Control method for electrical converter with LC filter

Abstract

An electrical converter is interconnected via a filter with an electrical load or an electrical power source. A method for controlling the converter comprises the steps of: receiving a reference flux (ψ*.sub.i) for the electrical converter; determining output signals (y) comprising currents and/or voltages measured in the filter; determining an estimated flux (ψ.sub.i) from the output signals (y); determining a corrective flux (ψ.sub.i,damp) from the output signals (y) based on a mathematical model of the filter and a quadratic cost function; determining control input signals (u) for the electrical converter based on a sum of the estimated flux (ψ.sub.i) and the corrective flux (ψ.sub.i,damp); controlling the converter with the control input signals (u); and algorithmic filtering of at least one of the output signals (y) by applying a signal filter to the at least one output signal, which is designed for amplifying the at least one output signal at a resonance frequency of the filter, whereby the corrective flux (ψ.sub.i,damp) is determined from the filtered output signals.

Claims

1. A method for controlling an electrical converter interconnected via a filter with an electrical load or electrical power source, the method comprising the steps of: determining output signals comprising currents and/or voltages measured in the filter; determining an estimated flux from the output signals; determining a corrective flux from the output signals based on a mathematical model of the filter and a quadratic cost function; determining control input signals for the electrical converter based on a sum of the estimated flux and the corrective flux; controlling the converter with the control input signals; algorithmic filtering of at least one of the output signals by applying a signal filter to the at least one output signal, which is designed for amplifying the at least one output signal at a resonance frequency of the filter, whereby the corrective flux is determined from the filtered output signals.

2. The method of claim 1, receiving a reference flux for the electrical converter; setting a flux error to a difference between the reference flux and the sum of the estimated flux and the corrective flux; determining control input signals for the electrical converter based on the flux error.

3. The method of claim 2, wherein the mathematical model is based on differential equations modelling the behaviour of the filter.

4. The method of claim 2, wherein the mathematical model of the filter is solved offline and the corrective flux is determined by integrating a filter voltage calculated via the minimization of the quadratic cost function subject to the mathematical model, with the result being a linear equation from the output signals.

5. The method of claim 2, wherein the output signals at least comprise one of: a converter current between the converter and the filter, a load current between the filter and the load or power source, a converter side filter voltage across the filter on the converter side, and a load side filter voltage across the filter on the load side or power source side.

6. The method of claim 2, wherein the filter comprises an inductor connecting the converter and the load or power source and a capacitor connected to the load or power source and the converter.

7. The method of claim 2, wherein the control input signals are determined by a model predictive control method and a second cost function; wherein the second cost function penalizes the flux error.

8. The method of claim 1, wherein the mathematical model is based on differential equations modelling the behaviour of the filter; and wherein the cost function is not quadratic in the corrective flux.

9. The method of claim 1, wherein the mathematical model of the filter is solved offline and the corrective flux is determined by integrating a filter voltage calculated via the minimization of the quadratic cost function subject to the mathematical model, with the result being a linear equation from the output signals.

10. The method of claim 1, wherein the output signals at least comprise one of: a converter current between the converter and the filter, a load current between the filter and the load or power source, a converter side filter voltage across the filter on the converter side, and a load side filter voltage across the filter on the load side or power source side.

11. The method of claim 1, wherein the filter comprises an inductor connecting the converter and the load or power source and a capacitor connected to the load or power source and the converter.

12. The method of claim 1, wherein the signal filter is designed for different phase shifting of at least one output signal at frequencies different from the resonance frequency.

13. The method of claim 1, wherein the control input signals are determined by a model predictive control method and a second cost function; wherein the second cost function penalizes the flux error.

14. The method of claim 13, wherein the control input signals are determined by modifying switching time instants of a selected pulse pattern for the converter such that the second cost function is minimized.

15. The method of claim 1, wherein the control input signals comprise switching states of the converter.

16. A controller for controlling an electrical converter interconnected via a filter with an electrical load or electrical power source, comprising: the controller structured to determine output signals comprising currents and/or voltages measured in the filter; the controller structured to determine an estimated flux from the output signals; the controller structured to determine a corrective flux from the output signals based on a mathematical model of the filter and a quadratic cost function; the controller structured to determine control input signals for the electrical converter based on a sum of the estimated flux and the corrective flux; the controller structured to control the converter with the control input signals; and the controller structured to perform algorithmic filtering of at least one of the output signals by applying a signal filter to the at least one output signal, which is designed for amplifying the at least one output signal at a resonance frequency of the filter, wherein the corrective flux is determined from the filtered output signals.

