• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    Predictive control method and system of a DC/AC converter. A method and a predictive control system of a DC/AC converter (1) for transmission in high voltage direct current are disclosed. The DC/AC converter (1) has an input voltage (VDC) and an output voltage vector {IMAGE-01} whose possible values define a plurality of hexagons. Also, the DC/AC converter (1) is configured to provide a load with a load voltage vector {PICTURE-02} and a load current vector {PICTURE-03}. The method and system comprise calculating an output voltage vector of reference {PICTURE-04}, starting from the grid voltage vector {PICTURE-05}, of the grid current vector {PICTURE-06}, and from the reference values of the active powers {PICTURE-07} and reactive {IMAGE-08}, to minimize a cost function that includes reactive and active power errors. (Machine-translation by Google Translate, not legally binding)
    公开号:ES2733738A1
    申请号:ES201930833
    申请日:2019-09-26
    公开日:2019-12-02
    发明作者:Prieto Dionisio Ramirez；Zarei Mohammad Ebrahim；Rodrigo Fernando Martinez；De Lucas Luis Carlos Herrero
    申请人:Universidad Politecnica de Madrid；Universidad de Valladolid；
    IPC主号:
    专利说明:

    [0001]
    [0002] Predictive control method and system of a DC / AC converter
    [0003]
    [0004] Object of the invention
    [0005]
    [0006] The present invention relates to the field of power electronics, and more specifically to the technical sector dedicated to electronic converters for the conversion between direct current "DC" and alternating current "AC".
    [0007]
    [0008] Background of the invention
    [0009]
    [0010] Continuous AC converters usually employ proportional intergral type (PI) regulators, which are easily implemented but are difficult to adjust and can lead to a slow response. They are responsible for controlling the active and reactive powers, the speed, the torque and the magnetic flux of the motors / generators, and the voltages and currents of the generators.
    [0011]
    [0012] Another solution that emerged about 30 years ago is that of model-based Predictive Control (MPC), which is considered one of the most important advances in control theory. This solution known in the state of the art is based on minimizing a cost function that includes, for example, the error between the reference and the actual value of the active / reactive power or current. The aforementioned state-of-the-art solution can be found in "IEEE Transactions on Industrial Electronics, Four-Switch Three Phase Operation of Grid Side Converter of Doubly Fed Induction Generator with Three Vectors Predictive Direct Power Control Strategy, ME Zarei, C. Veganzones, J. Rodríguez and D. Ramírez ' e “IEEE Transactions on Power Electronics, Three-phase four-switch converter for SPMS Generators based on Model Predictive Current Control for Wave Energy Applications, ME Zarei, D. Ramirez, C. Veganzones, J. Rodriguez. "
    [0013]
    [0014] There are two classes of MPC, Continuous Control Set MPC (CCS-MPC) and Finite Control Set MPC (FCS-MPC). In the first group, a modulator generates the switching states of the electronic converter from a continuous output of the predictive controller. In the second group, the known finite set of possible outputs that can be generated by the electronic converter is used to solve the optimization problem. The present invention fits into this second FSC-MPC group.
    [0015] The FSC-MPC predictive control systems obtain, as a result of their calculations, the application times of the spatial vectors generated by the electronic converter, making it unnecessary to use a Spatial Vector Modulation (SVM) stage.
    [0016]
    [0017] The present invention consists of a modification of the basic method of calculating the voltage reference vector in the alternating current part of the electronic converter, whose objective is to reduce and make the calculation time constant by adding an SVM step.
    [0018]
    [0019] In order to explain the invention, first of all it will be explained what the basic procedure consists of in a specific case where the cost function includes errors in the active and reactive powers. The explanation of the basic procedure is done step by step.
    [0020]
    [0021] 1. Calculate the slopes of the active power (SP1, SP2, SP0) and the reactive power (SQ1, SQ2, SQ0) that the electronic converter exchanges with the system to which it is connected:
    [0022] _ dP g
    [0023] SpJ = dt v j
    [0024] dQ g
    [0025] SQJ dt Vj
    [0026] and which are caused by each spatial vector Vx, V2 and V0 that the electronic converter generates.
    [0027]
    [0028] The vectors Vt, V2 are the two space vectors that define the first sector of the hexagon defined by the 6 non-null space vectors that a two-level converter is capable of generating, Fig. 1, and 0 ^) the null module space vector . These slopes can be calculated using the procedure indicated in "IEEE Transactions on Industrial Electronics, Four-Switch Three Phase Operation of Grid Side Converter of Doubly Fed Induction Generator with Three Vectors Predictive Direct Power Control Strategy, ME Zarei, C. Veganzones, J. Rodríguez and D. Ramírez ”.
    [0029]
    [0030] The electronic converter generates the spatial vectors V¡_,% and % successively over a period, Ts, which begins at the instant k and ends at the instant k + 1.
