WO1991011847A1

WO1991011847A1 - Method for control of a voltage stiff fundamental-frequency commutated convertor and convertor equipment with means for carrying out the method

Info

Publication number: WO1991011847A1
Application number: PCT/SE1991/000070
Authority: WO
Inventors: Lennart Ängquist
Original assignee: Asea Brown Boveri Ab
Priority date: 1990-02-02
Filing date: 1991-01-31
Publication date: 1991-08-08
Also published as: SE465546B; SE9000374L; SE9000374D0; AU7229891A

Abstract

Convertor equipment comprises a voltage stiff self-commutated and fundamental-frequency commutated convertor (SR) for connection to an alternating voltage network (N) and a direct voltage network (DP, DN, C). The equipment comprises path-determining members (FCC) adapted, based on the measured direct voltage (uD) of the convertor and a frequency reference (fbus), to determine characteristic data (γhex) of an estimated polygon path (HDEK) for a flux reference vector (γref), reference vector forming members (FCC) adapted, based on a phase angle reference arg (γref), to form a flux reference vector (γref) following this path, and commutation-controlling members (FCC) adapted to trigger the next commutation at such a time that the flux reference vector and the flux vector of the convertor will coincide at the point of intersection (C) between the path (A1, GC) of the flux vector and one side (EK) of said polygon path, which side has a direction which deviates from the direction of that polygon side (HD) along which the flux reference vector moves immdiately prior to the commutation.

Description

Method for control of a voltage stiff fundamental-frequency commutated convertor and convertor equipment with means for carrying out the method

TECHNICAL FIELD

The present invention relates to a method for control of a voltage stiff fundamental-frequency commutated convertor with alternating voltage terminals for connection to an alternating voltage network via inductive elements and with direct voltage terminals for connection to a direct voltage source, in which the flux vector of the convertor is formed and utilized for control of the commutation instants of the converter.

The invention also relates to convertor equipment comprising a voltage stiff fundamental-frequency commutated convertor with alternating voltage _^terminals for connection to an alternating voltage network via inductive elements, with, direct voltage terminals for connection to a direct voltage source and with means for forming the flux vector of the convertor.

BACKGROUND ART

A voltage stiff self-commutated convertor may generate an alternating voltage with a frequency and a phase position which are determined by the control equipment of the conver¬ tor. The fact that the convertor is voltage stiff means that it is intended for connection to a direct voltage net¬ work or a direct voltage source with low internal impedance . The magnitude of the direct voltage then determines the amplitude of the generated alternating voltage. In one type of such convertors, pulse width modulation is used for con¬ trol of the amplitude of the alternating voltage and poss¬ ibly also of its curve shape. In such a convertor a large number of commutations per alternating voltage period are made, and the commutating losses therefore become high. Another type of. this category of convertors are the so- called fundamental-frequency commutated convertors . In such a convertor, in principle, each valve in the convertor has only one single uninterrupted conduction interval per alternating voltage cycle. The number of commutations therefore becomes the lowest possible and the commutating losses low, which makes this type of convertor especially suited for use at high power ratings. In such a convertor the amplitude of the alternating voltage is, in principle, completely determined by the direct voltage amplitude and cannot therefore be affected by the control means of the convertor. In certain fundamental-frequency commutated convertors, occasional extra commutations per conduc¬ tion interval of the valves of the convertor occur for the purpose of reducing the harmonic content of the alternating voltage. However, this does not affect the characteristic properties of this type of convertors .