17. A converter system, comprising: an electrical converter for transforming a first electrical current into a second electrical current; a filter connected with the converter; and a controller for cant rolling the converter, the controller structured to determine output signals comprising currents and/or voltages measured in the filter; the controller structured to determine an estimated flux from the output signals; the controller structured to determine a corrective flux from the output signals based on a mathematical model of the filter and a quadratic cost function; the controller structured to determine control input signals for the electrical converter based on a sum of the estimated flux and the corrective flux; the controller structured to control the converter with the control input signals; and the controller structured to perform algorithmic filtering of at least one of the output signals by applying a signal filter to the at least one output signal, which is designed for amplifying the at least one output signal at a resonance frequency of the filter, wherein the corrective flux is determined from the filtered output signals.

18. The converter system of claim 17, wherein the electrical converter comprises at least one inverter and at least one active rectifier.

19. The converter system of claim 18, wherein the converter is connected to a load and/or a power source via the filter.

20. The converter system of claim 17, wherein the converter is connected to a load and/or a power source via the filter.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The subject-matter of the invention will be explained in more detail in the following text with reference to exemplary embodiments which are illustrated in the attached drawings.

(2) FIG. 1 schematically shows a converter system according to an embodiment of the invention.

(3) FIG. 2 schematically shows a converter system according to a further embodiment of the invention.

(4) FIG. 3 schematically shows a converter system according to a further embodiment of the invention.

(5) FIG. 4 shows a harmonic model of a converter system according to a further embodiment of the invention.

(6) FIG. 5 schematically shows a converter system according to a further embodiment of the invention.

(7) FIG. 6 shows a flow diagram for a method for controlling a converter according to an embodiment of the invention.

(8) FIG. 7 shows a diagram describing properties of a signal filter for a converter system according to a further embodiment of the invention.

(9) FIG. 8 schematically shows a converter system according to a further embodiment of the invention.

(10) FIG. 9 shows a flow diagram for a method for controlling a converter according to a further embodiment of the invention.

(11) FIG. 10 shows a diagram with pulse pattern for controlling a converter according to an embodiment of the invention.

(12) FIGS. 11A and 11B show diagrams with torque and currents of a converter system during a torque transient.

(13) FIGS. 12A and 12B show diagrams with torque and currents during a torque transient of a converter system according to an embodiment of the invention.

(14) The reference symbols used in the drawings, and their meanings, are listed in summary form in the list of reference symbols. In principle, identical parts are provided with the same reference symbols in the figures.

DETAILED DESCRIPTION OF EXEMPLARY EMBODIMENTS

(15) FIG. 1 shows a converter system 10 with an inverter (DC-to-AC converter) 12 connected on an output side via an LC filter 14 with a rotating electrical machine 16, such as a generator or electrical motor. As indicated, the converter 12 may have a three-phase output. However, the converter system 10 also may be a single phase system.

(16) The inverter 12 produces an N-level output voltage, which is smoothed by the LC filter 14, which comprises a filter inductor L.sub.f connected between the converter 12 and the rotating electrical machine 16. A filter capacitor C.sub.f is connected in parallel to the converter 12 and/or rotating electrical machine 16. It has to be understood that in a multi-phase system, the filter inductor L.sub.f and filter capacitor C.sub.f (as well as the components described below) comprise a number of physical inductors and capacitors corresponding to the number of phases.

(17) FIG. 2 shows a further converter system 10 that additionally has a long cable 18 between the LC filter 14 and the rotating electrical machine 16. The inductance of the long cable 18 is integrated into the LC filter 14 as inductor L.sub.t.

(18) FIG. 3 shows a converter system 10 with an active rectifier (AC-to-DC converter) 12 connected on an input side to a grid 20, which has an impedance L.sub.g. An optional transformer between the converter 12 and the grid 20 is shown as inductor L.sub.t.

(19) It is possible that the setup shown in FIG. 1 or 2 is combined with the setup of FIG. 3 with a DC link and that the converter system 10 has an input LC filter 14 on the input side and an output LC filter 14 on the output side. Also such a system may be controlled with a controller as will be explained with respect to FIG. 5.

(20) FIG. 4 shows a diagram visualizing a harmonic/mathematical model 22 of the converter system 10: an LC filter 14 is connected through an impedance Z(jω) to the total leakage impedance L of the machine 16 or impedance of the grid 20.