    [0031]
    [0032] The active and reactive powers at the end of said period Ts (instant k + 1) can be estimated as the sum of the effects that each spatial vector generated by the converter electronic causes:
    [0033]
    [0034]
    [0035]
    [0036] Q g ( k + 1) - Q g ( k) S Q í t V Í + Sq2ÍV2 S Qo t Vo
    [0037]
    [0038] Where S P 1 , S P2 and S P0 and S Q1 , S Q2 and S Q0 are the slopes in the active and reactive power caused by the successive voltage vectors generated by the electronic converter, and t v 1 , t v 2 , t V0 express their times during which the vectors have been applied within the period Ts.
    [0039]
    [0040] . Define a cost function that in this case includes the errors of the active and reactive powers at the end of each period, and where "*" means reference value,
    [0041]
    [0042]
    [0043]
    [0044]
    [0045] Calculate the optimal duration of the application times tV1, tv2, tvo of the vectors ( and [, V2, Vo), from the point of view of minimizing the cost function.
    [0046]
    [0047] If the calculated application times tv1, tv2, tvo are positive, that means that the voltage vector of the reference converter v * is in the first sector. Then, the solution vector that minimizes the cost function can be reconstructed by averaging the vectors % _,% and % using tv1, tv2, tvo and the period Ts using the equation:
    [0048]
    [0049]
    [0050]
    [0051]
    [0052] where Ts is the period of the PWM modulator and goes *, vp * the components in the horizontal and vertical axes of the solution vector.
    [0053]
    [0054] If any of the application times tv1, tv2, tvo calculated is negative, it means that the solution vector v * is not found in the first sector of the hexagon and must be searched in the second sector, where the spatial vectors that define are %,% and %. If the new times of application tv2, tv3, tvo are positive, then the solution vector v * is in the second sector, and if any of the times is negative, it is necessary to continue testing in the following sectors until selecting the correct sector of between the six possible within the hexagon until obtaining positive times (tvj, tvj + i, tvo).
    [0055]
    [0056] 4. Generate the solution vector v * using the calculated times corresponding to the correct sector.
    [0057]
    [0058] This procedure has two drawbacks. The first drawback is that it requires several iterations of calculation until the correct sector is found. The second drawback is that the number of iterations is not always the same. This makes the algorithm slower and has an indeterminate duration.
    [0059]
    [0060] When the converter is two levels, the maximum number of sectors in which the spatial vector solution v * can be found is 6, Fig. 1, but when the number of converter levels increases, the number of sectors increases exponentially, by For example, for five levels there are 96 sectors. This makes the procedure extraordinarily difficult.
    [0061]
    [0062] The present invention improves the procedure described above, since it manages to obtain the application times of the space vectors in a single iteration, resulting in a faster algorithm and of predetermined duration.
    [0063]
    [0064] This invention is applicable to continuous / alternating converters that are used in many different applications, such as control of alternating current motors, network connection of renewable energy sources, control of electric generators, STATCOM, HvDC transmission, etc. Although this invention is applicable to both two-level and multi-level converters, its advantage is all the greater the greater the number of levels.
    [0065]
    [0066] Description of the invention
    [0067]
    [0068] The present invention solves the problems, described above, to find the sector where the spatial vector solution v * is located and thus be able to calculate the application times of the spatial vectors that define said sector. The present invention manages to obtain the application times of the spatial vectors in a single iteration, resulting in a faster algorithm and of predetermined duration. Although this invention is applicable to both two-level and multi-level converters, its advantage is all the greater the greater the number of levels.