A conventional connection for the main circuits of such a convertor is shown in Figure 2. It has direct voltage connections DT+ and DT- for connection to a stiff direct voltage source, that is, a direct voltage source with low internal impedance. It has three phases a, b, c. Each phase, for example phase a, comprises two series-connected controllable semiconductor valves Tau and Tan. Of these, one of the valves is conducting during 180° of each alter¬ nating voltage period and the other valve during the remai¬ ning 180° of the period. Each valve is antiparallel- connected to a diode Dau and Dan, respectively. The three phase units operate mutually 120° offset in phase . In this way, a three-phase alternating voltage system is obtained at the alternating voltage terminals AT of the inverter. As shown in Figure 2, the valves of the convertor may consist of gate turn-off thyristors but they may alternatively consist of conventional thyristors which are provided with turn-off means, or of transistors. A thyristor and its antiparallel-connected diode may be integrated into a common device. Figure 1 shows such a convertor SR with its direct voltage terminals connected to a direct voltage network DP, DN. A capacitor C is connected between the direct voltage connec¬ tions of the convertor to achieve the necessary low internal impedance. The alternating voltage terminals AT of the convertor are connected to an alternating voltage network N via inductive elements with the resistance r and the reactance xτ_. These elements may consist of separate inductors or of the reactance of a convertor transformer. The network N typically consists of a three-phase power network or of a rotating three-phase machine. The main part of the difference between the instantaneous values of the line voltage, which in principle is sinusoidal, and the square-wave voltage of the convertor is taken up by the reactance x*_, (the voltage drop across the resistor r is normally of minor importance) .

A sinusoidal alternating voltage may be described using a vector which rotates in the complex plane. A corresponding flux, defined as the time integral of the voltage, will then be a vector which in the same complex plane, assuming a constant frequency and amplitude of the alternating voltage, moves in such a way that its tip follows a circular orbit at a constant peripheral velocity. The alternating voltage and the corresponding flux of a self-commutated convertor may in a corresponding way be described using vectors in a complex plane. This is known, for example, from

V G Tδrδk: "Near-Optimum on-line Modulation of PWM

Inverters",

IFAC Control and Power Electronics and Electrical

Drives, Luzern, Switzerland, 1985.

Lennart ngquist: "Stato^'r Flux Control of Asynchronous

Motor in the Field-Weakening Region",

Conference paper, "Evolution and Modern Aspects of

Induction Machines", Turin, July 8-11, 1986, M. Depenbrock: "Direkte Selbstregelung (DSR) fur hochdynamische Drehfeldantriebe mit Stromrichter- speisung",

ΞTZ Archiv, Band 7 (1985), Heft 7,

- M. Depenbrock: "Drehmomenteinstellung im Feldschwach- bereich bei stromrichtergespeisten Drehfeldantriben mit Direkter Selbstregelung (DSR)",

ETZ Archiv, Band 9 (1987), Heft 1,

M. Depenbrock: "Direct Self-Control (DSC) of Inverter- Fed Induction Machine,

IEEE Transactions on Power Electronics, Volume 3, No. 4, October 1988,

- US Patent No. 4,678,248.

From the above-mentioned US patent specification, a motor drive is known which comprises a convertor of the kind described above. With the aid of a torque controller, a reference value for the amplitude of the motor flux is obtained. The flux amplitude reference in turn affects the frequency of the convertor voltage and hence eventually also the phase position of this voltage. A method is proposed for changing, as quickly as possible, the flux amplitude from one value to another.

In convertors of the above kind, which are connected to an alternating voltage network, the path of the flux vector may for various reasons deviate from the desired path, which for a six-pulse convertor consists of an equilateral^' hexagon with its centre at the origin of the complex plane. Reasons for such disturbances in the path of the flux vector may, for example, be the following:

a) disturbances in the alternating voltage network,

b) deficiencies in* the convertor or its control system, c) rapid changes in the direct voltage of the convertor, for example because of transients during start-up or restart of the convertor.

A disturbance of the path of the flux vector means that the path will no longer be centered in the origin of the complex plane. This means that the alternating currents of the convertor form an unsymmetrical three-phase system with a risk of overload of certain valves and of interruption of the operation because of missing commutation. Further, a disturbance of the above-mentioned kind generally gives rise to a phase change of the alternating voltage of the convertor. Such a phase change will entail a change of' the active power flux between the network and the convertor and hence a rapid charge or discharge of the capacitor on the direct voltage side of the convertor, which in turn entails a rapid and undesired change of the direct voltage of the convertor. These disadvantages are especially serious in such applications where it is of the utmost importance that the operation of the convertor should be maintained to the utmost possible extent also during disturbances in the operation of the alternating voltage network. An example of such an application is a convertor which is used as static reactive power compensator. As shown in Figure 1, such a convertor is connected to an alternating voltage network via inductive elements. To the direct voltage side of the convertor a capacitor bank is connected, which maintains the direct voltage required for the operation.