(21) All the cases of FIGS. 1 to 3 may be modeled as shown in FIG. 4, as a general impedance Z(jω) (shown as impedance block 24 and an inductance L. L could represent the leakage inductance of an induction machine L.sub.σ (for example as in FIGS. 1 and 2) or the grid impedance L.sub.g (for example as in FIG. 3). The model 22 usually may be applicable to frequencies beyond the fundamental frequency of the converter system 10, for example less than 100 Hz. The impedance block 24 may model one or several storage components such as capacitors or inductors with an associated state, voltage or current, respectively, for example a long cable in between the converter 12 and the machine 16 or grid 20.

(22) The variables in FIG. 4 are named for the case of an inverter 12 connected to a machine 16, but may be also applicable to the other cases described above. FIG. 4 shows the inverter flux ψ.sub.i, which is the time derivative of the inverter voltage v.sub.i (at the output of the inverter 12 or the input of the filter 14), the inverter current i.sub.i (at the output of the inverter 12 or the input of the filter 14), the capacitor current if through the capacitor C.sub.f of the LC filter 14, the filter voltage of v.sub.f (after the LC filter 14), the stator voltage v.sub.s and the stator current i.sub.s.

(23) For simplicity, in the following it is focused on the cases of FIGS. 1 and 2 (with inverter) and L=L.sub.σ, and it is assumed that the impedance Z(jω) is the identity, i.e.,
v.sub.f=v.sub.s,i.sub.s=i.sub.i−i.sub.f

(24) However, the presented results also hold for the general case of FIG. 4. The transfer function relating the capacitor/filter voltage v.sub.f of to the inverter voltage v.sub.i is given by

(25) $\frac{υ_{f} (s)}{υ_{i} (s)} = \frac{(L_{σ} s) / / (\frac{1}{C_{f} s})}{L_{f} s + (L_{σ} s) / / (\frac{1}{C_{f} s})} = \frac{\frac{L_{σ} s}{L_{σ} C_{f} s^{2} + 1}}{L_{f} s + \frac{L_{σ} s}{L_{σ} C_{f} s^{2} + 1}} = \frac{\frac{1}{L_{f} C_{f}}}{s^{2} + \frac{L_{σ} + L_{f}}{L_{σ} L_{f} C_{f}}}$

(26) As such, the resonance frequency of the LC filter 14 is given by

(27) $f_{res} = \frac{1}{2 π \sqrt{\frac{L_{σ} L_{f} C_{f}}{L_{σ} + L_{f}}}}$

(28) As already mentioned, the LC filter 14 is added in order to attenuate unwanted harmonic content of the output currents (and possibly voltages depending on the type of grid codes used) of the converter 12 (rectifier or inverter). As multi-level converter 12 operates in discrete voltage levels that may be fractions of the full DC link voltage, which may produce harmonics at frequencies other than the fundamental frequency f0 (for example 50 Hz).

(29) On the one hand, the LC filter 14 may have a steep attenuation rate for the harmonic content beyond the resonance frequency f.sub.res (for example more than 100 Hz); thus the harmonic content for very high frequencies is almost eliminated. This positive effect is accompanied by a substantial magnification of the harmonic content around the resonance frequency. In particular, since there is no passive resistive element in the LC filter 14, it may create oscillations in the converter system 10, which may have a detrimental effect on stability and performance. This resonance may also cause drastic deterioration in the performance of any underlying controller being used. This is because the control relies on the measured signals to generate correcting actions, and these signals would be tainted with unwanted oscillations, if the filter resonance is left undamped.

(30) Therefore, as will be described in detail with respect to FIG. 5 and the following figures, an outer control loop is added that takes these oscillations into account and induces an artificial damping into the closed-loop system.

(31) FIG. 5 shows a converter system 10 with a converter 12 and an LC filter 14. The block 26 may be seen as an electrical drive of the system 10. The system 10 further comprises a controller 28 that also may be employed in the system shown in FIGS. 1 to 3.

(32) The controller 28 receives output signals y (for example measurement values of currents and/or voltages in the system 10) and generates control output signals u (for example switching instants for the converter 12) based on a (stator) reference flux ψ*.sub.s; and/or a (inverter) reference torque T*.sub.i. Note that as explained below, the stator related quantities all may be replaced with more general, for example grid related quantities and the inverter related quantities may be replaced with more general, for example rectifier related quantities.

