshold. Based on the result of the comparison, an alert may be generated, in particular if the THDI exceeds the preset threshold.

[0070] To this end, the control unit 124 may comprise a processor, a memory (RAM, ROM, flash memory, etc.) and an output interface for controlling the VSD 110 or for issuing and transmitting an alarm signal.

[0071] An example of sensorless control of a motor of the SynRM type, using signal injection, is described hereafter.

[0072] A model of the SynRM motor can be given by: dϕ SDQ dt = R − θ u Sαβ − R s L − 1 ϕ SDQ − Jωϕ SDQ dθ dt = ω ι Sαβ = R θ L − 1 ϕ SDQ where (θ) is a rotation function of angle theta in positive trigonometric direction.

[0073] The state of the motor is described by φ SDQ , which is the vector of the stator flux in a rotor-oriented DQ frame, and θ, which is the angular position of the rotor.

[0074] The vector of stator voltages in a stationary αβ frame, noted u Sαβ is the control input, while ω, the rotor speed, is a disturbance input, which is to be obtained to achieve a proper control of the SynRM motor.

[0075] When sensorless control is used, the sole available measurement is the vector of stator currents in the stationary αβ frame, noted I Sαβ .

[0076] Parameters of the model are the resistance R s of the stator and the matrix of inductances L = L D 0 0 L Q .

[0077] When signal injection is used, the vector of stator voltages is: u Sαβ t = u ‾ Sαβ t + u ˜ t ϵ , where u ˜ t ϵ is a high-frequency disturbance that is voluntarily superimposed on the control voltage vector u Sαβ . The high-frequency disturbance in the voltage vector creates a high-frequency in the stator flux, which becomes ϕ SDQ = ϕ ‾ SDQ + ϵ R − θ U ˜ t ϵ and in turn creates a disturbance in the measured current, which becomes: where Ũ is the zero-mean primitive of ũ and the undisturbed variables follow the original model: dϕ SDQ dt = R − θ u ‾ Sαβ − R s L − 1 ϕ ‾ SDQ − Jω ϕ ‾ SDQ dθ dt = ω I ‾ Sαβ = R θ L − 1 ϕ ‾ SDQ

[0078] Estimating I Sαβ and (θ) (respectively corresponding to y 0 and y 1 mentioned above) using the invention detailed above, allows to retrieve theta from (θ) = (θ)L -1< (-θ) and use it together with I Sαβ to compute u Sαβ in the control law.

[0079] Figure 2 is a diagram showing the steps of a method according to the invention as claimed.

[0080] At a step 200, the VSD 110 is controlled (by the control device 120 or by any other entity) to inject a motor voltage or motor voltages depending on a control law voltage and an additional voltage component, such as a high-frequency component for example.

[0081] At step 201, the measuring unit 121 acquires an input signal such as measurements of the current (or currents for different phases) flowing through the motor 100. The input signal is an analog signal.

[0082] At step 202, the analog input signal y is processed by the delta-sigma modulation unit 122 to obtain the digital signal y ΔΣ .

[0083] At step 203, in parallel to step 202, the one or several periodic signals s i (t) is / are computed or received from an external entity.

[0084] At one or several steps 204.1-204.n, each of the estimation units 123 estimates a signal y i based on the periodic signal received / computed at step 203 and based on the digital signal y ΔΣ .

[0085] At step 205, the control unit 124 adapts the control law of the VSD 110 and / or extracts harmonics at known frequencies to compute a monitoring value, for example to calculate RMS or THDI as explained above.

[0086] The calculation of RMS or THDI can be followed by an optional step 206 of comparison of the RMS or THDI value with a preset threshold, and an alarm can be generated based on the result of the comparison, as explained above.

[0087] Although the present invention has been described above with reference to specific embodiments, it is not intended to be limited to the specific form set forth herein. Rather, the invention is limited only by the accompanying claims.

Claims

1. A method for estimating amplitudes of a periodic component in a measured signal, the measured signal corresponding to measurements of current flowing through an electric motor (100), wherein the method is performed by a control device (120) configured to control a variable speed drive (110) in charge of the electric motor (100), wherein the control device (120) comprises at least two filters, including a first filter and a second filter, the method comprising the following consecutive operations: - controlling (200) the variable speed drive (110) to inject a motor voltage comprising a control law voltage and an additional voltage component into the electric motor (100); - receiving (201) said measured signal, the measured signal being analog; - performing (202) a delta-sigma modulation on the received analog signal to obtain a digital signal; - applying (204.1-204.n) the first filter to the product of the digital signal and a first periodic signal to estimate a first amplitude signal among said amplitudes, and applying the second filter to the product of the digital signal and a second periodic signal to estimate a second amplitude signal among said amplitudes, wherein the first and second periodic signals are independent orthogonal functions, and wherein the at least two filters are Finite Impulse Response Filters, being a linear combination of basic blocks M, basic blocks M being iteratively defined by: M 0 f t = f t M k + 1 f t = 1 T ∫ t − T t M k f τ dτ where T is an integer multiple of the period of the periodic signals, and where f is a dummy function variable; - adapting (205) a control law of the electric motor (100) based on at least the first and second amplitude signals, or calculating (205) a monitoring value based on at least the first and second amplitude signals.

2. The method according to claim 1, wherein the periodic signals are calculated by a computing unit (125) of the control device (120) based on a clock signal.

3. The method according to claim 1, wherein the periodic signals are received by the control device (120) from an external device.

4. The method according to claim 1, wherein the additional voltage component is a high-frequency component.

5. The method according to claim 1 or 4, wherein adapting the control law comprises estimating a rotor position and / or speed of the electric motor (100) based on at least the first and second amplitude signals.

6. The method according to one of the preceding claims, wherein the monitoring value calculated based on at least the first and second amplitude signals is: - a Root Mean Square, RMS, value; and / or - a Total Harmonic Distortion, THDI, value.

7. The method according to one of the preceding claims, further comprising comparing (206) the calculated monitoring value with a preset threshold to decide to issue or not an alert.

8. A non-transitory computer-readable storage medium, with a computer program stored thereon, said computer program comprising instructions for, when executed by a processor, carrying out the steps of the method according to one of claims 1 to 7.

9. A control device (120) configured for performing the method according to one of claims 1 to 7, the control device (120) comprising: a measuring unit (121) configured for receiving the measured analog signal; a delta-sigma modulation unit (122) configured to perform the delta-sigma modulation on the received analog signal to obtain the digital signal; a first estimation unit (123.1) comprising the first filter to estimate the first amplitude signal among said amplitudes; a second estimation unit (123.2) comprising the second filter to estimate the second amplitude signal among said amplitudes; a control unit (124) configured for: - adapting the control law of the electric motor (100) based on at least the first and second amplitude signals; or - calculating the monitoring value based on at least the first and second amplitude signals, and comparing the calculated monitoring value with a preset threshold to decide to issue or not an alert.