SUMMARY OF THE INVENTION

The invention aims to provide a method and convertor equip¬ ment of the kind described in the preamble to the specification, which provides:

a) the fastest possible recovery after a disturbance in the path of the flux vector permitting the operation of the convertor to be maintained also during disturbances in, for example, the line voltage,

b) the fastest possible symmetrization of the alternating voltage of the convertor after a disturbance, which provides the lowest possible current load on the valves and the lowest possible requirements on the commutating capability of the valves,

c) a possibility of maintaining a desired degree of Constance of the direct voltage of the convertor also with moderate values of the capacitance of this capacitor.

The above advantageous properties are obtained by generating, according to the invention, a flux reference vector in dependence on the line voltage, determining the flux vector of the convertor, and controlling the commutations of the convertor such that the flux vector of the convertor, in the fastest possible way after a disturbance, is brought to pass to a new symmetrical path, while at the same time restoring the desired phase position of the convertor voltage relative to the line voltage. Since in this way undesired deviations of the phase position of the convertor voltage will be restored in the fastest possible way, undesired charges and discharges of the capacitor of the direct voltage side will be maintained at a minimum.

What characterizes a method and convertor equipment according to the invention will be clear from the appended claims.

BRIEF DESCRIPTION OF THE DRAWINGS

In the following the invention will be described in greater detail with reference to the accompanying Figures 1-6. Figure 1, the main circuits of which have been described above, shows a convertor of the relevant kind connected to an alternating voltage network. Figure 2, which has been described above, shows an example of the configuration of the main circuits of the convertor. Figure 3 shows the path in the complex plane for the flux vector of the convertor. Figure 4 shows the- paths of the flux vectors in the complex plane and illustrates how the determination of the instant of the next commutation is done according to the invention. Figure 5 shows an example of the configuration of the circuit which is used for forming the flux vector of the convertor in the equipment shown in Figure 1. Figure 6 shows the mode of operation of the calculating member in the equipment according to Figure 1, which is used for determining the instant of the next commutation.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

Figure 1, the main circuits of which have been described above, shows a convertor which is voltage stiff, self- com utated and fundamental-frequency commutated. The convertor is connected to a three-phase alternating voltage network N. This may consist of a three-phase power network and the convertor, for example, be intended for static reactive power compensation. In this case, the direct voltage network connected to the direct voltage terminal of the convertor consists only of the busbars DP and DN and the capacitor bank C. With the aid of a voltage transformer UM, signals -_bus, which depict the network alternating voltage, are generated. With the aid of current transformer equip¬ ment IM, a three-phase signal i is generated which depicts the alternating currents flowing between the network and the convertor. These signals are supplied to a calculating circuit MC, the function of which will be described with reference to Figure 5. From the calculating circuit a signal ψconv s obtained, which indicates the position in the complex plane of the flux vector of the convertor. Further, the calculating circuit generates a signal ffc_>us which is a measure of the frequency of the network alternating voltage. The calculating circuit also supplies a signal arg(ψ_bUS) which is a measure of the argument of that flux vector which corresponds to the network voltage. A direct voltage measuring device DM, for example a voltage divider connection, generates a measured signal UD which is a measure of the direct voltage of the convertor. This signal is supplied to a summator S2 as well as to a -control circuit FCC. The circuit FCC supplies control signals SS to the convertor, which trigger commutations in the convertor. A reference voltage generator RS, which in its simplest form may consist of a potentiometer, delivers a voltage reference signal o_ref, which is a measure a of the desired direct voltage of the direct voltage side of the convertor. This reference is compared in the summator S2 with the measured voltage UD, and the difference is supplied to a voltage regulator UR, which has a PI characteristic. The output signal δ_ref of the regulator is in steady-state operation that phase angle between the flux vectors of the network and the convertor which is required for the capacitor voltage to be maintained constant at the reference value given by the reference generator RS. The output signal of the regulator is added in a summator SI to the signal arg(ψ_-_>us) , and the sum of these two signals constitutes the argument arg(ψ_ref) for a reference flux vector.