(33) The controller 28 comprises a damping block/damping module 30 that generates a corrective (inverter) flux ψ.sub.i,damp from the output signals y that is added to an estimated (inverter) flux ψ.sub.i. In particular, the damping module 30 comprises a signal filter 32 for filtering the output signals y and a regulator 34 for determining the corrective flux ψ.sub.i,damp

(34) The estimated flux ψ.sub.i is provided by an estimator 36 and is calculated from the control input signals u and the output signals y.

(35) The corrected estimated flux, i.e. the sum of corrective flux ψ.sub.s,damp and estimated flux ψ.sub.i is input to a converter controller 38 that need not be aware of the LC filter 14. The converter controller 38 generates the control input signals u based on the reference flux ψ*.sub.s and/or the reference torque T*.sub.i.

(36) FIG. 6 shows a flow diagram for a control method that may be performed by the controller 28.

(37) In step S10, output signals y are determined, which may comprise currents i.sub.i, i.sub.s, i.sub.f and/or voltages v.sub.i, v.sub.f, v.sub.s, measured in the LC filter 14 and/or derived from such currents and/or voltages.

(38) In step S12, at least one of the output signals y is filtered by the signal filter 32, which may be designed for amplifying the at least one output signal at a resonance frequency of the LC filter 14.

(39) In step S14, a corrective flux ψ.sub.i,damp is determined from the filtered output signals y by the regulator 34, which may use the mathematical model 22 of the LC filter 14 and a quadratic cost function as will be explained below.

(40) In step S16, an estimated flux ψ.sub.i is determined from the output signals by the estimator 36.

(41) In step S18, control input signals u are determined by the converter controller 38 for the electrical converter 12 based on a sum of the estimated flux ψ.sub.i and the corrective flux ψ.sub.i,damp as corrected estimated flux.

(42) In step S20, the converter 12 is controlled with the control input signals u. For example, switching instants may be applied to the semiconductor switches of the converter 12.

(43) Embodiments of the modules/blocks 32, 34, 36 and 38 will be described in the following.

(44) Signal Filter

(45) The signal filter module 32 receives outputs signals y, in particular the measured output signals [i.sub.i v.sub.f i.sub.s].sup.T. More precisely, one is interested in extracting the frequency content of all measured signals around the resonance frequency of the LC filter 14. With such a signal filter 32, a desired attenuation outside a certain frequency band around a resonance frequency f.sub.res of the LC filter 14 may be achieved.

(46) FIG. 7 shows two diagrams with a damping of signals (upper diagram) and a phase shift (lower diagram) of the LC filter 14 and the signal filter 32. The frequency of the signals is depicted to the right. The signal filter 32 is designed such that the output signals y are strongest attenuated at the resonance frequency f.sub.res and/or peak 40 of the LC filter 14. Furthermore, the signal filter may compensate for the phase shift of the LC filter 14.

(47) The signal filter module 32 may be based on an n-th order band-pass filter that extracts the content around the resonant peak 40 of the LC filter 14. A choice of such filter with the gain c, m zeros, and n poles is given by the transfer function

(48) $H (s) = \frac{{c (s + \frac{1}{T_{z}})}^{m}}{{(s + \frac{1}{T_{p}})}^{n}}$

(49) The constant 1/T.sub.z dictates the location of the zeros, which should typically fall below the resonance frequency f.sub.res. The constant 1/T.sub.p dictates the location of the poles, which should typically be chosen such that a certain phase shift is achieved at the resonance frequency f.sub.res. The constant c provides a degree of freedom to adjust the DC-scaling. The number of zeros should be less or equal to the number of poles, i.e., m≦n. All parameters may be chosen, for example, to have almost zero phase angle around the resonance frequency f.sub.res.

(50) Regulator

(51) The regulator module 34 is based on the mathematical model 22 as depicted in FIG. 4, which may rely on the filtered version of the measured output [i.sub.i v.sub.f i.sub.s].sup.T provided by the signal filter module 32.

(52) The model 22 may be only valid for frequencies significantly higher than the fundamental frequency of the system 10, including the resonance frequency f.sub.res of the LC filter 14. In mathematical formulas, the model 22 may be described by a linear time-invariant state-space model in the continuous-time domain

(53) $\frac{d}{d t} x = A x + B υ_{i}$

(54) The system matrices A and B may be derived from the underlying circuit for the converter system 10. For example, the mathematical model describing the circuit shown in FIG. 4 without the impedance Z(jω) is given by