    [0069] Therefore, in a first aspect of the invention, a predictive control method of a DC / AC converter is disclosed. The DC / AC converter comprises an input voltage (vdc) and an output voltage vector ( v MMC) with a plurality of possible values, the DC / AC converter being configured to provide a load with a load voltage vector (vp and a current charge vector (ip. The method comprises the following steps:
    [0070] • measure the grid voltage vector (vp and the grid current vector (ig);
    [0071] • calculate the horizontal (vga) and vertical (vgp) components of the grid voltage vector (vp, and the horizontal (iga) and vertical (igp) components of the grid current vector (ip;
    [0072] • calculate the active (Pg) and reactive (Qg) powers delivered to the network, from the horizontal (vga) and vertical (vgp) components of the grid voltage vector (vp, and horizontal (iga) and vertical components (igp) of the current network vector (ip, using the following equations:
    [0073]
    [0074] Pg = 1.5 ■ {Vg a Íg a Vg p i g f i)
    [0075] Qg = 1 .5 • ( vgpiga - vg a ig p)
    [0076]
    [0077] • calculate the slopes of active power (Sp1, S p2, S p0) and reactive (Sq1, S q2, S q0), based on the vertical and horizontal components of the grid current vector (ip and the grid voltage vector (vp, using the following equations:
    [0078]
    [0079]
    [0080]
    [0081] for j = 1, 2, 0;
    [0082] estimate the active and reactive powers at the end of the period (instant k + 1) as the sum of the effects of each vector used in the modulation:
    [0083]
    [0084] Pg ( k + 1) - PgiK) Rpi ^ l Rp2 ^ 2 Rpo ^ o
    [0085] Qgik + 1) - Qg ( k) S qi Í i + SQ2Í2 SQoto
    [0086]
    [0087] where SP1, SP2 and SP0 are the slopes of the active power of the corresponding voltage vectors; SQ1, SQ2 and SQ0 represent the slopes of the reactive power for the selected space vectors; and, t1, t2, t0 express their duration times;
    [0088] • define a cost function that includes the errors of the active and reactive powers at the end of each period, where "*" means reference value, where the cost function comprises the following expression:
    [0089]
    [0090] F ( k + 1) = ( pg ( k + 1) - Pg * ( k) f ( Qg ( k + 1) - Q * g ( k) f
    [0091]
    [0092] • calculate the optimal duration of the application times (t1, t2, te) of the vectors ( and [,%, V q ), to minimize the cost function:
    [0093]
    [0094]
    [0095]
    [0096] tvo = Ts - tV1 - tV2
    [0097] where Ts is the switching period of the generated spatial vectors;
    [0098]
    [0099] • calculate the components on the horizontal ( vmmc < x *) and vertical ( vmmcp *) axes of the output voltage of the reference converter ( v MM c *), from the application times (ti, t2, te) and of the vectors (Vj *, %, V q ,), using the equation:
    [0100]
    [0101]
    [0102]
    [0103]
    [0104] generate configuration parameters so that the output voltage vector of the DC / AC converter ( v MMC) coincides on average with the reference output voltage vector ( vMmc *), applying a multilevel SVM algorithm that selects the appropriate hexagons in each cycle , for which, three-phase components of the output voltage vector ( vMMC), and / or switching times of transistors that make up the DC / AC converter, so that the output voltage ( v MMC) averages the reference vector vMMc *.
    [0105]
    [0106] In an embodiment of the predictive control method of a DC / AC converter, the method additionally comprises, after a predetermined period, repeating the previously defined steps for the first aspect of the invention, for the calculation and application of a new reference output voltage vector ( v M mc *) -
    [0107]
    [0108] In an embodiment of the predictive control method of a DC / AC converter, the method additionally comprises generating configuration parameters is carried out by means of a multilevel spatial vector modulator.
    [0109]
    [0110] In a second aspect of the invention there is presented a predictive control system of a DC / AC converter comprising calculation means and configuration means adapted to determine and apply configuration parameters of a continuous / alternating converter from a voltage of input, a load voltage vector, a load current vector and a plurality of possible values of an output voltage vector.
    [0111]
    [0112] In one embodiment of the second aspect of the invention, the DC / AC converter is a multi-level modular converter, which comprises an input voltage ( v dc) and an output voltage vector ( vMMC) with a plurality of possible values. In this case, the modular multilevel converter is configured to provide a load with a load voltage vector (vp and a load current vector (ip. The predictive control system of a DC / AC converter comprises:
    [0113] • measuring means configured to measure the load voltage vector (vp and the load current vector (ip and generate control parameters; and,
    [0114] • a set of means configured to carry out the steps of the method defined in any one of the embodiments of the first aspect of the invention, based on the control parameters.
    [0115]
    [0116] In one embodiment, the set of means comprises: a first conversion element; a second conversion element; first means of calculation; second means of calculation; and, a multilevel spatial vector modulator.
    [0117]
    [0118] In particular, the media set is configured to:
    [0119] - Calculate the application times of the spatial vectors (SVM) of the first sector, as described in the first aspect of the invention.
    [0120] - Calculate the reference voltage vector, as described in the first aspect of the invention.
    [0121] For its part, the calculation means calculate configuration parameters (for example, the duration of the IGBT trips generated and calculated by the SVM itself) from the reference values (normally active / reactive powers or currents) and apply the configuration parameters to the converter. In addition, preferably, the system further comprises a subset or all of the following elements:
    [0122]
    [0123] - A spatial vector modulator configured to calculate the previously described configuration parameters.
    [0124]
    [0125] Note that any preferred option or particular implementation of the system of the invention can also be applied to the method of the invention. Likewise, the elements of said system can be adapted or configured to implement any step of the method of the invention, according to any particular implementation of both.
    [0126]
    [0127] Finally, in a third aspect of the invention there is a computer program comprising computer program code means adapted to implement the described method, when an application-specific integrated circuit is executed in a digital signal processor, a microprocessor, a microcontroller or any other form of programmable hardware. Note that any preferred option and particular implementation of the device and system of the invention can be applied to the method and computer program of the invention, and vice versa.