Figure 3 shows the flux vector ψconv and its path in the complex plane. In steady and undisturbed state, the path of the tip of the flux vector consists of a symmetrical hexagon. The flux vector ψconv is formally defined by the following equation:

where u_conv is the voltage vector of the convertor and ω_n is the angular frequency of the network voltage. The length ψconv of the side of the hexagon is determined by the condition that the circumference of the hexagon is to be traversed during one of the network alternating voltage cycles at a peripheral speed determined by the amplitude of the alternating voltage of the convertor, that is, by its direct voltage UD. The side length is obtained from the following equation:

A ^2*π'υ

Figure 4 illustrates the principle of the invention. According to the invention, the flux vector ψconv of the convertor is continuously determined, the tip of the vector at a certain time being assumed to be at point A. Based on the measured value UD of the direct voltage of the ^■ convertor, an estimation is made of the direct voltage value which will prevail up to the commutation which succeeds the immediately following commutation. This estimation is made in accordance with a suitable algorithm. In its simplest form, the estimation consists of the assumption that the direct voltage will be constantly equal to the measured value. Further, according to a suitable algorithm, an estimation is made of the value of the network frequency _bus which will prevail for two commutations onwards in time. In its simplest form, also this estimation consists of the assumption that the frequency will be constantly equal to the value measured at that particular instant. Based on the values thus estimated, the side length ψ ex of an equilateral hexagon, centered at the origin of coordinates, is determined, which hexagon constitutes the path of a reference flux vector ψ_ref, which at the current values of direct voltage and network frequency and in case of undisturbed steady-state operation, would follow this path. Of this hexagon, the upper half HDEK is shown in Figure . The circumference of the path is parameterized with a parameter 0 to 2π for one revolution, and the magnitude of the vector is obtained by finding that point on the path whose parameter value corresponds to the argument arg(ψ_ref) according to Figure 1. The reference flux vector assumes the shown position with its tip at the point 3 while at the same time the flux vector of the convertor has its tip at the point A. The reference vector and the flux vector of the convertor will move in parallel in a direction towards the points D and i, respectively, and at the same velocity, which is determined by the currently measured value of the direct voltage of the convertor. The next commutation is assumed to take place when the flux vector of the convertor is at point Ai. The flux vector will thereafter move along the path AiGC, whereas the reference flux vector moves along the path BiDEC. According to the invention, the instant of the commutation is selected such that the length of the path AiGC equals the length of the path BiDEC. Since the peripheral speeds of the flux vectors are equal, they will therefore occur at the point C simultaneously. The reference flux vector and the flux vector of the convertor will then coincide, that is, the flux vector of the convertor has been brought to assume the desired position in relation to the flux vector of the network. A change of the position of the flux vector, caused by a disturbance, in relation to the reference flux vector will in this way be eliminated in the shortest possible time and completely without overshoot.

As will be clear from Figure 4, the distances CE = CG and ED = GF, that is, DEC = FGC. If the commutation is performed at the moment when B_JD = A_,F, the above-mentioned condition relating the equality of the path lengths will therefore be fulfilled. According to the invention, during the com¬ pletion of the polygon side HD by the reference flux vector, the distance betwen the tip B of the vector and the next corner D of the polygon, namely (l-λ)ψ_hex, and the simul¬ taneous distance between the tip A of the convertor flux vector and the polygon side HD in a direction which is parallel to the next polygon side DE (this distance becomes the same as A*_F in Figure 4) are determined continuously one or several times . The distance in the flux plane from the position of the reference flux vector at the time of measurement and its position at the time of commutation are given by the difference between the measured values calculated above. The distance in the flux plane may be converted into distance in time since the velocity of travel in the flux plane is known by the estimation uoestim- The calculated commutating time t_COm can now be calculated if the time for the measurement has been recorded. This will make it possible to set, in-a running clock, the time when triggering of the commutation is to take place. If the distance to the commutation is sufficiently great, there will be time to carry out several calculations, and the time set for commutation can thus be adjusted several times upon passage of a hexagon side.