(55) $\frac{d}{d t} [\begin{matrix} i_{i} \\ υ_{f} \\ i_{s} \\ ψ_{i, damp} \end{matrix}] = [\begin{matrix} 0 & \frac{- 1}{L_{f}} I & 0 & 0 \\ \frac{1}{C_{f}} I & 0 & \frac{- 1}{C_{f}} I & 0 \\ 0 & \frac{1}{L_{σ}} I & 0 & 0 \\ 0 & 0 & 0 & 0 \end{matrix}] [\begin{matrix} i_{i} \\ υ_{f} \\ i_{s} \\ ψ_{i, damp} \end{matrix}] + [\begin{matrix} \frac{1}{L_{f}} I \\ 0 \\ 0 \\ I \end{matrix}] υ_{i}$ $y = Cx = [\begin{matrix} I & 0 & 0 & 0 \\ 0 & I & 0 & 0 \\ 0 & 0 & I & 0 \end{matrix}] [\begin{matrix} i_{i} \\ υ_{f} \\ i_{s} \\ ψ_{i, damp} \end{matrix}]$

(56) The model is based on differential equations of the (filtered) output signals [i.sub.i v.sub.f i.sub.s].sup.T and the corrective flux ψ.sub.i,damp as states. The mathematical model may be simulated on a computing unit and as such we have access to the corrective flux ψ.sub.i,damp.

(57) Using the mathematical model described above, we define the associated quadratic objective or cost function
custom character =∫(x.sup.TQx+v.sub.i.sup.TRv.sub.i)dt

(58) where Q=C.sup.TC which does not have any penalty on the magnitude of the corrective flux ψ.sub.i,damp. R may be chosen as identity matrix. The problem posed by the mathematical model and the cost function may be seen as a linear quadratic regulator (LQR) control problem.

(59) In the case, the matrices A, B, C, Q and R are not time-dependent, the problem may be solved offline (being a quadratic problem for minimizing the cost function with the above equations as constraints). In this case, the optimal control input (with respect to this problem but not of the controller 28) is then given by
v.sub.i=K.sub.LQRx=−R.sup.−1B.sup.TPx

(60) where the matrix P is positive-definite symmetric and solves the algebraic Riccati equation
0=A.sup.TP+PA+Q−PBR.sup.−1B.sup.TP

(61) The resulting corrective flux ψ.sub.i,damp, which is the integral of the voltage v, is used by the converter controller 38, as will be explained in detail below.

(62) Estimator

(63) The estimator module 36 may have access to measured or computed quantities, based on the state of storage elements in the converter: DC-link voltages, phase capacitor voltages, filter inductor current, etc.

(64) In particular, the estimator module 32 receives or computes the inverter output voltage v.sub.i (possibly by reconstructing it based on the DC link voltage and the switching instants of the control input u). Furthermore, it may receive the measured inverter current i.sub.i and the filter capacitor current i.sub.f. The estimator module 36 may produce at least one or more of the following quantities: The estimated inverter flux ψ.sub.i at the output of the inverter 1 and/or the input of the LC filter 12. The estimated stator and/or filter output flux ψ.sub.s at the output of the LC filter 12 and/or the input to the electrical machine 16. The estimated virtual filter and/or inverter torque T.sub.i at the output of the inverter 12 and/or the input of the LC filter 12. The torque T.sub.i may be computed based on the cross product between the estimated inverter flux ψ.sub.i and the inverter current i.sub.i, or the cross product between the estimated inverter flux ψ.sub.i and the estimated filter capacitor flux. The computation differs by a constant that depends on the filter inductance L.sub.f. The estimated stator or filter output speed ω.sub.s.

(65) Converter Controller

(66) In general, the converter controller 38 may be based on any control method, for instance DTC (direct torque control) and/or PWM (pulse width modulation).

(67) The converter controller 38 may be based on model predictive control, i.e. may comprise a further mathematical model and a further cost function that is optimized with respect to the equations of the mathematical model as constraints. For example, the converter controller 38 may be based on MPDTC (model predictive direct torque control) and/or MP3C (model predicted pulse pattern control).

(68) In the following, a converter controller 38 based on MP3C will be described in more detail.

(69) MP3C is a method to achieve fast closed-loop control of a rotating machine 16 with a voltage source inverter 12 using discrete voltage levels. MP3C relies on optimized pulse patterns (OPPs) with low total harmonic distortion factors that are computed offline. The OPPs are used to generate reference flux trajectories that are to be followed. The core of MP3C is an online computational stage that adjusts the switching instants in the OPPs so as to maintain the flux on the reference trajectory in closed-loop. The stator flux may be controlled based on the estimated stator and rotor fluxes, respectively.