    [0128]
    [0129] The method, system and computer program of the invention calculate the voltage value necessary to generate the variables (normally active / reactive powers or currents) desired in each application quickly, reducing the computational load, the amplitude of the generated harmonics, the losses in generators and the mechanical stress produced in the windings. These and other advantages of the invention will be apparent in light of the detailed description thereof.
    [0130]
    [0131] Description of the figures
    [0132]
    [0133] In order to help a better understanding of the features of the invention according to a preferred example of practical realization thereof, and to complement this description, the following figures are attached as an integral part thereof, the character of which is illustrative and non-limiting:
    [0134] Figure 1 shows the hexagon that defines the spatial vectors that a two-level electronic converter is capable of generating in alternating current.
    [0135]
    [0136] Figure 2 shows the hexagon defining the spatial vectors that a two-level electronic converter is capable of generating in alternating current, expressed as a function of only VÍ and v2.
    [0137]
    [0138] Figure 3 shows a schematic of a three-phase MMC, of four switches per branch, known in the state of the art.
    [0139]
    [0140] Figure 4 presents an example of the implementation of an MMC module, in semi-bridge configuration, known in the state of the art.
    [0141]
    [0142] Figure 5 exemplifies a connection of the MMC to a power grid through inductances and resistors, known in the state of the art.
    [0143]
    [0144] Figure 6 exemplifies the possible values of the output voltage vector of a five-level MMC converter. The vectors (VÍ, v2 and V0) that limit the first sector are indicated.
    [0145]
    [0146] Figure 7 shows the possible spatial vectors that a multilevel converter of four switches per branch (which gives rise to 5 voltage levels in the output waveform) can generate within the hexagon and shows the outer hexagon defined exclusively by the vi vectors and %.
    [0147]
    [0148] Figure 8 shows the voltage and current measurement elements used by a preferred embodiment of the present invention, as well as a possible application scenario formed by an MMC and the network to which it is connected.
    [0149]
    [0150] Figure 9 shows the control elements that determine the gate signals of the MMC converter from the measured parameters, in accordance with a preferred embodiment of the invention.
    [0151]
    [0152] Figure 10 presents the joint control and power scheme. Figure 10 shows first the measurement of the voltages and load currents to calculate with them the instantaneous active and reactive powers. Next, along with the vectors y1, y2, the load voltage and frequency are used to obtain the slopes of said powers. TO
    [0153]
    [0154]
    [0155] Then, the slopes are used together with the real and reference values of the powers, to calculate the application times of the vectors ^, V2, y0 and the voltage reference vector, v *. Finally, the solution vector v * is reproduced using a spatial vector modulator that triggers the inverter elements accordingly.
    [0156]
    [0157] Figure 11 describes in a flow chart the process described for Figure 10.
    [0158]
    [0159] Preferred Embodiment of the Invention
    [0160]
    [0161] List of references.
    [0162]
    [0163] 1. - Multilevel modular converter - MMC converter.
    [0164] 2. - Configuration parameters.
    [0165] 3. - First conversion element.
    [0166] 4. - Second conversion element.
    [0167] 5. - First means of calculation.
    [0168] 6. - Second means of calculation.
    [0169] 7. - Control means.
    [0170]
    [0171] In this text, the term "comprises" and its derivations (such as "understanding", etc.) should not be understood in an exclusive sense, that is, these terms should not be construed as excluding the possibility that what is described and defined can include more elements, stages, etc. Note also that vectors and components accompanied by the term "reference" should be understood herein as objective values for said vectors and components, said objectives being calculated by the method and system of the invention. That is, said "reference" vectors and components are not external fixed value references, but rather have variable values that can change in successive iterations of the calculations performed by the invention.
    [0172]
    [0173] The preferred embodiment of the invention is presented in the case where the alternating continuous converter is a modular multilevel converter (MMC), commonly used for transmission of electric power in high voltage direct current (HVDC) . However, the present invention is valid for any three-phase converter, multilevel or not. The equuacones do not change nor does the procedure. Only the space vector modulator (known in the state of the art) and the electronic converter itself changes.
    [0174] In an embodiment of the predictive control method of a continuous / alternating converter, the method comprises:
    [0175]
    [0176] - Measure a load voltage vector and a load intensity vector supplied to an external load.
    [0177]
    [0178] - Calculate the variables involved in the cost function from the load voltage and current using in all cases the spatial vectors that define the first sector, V % and V0, Fig. 1.
    [0179]
    [0180] - Calculate a cost function to minimize the error of the quantities involved in it (normally active / reactive powers or currents).
    [0181]
    [0182] - Always calculate the application times tv i, tv2, tvo of the spatial vectors of the first sector, V , % and V0 that allow to minimize the cost function.