Figure 5 shows the configuration of the calculating circuit MC in Figure 1. From the measuring devices IM and UM the circuit is supplied with the instantaneous values of the three phase currents i_{a r} i__ , i_c and the phase voltages u_a, _b, u_c. The reference axis for phase a is assumed to coincide with the positive direction of the real axis of the complex plane. According to known relationships, the real part i_RE and the imaginary part im of the complex vector, which corresponds to the current flowing from the convertor to the network, are now formed. This is done by addition of the instantaneous values of the phase currents in the operational amplifiers Fl and F2 with the proportionality constants indicated in the operational amplifier symbols. In a corresponding way, with the aid of the operational amplifiers F3 and F4, the real part URE and the imaginary part UIM of the vector which describes the network voltage are formed. With the aid of integrators INI and IN2, the real part ψ_bus_RE and the imaginary part ψ_bus_lM of that flux vector ψ us which corresponds to the network voltage are obtained. With the aid of the multipliers MU1 and MU2,

SUBSTITUTE SHEET the quantities i _E ^■ X and im ^• xL are formed. These quantities are added with the aid of summators S3 and S4 to the real and imaginary parts of the network flux vector. The output signals from these summators constitute the real part ψconvRE and the imaginary part ψconvi_M/- respectively, of the flux vector of the convertor. These signals are supplied to the control circuit FCC in the manner shown in Figure 1. The network alternating voltage is supplied to a frequency measuring device FM, which delivers a signal f us which is a measure of the frequency of the network alternating voltage. A calculating circuit AB forms the argument for the flux vector of the network according to the relationship

*»<*bu,s_β>' = *-*^*<-*-B» *?**** busRE

Figure 6 illustates the mode of operation of the control circuit FCC in Figure 1. The circuit may, for example, consist of a fast processor adapted to operate according to the flow chart shown in Figure 6. The circuit first determines the quantity k__, which indicates the serial number of that side in the reference polygon on which the reference flux vector is currently located. For, for example, the polygon side HD, kjj = 0, for the side DE, k__ = 1, etc. Thereafter, a quantity λ is determined, which indicates how great a part of the current polygon side that has been passed. Immediately after each commutation, λ = 0 and immediately prior to a commutation, λ = 1. Thereafter, the side length ψ_hex of the reference polygon is determined according to the condition shown, where U^'D, estim is the estimated value of the direct voltage which is used for the calculation and f us, es_tim is the estimated value of the network frequency which is used for the calculation. As has been stated above, as estimated values of these quantities, the currently measured values of the quantities may be used. Then, the complex reference flux vector ψ_ref(t) is formed. In the next functional block, the- quantity Δψ_COm is formed. This quantity denotes the distance in the flux plane from the current positions of the flux vectors to the time of commutation, that is, the distances BBi = A i in Figure 4. The time remaining till commutation, Δt_COm, is obtained by division of the latter quantity by the velocity of the flux vectors along the path. The remaining time is added to the time t_m t at the time of measurement and a calculated commutating time t_Com is obtained, which may be loaded in a register in a system clock (Load t_Com) • When the system clock (t) passes the loaded-in time, the commutation (COM) which is next in turn is triggered. The calculations and the updating of the commutating time may be made one time or several times per polygon side, depending on the need of calculating time.

A method and a convertor according to the invention have been described above for the case where the convertor is used as a reactive power compensator. The described control will then have to be supplemented, in conventional manner, with control members which control the amplitude of the alternating voltage of the convertor such that the desired reactive power flow between the convertor and the network is obtained. This can be made by influencing the direct vol¬ tage reference u re_f n Figure 1. However, the method and the convertor equipment according to the invention may be used for other purposes than for power factor correction.

In the embodiments described above, the convertor consists of a six-pulse convertor. However, the invention may be applied to other pulse numbers as well, for example to twelve-pulse convertor connections, in which the path of the flux reference vector in steady state becomes an equilateral twelve-sided polygon. The invention can also be applied when six-pulse bridges according to the invention are adapted to cooperate relative to a common network.