(70) As shown with respect to FIG. 2, the rotating machine 16 is connected to the inverter 12 through an LC filter 14 and possibly long cables 18 and a step-up transformer. Therefore, in the present disclosure, the MP3C method is adapted to handle the oscillations resulting from this resonant overall system.

(71) Inherent to the MP3C mechanism is a core online functionality that compares the estimated flux to the reference flux that is generated from the online-computed OPPs. For example, the estimated stator flux may be compared with the reference stator flux. In the present disclosure it is focused on controlling the inverter flux.

(72) FIG. 8 shows an MP3C converter controller 38 as inner control loop with an active damping module 30 as described above as outer control loop.

(73) The MP3C controller 38 comprises a speed controller (module) 42, a torque controller (module) 44, a flux controller (module) 46, a pattern selector (module) 48, a flux reference controller (module) 50, and a pattern controller (module) 52.

(74) The operation of these modules will be described with reference to FIG. 9, which shows a flux diagram for a method to be performed by the controller 28 of FIG. 8.

(75) The controller 28 may operate in the discrete time domain and/or may be activated at equally spaced time-instants kT.sub.s, with the natural number k being the discrete time-step and T.sub.s denoting the sampling interval. The control problem may be formulated and solved in stationary orthogonal coordinates. The algorithm comprises the following six steps, which are executed at the time-instant kTs.

(76) In step S30, the estimator 36 estimates the estimated inverter flux (vector) ψ.sub.i and estimated stator flux (vector) ψ.sub.s in the stationary reference frame.

(77) Let <ψ denote the angular position of a flux vector and ∥ψ∥ its magnitude. Note, that in case there is a long cable 18 at the output of the inverter 12, then ψ.sub.s would be replaced with the estimated filter flux (vector) ψ.sub.f corresponding to the filter voltage v.sub.f.

(78) In step S32, the flux controller 46 and the pattern selector 48 select a pulse pattern P(m, d), wherein m is the modulation index and d is the pulse number, i.e. number of switches per quarter period of the fundamental frequency.

(79) In step S34, the reference inverter flux (vector) ψ*.sub.i is determined.

(80) The speed controller 42 determines a so-called reference inverter torque T*.sub.i from the difference of a reference speed) ω*.sub.s and an estimated speed) ω.sub.s provided by the estimator 36.

(81) Therefrom, the torque controller 44 then determined the angle or phase<ψ*.sub.i of the reference inverter flux ψ*.sub.i. The inverter torque T.sub.i can be written as T.sub.i=1/L.sub.f∥ψ.sub.s∥∥ψ.sub.i∥ sin γ, where L.sub.f is the resonant filter inductance, and γ is the angle between the inverter flux vectors. For a given value of the stator flux magnitude and a given torque reference, the desired angle between the stator and rotor flux vectors is

(82) $γ^{*} = \sin^{- 1} (\frac{L_{f} T_{i}^{*}}{.Math. ψ_{i}^{*} .Math. .Math. ψ_{s} .Math.})$

(83) The reference flux vector ψ*.sub.i is then obtained by the flux reference controller 50 by integrating the chosen nominal three-phase pulse pattern that is generated by the pattern selector 48 (see below). The reference angle<ψ.sub.s+γ* constitutes the upper limit of the integral. The resulting instantaneous reference flux vector has, in general, a magnitude and angle that slightly differ from their respective values on the unitary circle.

(84) In step S36, the inverter flux error ψ.sub.i,err is computed, which is the difference between the inverter reference flux vector ψ*.sub.i and the estimated inverter flux vector ψ.sub.i,err, with a further correction term ψ.sub.i,damp coming from the active damping module 30 as described above.
ψ.sub.i,err=ψ*.sub.i−ψ.sub.i−ψ.sub.i,damp

(85) In step S38, an optimized pulse pattern is determined by pattern controller 52 from the selected pattern P(m, d). The optimized pulse pattern and the corresponding switching instants are the control input u for the converter 12.

(86) The MP3C control problem can be formulated as an optimization problem with a quadratic objective function and linear constraints, a so-called quadratic program (QP). The objective function penalizes both the corrected flux error ψ.sub.i,err (the controlled variable) and the changes of the switching instants (the manipulated variable), using the diagonal weight matrix Q, whose components are very small. Specifically, the QP is formulated as
min.sub.Δt(∥ψ.sub.i,err−ψ.sub.i,corr(Δt)∥.sub.2.sup.2+Δt.sup.TQΔt)
s.t. kT.sub.s≦t.sub.a1≦t.sub.a2≦ . . . ≦t.sub.an.sub.a≦t*.sub.a(n.sub.a.sub.+1)
kT.sub.s≦t.sub.b1≦t.sub.b2≦ . . . ≦t.sub.bn.sub.b≦t*.sub.b(n.sub.b.sub.+1)
kT.sub.s≦t.sub.c1≦t.sub.c2≦ . . . ≦t.sub.cn.sub.c≦t*.sub.c(n.sub.c.sub.+1).