    [0183]
    [0184] If the three times tv i, tv2, tvo are positive, they can be used to average the solution vector v * without the need for a spatial vector modulator (SvM) in the next stage.
    [0185]
    [0186] If any of the three times tv i, tv2, tvo are negative, it is not possible to use them to perform the modulation of the solution vector v * in any case, which would force iterations until only positive values are obtained.
    [0187]
    [0188] However, the advantage of this invention is that, if any of the three times tv1, tv2, tv0 is negative, it is always possible to correctly rebuild, without making a mistake, the solution vector vector v *
    [0189]
    [0190]
    [0191]
    [0192]
    [0193] although some of the times tv i, tv2, tvo is negative. This always allows modulation in an SvM known in the state of the art, without having to perform the iterations described above to find the sector containing the voltage reference vector that minimizes the cost function.
    [0194] The reason is that, as seen in Fig. 1, each sector of the hexagon is defined by two positive module vectors at 600 but, mathematically, those two vectors can always be vi, % if they are allowed to have a negative module, as seen in Fig. 2.
    [0195]
    [0196] Thus, the x, y coordinates of the solution vector v * can always be calculated based on the vectors % _,% for any position within the hexagon:
    [0197]
    [0198] Sector I
    [0199] v x = Vl x - - f V2 x - - f 0 - - f
    [0200]
    [0201]
    [0202]
    [0203]
    [0204] „_ T / Vy = V4y ^ { V ^ 4 + t V j Sy ^ t V ^ 5 - 0 - ^ V 2y - ^ 2 <-
    [0205] Sector V
    [0206]
    [0207] Vx = V 5 x ^ - f V6x ^ - f ■ V1x ^ - Xp V 2 x ^^ 2 <0
    [0208]
    [0209] Vy = V 5y ^ - f V 6 y ^ - f ■ 0 ^ p V 22yy ^ ^ 2 p <-
    [0210]
    [0211]
    [0212]
    [0213]
    [0214] The function of vector V0 is to average the module of the solution vector, but it does not define the hexagon sector. Besides tv0 only influences the calculation of tvj, tvj + i but does not add
    [0215]
    [0216]
    [0217] value to the components "x, y" of the solution vector v * since the vector V0 has a null module.
    [0218]
    [0219] The fact that it is possible to reconstruct the solution vector even though some time is negative, always allows to use VÍ,% and Vq and calculate (ti, t2, to) even though the solution vector v * is actually in another sector of the hexagon different from sector 1. That is, although the solution vector v * is found, for example, in sector III defined by %, í £, it can be calculated using - V , v2, as seen in the Fig. 2.
    [0220]
    [0221] Therefore, in this invention, the application times of the vectors tvj, tvj + i tvo result of the FSC-MPC are not used to perform the modulation of the solution vector v * but to calculate the components of the solution vector v *, which subsequently it is modulated by adding an SVM stage.
    [0222]
    [0223] - Rebuild the reference voltage vector v * using the application times (ti, t2, to) previously calculated and the period Ts.
    [0224]
    [0225] - Calculate the position of the solution vector v * by any system known in the state of the art, to choose 3 spatial vectors with which to reproduce it by modulation. In this way, the linear combination of the 3 spatial vectors of output voltage of the electronic converter averaged during its application times generates an output voltage that equals on average the desired reference output voltage vector.
    [0226]
    [0227] - Perform with the electronic converter the spatial vector modulation that averages the solution vector v *.
    [0228]
    [0229] Figure 3 shows an MMC converter (1) for HVDC known in the state of the art, which serves as an example of application of the method and system of the invention. In particular, it is a three-phase MMC with six branches (R1, R2, R3, R4, R5, R6) and five modules (also called SM, of the English 'Switching-Module') per branch (SM1, SM2, SM3, SM4). The DC area of the converter has an input voltage (vdc), divided into two symmetrical voltages (+ v dc / 2, - v dc / 2). The first branch (R1), second branch (R2) and third branch (R3) are connected to the positive symmetric voltage (+ vdc / 2); while the fourth branch (R4), the fifth branch (R5) and the sixth branch (R6) are connected to the negative symmetric voltage (-vdc / 2). In turn, the branches are connected two by two through pairs of inductances (L), establishing a vector output voltage (vp with three three-phase components ( ^ or > V ob ' V oc ).
    [0230]
    [0231] Figure 4 shows a possible realization of an SM (denoted as SMk for being valid for any of the five sub-modules of each branch: SM1, SM2, SM3, SM4) called semipuente (in English 'half-bridge'). The semipuent topology comprises a first transistor (T1) connected in parallel to a first diode (D1) and a second transistor (T2) connected in parallel to a second diode. The anode of the first diode (D1) is connected to the cathode of the second diode (D2) serving as the entry point of the submodule current ( i sm) and first terminal of the submodule voltage ( v sm). In turn, the cathode of the first diode (D1) is connected to the anode of the second diode (D2) through a capacitor (C) with a capacitor voltage (vc). Said anode of the second diode in turn acts as the second terminal of the submodule voltage ( v sm).