Claims

1. A method for control of a voltage stiff fundamental- frequency commutated convertor (SR) with alternating voltage terminals (AT) for connection to an alternating voltage network (N) via inductive elements (r, XL) and with direct voltage terminals (DT+, DT-) for connection to a direct voltage source (DP, DN, C) , at which the flux vector (ψcon_v) of the convertor is formed and utilized for control of the commutation instants of the convertor, characterized in that

a) the voltage (UQ) of the direct voltage source is measured, b) a polygon path (HDEK) for the reference flux of the convertor in a complex vector plane (re, im) is estimated based on the measured direct voltage and a frequency reference (f_bus) r c) based on a phase angle reference arg(ψ_ref), a flux reference vector (ψref) °ⁿ this path is determined , d) the next commutation is carried out at such a time that the flux reference vector and the flux vector of the convertor will coincide at the point of inter¬ section (C) between the path (Ai, GC) of the flux vector and one side (EK) of said polygon path, which side has a direction which deviates from the direction of that polygon side (HD) along which the flux reference vector moves immediately prior to the commutation.

2. A method according to claim 1, characterized in that the next commutation is made when the estimated remaining path lengths (BiDEC and AiFGC, respectively) of the flux reference vector and the flux vector of the convertor, respectively, from the time of commutation up to the time when these vectors coincide at said point of intersection, correspond.

3. A method according to claim 1 for control of a six-pulse convertor, characterized in that the estimated polygon path consists of an equilateral hexagon.

4. A method according to any of the preceding claims, characterized in that the side length (ψ_hex) of the estimated polygon path is proportional to the measured direct voltage (UD) •

5. A method according to any of the preceding claims, characterized in that the frequency (f_bus) of the alternating voltage network is measured and constitutes said frequency reference.

6. A method according to any of the preceding claims, characterized in that the phase angle reference is determined in dependence on the difference between the measured direct voltage (UD) and a voltage reference (uπre_f) •

7. Convertor equipment comprising a voltage stiff fundamental frequency commutated convertor (SR) with alternating voltage terminals (AT) for connection to an alternating voltage network (N) via inductive elements (r, XL) _r with direct voltage terminals (DT+, DT-) for connection to a direct voltage source (DP, DN, C) , and with members (MC) for forming the flux vector (ψconv) of the convertor, characterized in that it comprises

voltate measuring members (UM) for determining the direct voltage (UD) of the direct voltage source,

frequency reference forming members (MC) for forming a frequency reference (f_bus) ■

path-determining members (FCC) adapted, based on the measured direct voltage (up) and the frequency reference (^fbus) _r to determine characteristic data (ψhex) of ^an estimated polygon path (HDEK) for a flux reference vector

(ψref) ,

reference vector forming members (FCC) adapted, based on a phase angle reference (arg(ψ_ref)), to form a flux reference vector (ψre_f) following said path,

commutation-controlling members (FCC) adapted to trigger the next commutation at such a time that the flux reference vector and the flux vector of the convertor will coincide at the point of intersection (C) between the path (Ai, GC) of the flux vector and one side (EK) of said polygon path, which side has a direction which deviates from the direction of that polygon side (HD) along which the flux reference vector moves immediately prior to commutation.

8. Convertor equipment according to claim 7, characterized in that the commutation-controlling members are adapted to trigger the next commutation when the estima¬ ted remaining path lengths (BiDEC and AiFGC, respectively) for the flux reference vector and the flux vector of the convertor, respectively, from the time of commutation to the time when these vectors coincide at said point of inter¬ section, correspond.

9. Convertor equipment according to claim 7 or 8, in which the convertor is a six-pulse convertor, characterized in that the path-determining members are adapted to form characteristic data for an estimated polygon path in the form of an equilateral hexagon.

10. Convertor equipment according to any of claims 7-9, characterized in that the path-determining members are adapted to form the estimated polygon path with a side length (ψ_hex) which is^' proportional to the measured direct voltage (up) .

SUBSTITUTE SHEET 1 7

11. Convertor equipment according to any of claims 7-10, characterized in that said frequency reference forming members are adapted to form the frequency reference by measuring the frequency (f_bus) of the voltage in the alternating voltage network (N) .

12. Convertor equipment according to any of claims 7-11, characterized in that it comprises a direct voltage regulator (UR) adapted to form said phase angle reference arg (ψ_ref) in dependence on the difference between the measured direct voltage (UD) and a voltage reference (upre_f) and on the argument (arg(ψ_bus)) of a flux vector corresponding to the network alternating voltage.