(87) The corrections of switching instants are aggregated in the vector
Δt=[Δt.sub.a1Δt.sub.a2 . . . Δt.sub.an.sub.aΔt.sub.b1 . . . Δt.sub.bn.sub.bΔt.sub.c1 . . . Δt.sub.cn.sub.c].sup.T

(88) For phase a, for example, the correction of the i-th transition time is given by Δt.sub.ai=t.sub.ai−t*.sub.ai, where t*.sub.ai denotes the nominal switching instant of the i-th transition u.sub.ai. Again, the latter is defined as Δtu.sub.ai=u.sub.a(t.sub.ai)−u.sub.a(t*.sub.ai−dt) with dt being an infinitesimally small time step. Moreover, n.sub.a denotes the number of switching transitions in phase a that are within the prediction horizon, and t*.sub.a(n.sub.a+1.sub.) refers to the first nominal switching transition beyond the horizon. The quantities for phases b and c are defined accordingly.

(89) The switching instants cannot be modified arbitrarily. For the three phases, the set of constraints is imposed, which constrains the switching instants in two ways. Firstly, by the current time-instant kT.sub.s, i.e. transitions cannot be moved into the past. Secondly, by the neighboring switching transitions in the same phase, ensuring that the correct sequence of switching transitions is kept.

(90) FIG. 10, which refers to a model predictive pulse pattern control (MP3C) problem for a three-phase three-level pulse pattern, provides an example to illustrate this. Six switching transitions fall within the horizon T.sub.p, which is of fixed length. The lower and upper bounds for the nominal switching instants are depicted by arrows.

(91) The first switching transition in phase b, for example, is constrained to lie between kT.sub.s and the nominal switching instant of the second transition in phase b, t*.sub.b2. The second switching transition in phase b can only be delayed up to the nominal switching instant of the third transition in the same phase, t*.sub.b3. In this example, the number of transitions that fall within the prediction horizon are n.sub.a=2, n.sub.b=3 and n.sub.c=1. Note that the transitions in a given phase may be modified independently from those in the other phases.

(92) The horizon length T.sub.p is a design parameter. If required, T.sub.p may be increased so as to ensure that switching transitions in at least two phases fall within the horizon. Consider again FIG. 10. In case T.sub.p is smaller than t*.sub.b2−kT.sub.s, it may be increased to this value.

(93) In the end, the pattern controller 52 removes the switching transitions from the QP that will occur within the sampling interval. This may be accomplished by updating a pointer to the look-up table that stores the switching angles of the OPP and the respective three-phase potential values.

(94) In step 40, the pattern controller 53 derives the switching commands over the sampling interval, i.e. the switching instants and the associated switch positions. The switching commands are sent to the gate units of the semiconductor switches in the inverter 12.

(95) Note that the same approach of active damping may also hold for the Dead Beat version of the MP3C method.

(96) Simulation Results

(97) A simulation has been performed for a medium-voltage drive system 10, encompassing a five-level active neutral point clamped (ANPC) inverter 12, an LC filter 14, a short cable 18 and a 6 kV squirrel-cage induction machine 16 rated at 1 MVA with a total leakage inductance of L.sub.σ=0.18 pu.

(98) The rated values of the machine 16 are summarized in the following table.

(99) TABLE-US-00001 Parameter SI value Voltage V.sub.R 6000 V Current I.sub.R 98.9 A Real power P.sub.R 850 kW Stator frequency f.sub.sR 50 Hz Rotational speed ω.sub.mR 1494 rpm

(100) The pu system is established using the base quantities V.sub.B=√{square root over (⅔)}V.sub.R=4899 V, I.sub.B=√{square root over (2)}I.sub.R=139.9 A and f.sub.B=f.sub.R=50 Hz.

(101) The machine, filter and inverter parameters are summarized in Table 2 as SI quantities and pu values, along with their respective symbols.