    [0232]
    [0233] Figure 5 exemplifies the connection of the MMC converter (1) to a network (acting as a load) by means of three filter inductances (Lf) and three resistors (Rf). After the three filter inductances (Lc) and the three resistors (Rf), the output voltage vector (tp) becomes the load voltage vector (ip with three-phase component paths ( vga, vgb, v gc). figure 5 shows also reflected the input current (AD), the sinusoidal components (iga 'IGB' igc) of the vector load current (p; voltage upper branch (vupa) corresponding to the first branch (R1); and the lower branch voltage (vlowa) corresponding to the fourth branch (R4).
    [0234]
    [0235] In order to exemplify the vector calculations described, Figure 6 shows the set of values that the output voltage vector can take in a five-level converter. Each possible vector therefore comprises three components, each component being between levels 0 and 4. Each of said possible vectors forms the vertex of a triangle which, in turn, forms a plurality of hexagons. The vectors (V1, V2 and V0) that limit the first sector are indicated. The reference voltage vector (vref) is obtained as a linear combination of vectors V1, V2 and V0.
    [0236]
    [0237] Figure 7 shows how all the vectors within the hexagon of a five-level converter can be obtained by the linear combination of the vectors VÍ and % provided that negative modules are admitted for them.
    [0238]
    [0239]
    [0240] Figure 8 shows the measuring means comprised by a preferred embodiment of the system of the invention. Also shown in Figure 8 are control parameters (2) of the MMC converter (1), calculated from said horizontal component (vga) and vertical component (vgp) of the grid voltage vector (p; and of the horizontal component (iga) and vertical component (igp) of the current network vector ( p .
    [0241]
    [0242] Figure 9 shows, in turn, a preferred embodiment of the elements that calculate the control parameters (2) which in turn execute steps of preferred embodiments of the method and the computer program of the invention. In particular, the system comprises a first conversion element (3) and a second conversion element (4) that transfer the inputs of three-phase axes (abc) to horizontal-vertical axes (ap). In particular, the first conversion element (3) calculates a horizontal component (iga) and a vertical component (igp) of the current network vector ( p from the three three-phase components (iga, igb, igc ) of the current vector of network ( p . The second conversion element (4) calculates a horizontal component (vga) and a vertical component (vgp) of the network voltage vector (p from the three-phase components (vga, vgb, v gc) of the voltage vector load ( p .
    [0243]
    [0244] The vertical and horizontal components of the mains intensity vector (p and of the mains voltage vector ( p are introduced in the first calculation means (5), obtaining active (SPabc) and reactive (SQabc) power slopes, and values of active power (P) and reactive (Q). The slopes of active (SPabc) and reactive (SQabc) power, active (P) and reactive (Q) power values, and active power (P *) references. and reactive (Q *), are introduced in a few seconds calculation means (6), obtaining horizontal (vMMCa *) and vertical ( v mmcp *) components of the reference output voltage vector (i; MMC *). Calculation means (6) are preferably a predictive controller Finally control means (7), preferably formed by a multilevel space vector modulator (SVM), calculates the control parameters (2) from the reference output voltage vector (vMMC *).
    [0245]
    [0246] Figure 10 shows the joint voltage and control scheme. The measurement of the network voltages and currents, which are transformed into the components of the mains voltage (vgap) and mains current (igap) vectors, which are used for the calculation of active powers (Pg) is observed. ) and reactive (Qg). Subsequently, the slopes of the active (SP1, SP2, SP0) and reactive (SQ1, SQ2, SQ0) powers are calculated. Next, the components (vMMCa *, v mmcp *) of the reference output voltage vector are calculated. SVM modulator
    [0247]
    [0248]
    [0249] find the triangle of Fig. 7 in which the output voltage vector is located, and generate the application times of the three vectors that delimit said triangle. Finally, the power part is formed by the converter, the inductances and coupling resistors, and the network.
    [0250]
    [0251] The steps of a preferred embodiment of the method of the invention are detailed below, which are in turn implemented by a preferred embodiment of the first and second calculation means (5, 6) of the system. The method comprises:
    [0252]
    [0253] 1. Calculate the horizontal (vga) and vertical (vgp) components of the grid voltage vector (ip, and the horizontal (iga) and vertical (igp) components of the grid current vector (Q, from the voltages (vga, vgb, vgc) and network currents (iga, igb, igc).