(102) TABLE-US-00002 Parameter SI value pu value Stator resistance R.sub.s 203 mΩ R.sub.s 0.0057 pu Rotor resistance R.sub.r 158 mΩ R.sub.r 0.0045 pu Stator leakage inductance L.sub.ls 9.968 mH L.sub.ls 0.0894 pu Rotor leakage inductance L.sub.lr 10.37 mH L.sub.lr 0.0930 pu Main inductance L.sub.m 277.8 mH L.sub.m 2.492 pu Number of pole pairs p 2 Filter inductance L.sub.f 12.75 mH L.sub.f 0.114 pu Filter capacitor C.sub.f 31.0 μF C.sub.f 0.341 pu Dc-link voltage V.sub.dc 9.8 kV V.sub.dc 2.000 pu Dc-link capacitor C.sub.dc 200 μF X.sub.dc 2.201 pu Phase capacitor C.sub.ph 140 μF X.sub.ph 1.541 pu

(103) Note that the value of the dc-link capacitance refers to one half of the dc-link, i.e. either the upper or the lower half. The cable 18 is with 100 m very short and can thus be neglected.

(104) The LC filter 14 has very small, effectively negligible, resistors associated with the filter inductor and capacitor. As such, the only passive damping provided by the circuit is due to the machine's stator resistance. Since the stator resistance is very small in a medium voltage setting, the passive damping provided by the sum of resistances in the system is almost zero. This in fact is highlighted by simulation results shown in FIGS. 11A and 11B, which show the electromagnetic torque and the stator currents in pu. At nominal speed, a torque reference ramp from 1 to 0 pu is applied for 10 ms. The MP3C converter controller 28 (without the damping module 30) manipulates the inverter flux vector such that the torque accurately follows this ramp. The ramp excites the filter resonance at 320 Hz. When using baseline MP3C without the outer damping loop provided by the active damping module 30, the resonance is not actively dampened. The passive damping tends to reduce the amplitude of the oscillations, as can be seen in FIG. 11A, but the rate of decay is very slow, amounting to a few seconds.

(105) FIGS. 12A and 12B show corresponding diagrams but for a combination of MP3C controller 28 with an active damping loop provided by the damping module 30. It can be seen that oscillations are quickly and effectively removed within about 10 ms. The remaining minor ripple is due to the 5th and 7th harmonics in the optimized pulse pattern.

(106) When using long cables of lengths amounting to several km or even tens of km, significant ohmic resistance is added to the system, which provides passive damping. As such, the case of an LC filter 14 without a cable of significant length may constitute the worst case, in the sense that this setup provides the least passive damping.

(107) While the invention has been illustrated and described in detail in the drawings and foregoing description, such illustration and description are to be considered illustrative or exemplary and not restrictive; the invention is not limited to the disclosed embodiments. Other variations to the disclosed embodiments can be understood and effected by those skilled in the art and practising the claimed invention, from a study of the drawings, the disclosure, and the appended claims. In the claims, the word “comprising” does not exclude other elements or steps, and the indefinite article “a” or “an” does not exclude a plurality. A single processor or controller or other unit may fulfil the functions of several items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage. Any reference signs in the claims should not be construed as limiting the scope.

LIST OF REFERENCE SYMBOLS

(108) 10 converter system 12 electrical converter 14 LC filter 16 electrical load 18 cable 20 electrical grid 22 model 24 impedance block 26 electrical drive 28 controller 30 damping module 32 signal filter module 34 regulator module 36 estimator module 38 converter control module 40 resonant peak 42 speed controller module 44 torque controller module 46 flux controller module 48 pattern selector module 50 flux reference controller module 52 pattern controller module

Control method for electrical converter with LC filter

Assignee

Inventors

Cpc classification

Classification Explorer

H02P23/14

ELECTRICITY

Classification Explorer

H02M1/0003

ELECTRICITY

Classification Explorer

G05B2219/41406

PHYSICS

Classification Explorer

H02M7/53873

ELECTRICITY

Classification Explorer

H02P21/141

ELECTRICITY

Classification Explorer

H02M1/126

ELECTRICITY

Classification Explorer

H02M7/53875

ELECTRICITY

Classification Explorer

H02P21/30

ELECTRICITY

Classification Explorer

H02P23/30

ELECTRICITY

International classification

Classification Explorer

H02M1/12

ELECTRICITY

Classification Explorer

H02P21/14

ELECTRICITY

Classification Explorer

H02P23/14

ELECTRICITY

Classification Explorer

H02P21/30

ELECTRICITY

Classification Explorer

H02P23/30

ELECTRICITY

Abstract

Claims

Description