    [0254]
    [0255] two . Calculate the active (Pg) and reactive (Qg) powers delivered to the network, from the horizontal (vga) and vertical (vgP) components of the grid voltage vector (ip, and the horizontal (iga) and vertical ( igp) of the current network vector (p, using the following equations:
    [0256]
    [0257] Pg = 15 • ( Vga Íga V g ^)
    [0258] Qg = L 5 • ( V gpiga - vg a ig p)
    [0259]
    [0260] 3. Calculate the slopes of active power (Sp1, S p2, S po) and reactive (Sq1, S q2, S q0), from the vertical and horizontal components of the network current vector (p of the network voltage vector (p, using the following equations:
    [0261]
    [0262] c _ dP9 1.5
    [0263] SpJ ~~ dF j
    [0264] Vj LZ
    [0265]
    [0266]
    [0267]
    [0268] for j = 1, 2, 0.
    [0269] These equations have been obtained from the theory known in the state of the art that allows calculating voltages, currents and powers expressed as spatial vectors:
    [0270]
    [0271]
    [0272] To calculate the slopes of the active and reactive powers, the derivatives of the voltage and the current in the horizontal (a) and vertical (P) axes must be calculated. The tension of a balanced network can be represented as:
    [0273] Vga = Vg Sin a st
    [0274] Vg¡3 = -Vg COS Mst
    [0275] where ws is the frequency of the network.
    [0276] The derivative of the grid voltage in the horizontal (a) and vertical (P) axes is:
    [0277] dVga
    [0278] —— = WSVg COS Wst = - teSVgp
    [0279] d-VgQ
    [0280] = wsVg sin a st = wsvga
    [0281]
    [0282] Taking into account the equation of the coupling of the MMC converter to the network,
    [0283] ____> ^ dlg ^
    [0284] VMMC = vg ^ f l3
    [0285] the derivative of the network current can be expressed as
    [0286]
    [0287]
    [0288]
    [0289] Then, those derived from the active and reactive powers can be estimated by:
    [0290]
    [0291]
    [0292]
    [0293]
    [0294] Estimate the active and reactive powers at the end of the period (instant k + 1) as the sum of the effects of each vector used in the modulation:
    [0295]
    [0296] P g ( k 1) = P g ( k) S P í t í + S P 2t2 S P 0 t 0
    [0297] Q g ( k + 1) = Q g ( k) S Q í t í + S Q 2t2 S Q 0 t 0
    [0298]
    [0299] Where S P 1 , S P 2 and S P0 are the slopes of the active power of the corresponding voltage vectors, S Q1 , S Q 2 and S Q0 represent the slopes of the reactive power for the selected space vectors, and t 1 , t2, t 0 express their duration times.
    [0300]
    [0301]
    [0302] 5. Define a cost function that includes the errors of the active and reactive powers at the end of each period, where * means reference value,
    [0303]
    [0304] F ( k + 1) = ( pg ( k + 1) - Pg * ( k) f ( Qg ( k + 1) - Q * g ( k) f
    [0305]
    [0306] 6. Calculate the optimal duration of the application times (t1, t2, te) of the vectors ( and [,%, V0, to minimize the cost function:
    [0307]
    [0308]
    [0309]
    [0310] tvo = Ts - tV1 - tV2
    [0311] where Ts is the switching period of the generated spatial vectors.
    [0312]
    [0313] Calculate the components on the horizontal ( vmmc «*) and vertical ( vmmcp *) axes of the output voltage of the reference converter ( vMmc *), from the application times (ti, t2, te) and vectors (^, %, Vq,), using the equation:
    [0314]
    [0315]
    [0316]
    [0317]
    [0318] Generate the configuration parameters (2) so that the converter output voltage vector ( v MMC) coincides on average with the reference output voltage vector ( vmmc *) ■ In particular, a multilevel SVM algorithm chooses the hexagons in each cycle suitable, that is, the configuration parameters (2) of the three-phase components of the output voltage vector ( vMMC), and / or the switching times of the transistors that make up the MMC converter modules (1), so that the voltage Output averages the reference vector.
    [0319]
    [0320] After a cycle, the steps described for the calculation and application are repeated
    [0321]
    [0322]
    [0323] of a new vecto r reference output voltage ( i MM¿ *).
    [0324]
    [0325] The person skilled in the art will be able to tend that the invention has been described according to some preferable embodiments of the same, but that more variations may be introduced in said preferred embodiments, without leaving the object of the invention as and has been claimed.
    [0326]
    [0327]
    two
    权利要求:
    Claims (6)
    [1]
    1.- Predictive control method of a DC / AC converter, where the converter comprises an input voltage (vdc) and an output voltage vector ( vMMC) with a plurality of possible values, and the DC / AC converter is configured to providing a load with a load voltage vector (eg) and a load current vector (ig); where the method is characterized by comprising:
    • measure the network voltage vector (eg) and the network current vector (ig);
    • calculate the horizontal (vga) and vertical (vgp) components of the grid voltage vector (vg), and the horizontal (iga) and vertical (igp) components of the grid current vector (ig);
    • calculate the active (Pg) and reactive (Qg) powers delivered to the network, from the horizontal (vga) and vertical (vgP) components of the grid voltage vector (vg), and of the horizontal (iga) components and vertical (igp) of the current network vector (Ig), using the following equations:
    Pg = 1. 5 - ( Vg to Íg to Vg pigfi)
    Qg = L 5 • ( vgpiga - vg a ig p)
    • calculate the slopes of active power (Sp1, S p2, S p0) and reactive (Sq1, S q2, S q0), from the vertical and horizontal components of the network current vector (Tg) and the voltage vector of network (eg), using the following equations:

    [2]
    2. - Predictive control method of a DC / AC converter, according to claim 1, characterized in that it additionally comprises, after a predetermined period, repeating the steps defined in claim 1 for the calculation and application of a new voltage vector of reference output (i7MMC *).
    [3]
    3. - Predictive control method of a DC / AC converter, according to claim 1, characterized in that the step of generating configuration parameters is carried out by means of a multilevel spatial vector modulator.
    [4]
    4. - Predictive control system of a DC / AC converter, where the DC / AC converter is a multi-level modular converter (1), which comprises an input voltage ( v dc) and an output voltage vector (■ v MMC ) with a plurality of possible values, and the multilevel modular converter (1) being configured to provide a load with a load voltage vector (vp and a load current vector (ip; characterized in that it additionally comprises:
    • measuring means configured to measure the load voltage vector (vp and the load current vector (ip and generate control parameters (2); and,
    • a set of means (3,4,5,6,7) configured to carry out the steps of the method defined in any one of claims 1 to 3 from the control parameters (2).
    [5]
    5. - Predictive control system of a DC / AC converter according to claim 4, characterized in that the set of media (3,4,5,6,7) comprises
    • a first conversion element (3);
    • a second conversion element (4);
    • first means of calculation (5);
    • second means of calculation (6); Y,
    • a multilevel spatial vector modulator (7).
    [6]
    6. - Computer program comprising instructions which, when the program is executed on a computer, causes the computer to carry out the method of any one of claims 1 to 3.
    two
    类似技术:
    公开号 | 公开日 | 专利标题
    Kwak et al.2014|Predictive-control-based direct power control with an adaptive parameter identification technique for improved AFE performance
    ES2701706T3|2019-02-25|Controller of a modular power converter with distributed electric energy accumulators
    Gupta et al.2008|Switching characterization of cascaded multilevel-inverter-controlled systems
    ES2643896T3|2017-11-27|Production of an active power balance of the phase modules of a converter
    US10516328B2|2019-12-24|Controlling a three-phase electrical converter
    BRPI0710957A2|2012-02-14|matrix converter and control method for matrix converter
    US10554142B2|2020-02-04|Modulation of AC/AC MMC
    JP2015015885A|2015-01-22|Apparatus for modifying voltage command for detecting output current in inverter
    Lupon et al.2014|FPGA implementation of a PWM for a three-phase DC–AC multilevel active-clamped converter
    US20170179844A1|2017-06-22|Electronic Power Converter and Computer Program
    ES2733738A1|2019-12-02|METHOD AND SYSTEM FOR PREDICTIVE CONTROL OF A DC/AC CONVERTER |
    Christe et al.2017|Virtual submodule concept for fast semi-numerical modular multilevel converter loss estimation
    Dahidah et al.2005|A hybrid genetic algorithm for selective harmonic elimination control of a multilevel inverter with non-equal dc sources
    Fuchs et al.2018|Analysis and reduction of the output voltage error of PWM for modular multilevel converters
    CN107017793B|2019-09-13|A kind of space vector modulating method and system of three-phase tri-level inverter circuit
    RU2522036C2|2014-07-10|Method for control of three-phase voltage inverter with current stabilisation at transfer to overload mode
    Arulmurugan2019|Design and testing of stability improvement of nine multi-level H-inverter for distribution system
    ES2600757B1|2017-12-27|Multilevel converter with adaptive voltage
    Teymouri et al.2020|Sensor malfunction detection and mitigation strategy for a multilevel photovoltaic converter
    Kolmakov et al.2015|Application of hysteresis voltage control for three-level neutral point clamped voltage source inverter
    CN103984826A|2014-08-13|SVPWM | module generation system and method
    Zinoviev et al.2013|Unified analysis technique for energy quality factors estimation of NPC multilevel VSC for energy storage systems
    Lee2015|Advanced static over-modulation scheme using offset voltages injection for simple implementation and less harmonics
    Karthik et al.2014|Harmonic elimination in cascade multilevel inverters using Firefly algorithm
    ES2616274A1|2017-06-12|Method and control system of a multilevel modular converter of high voltage direct current |
    同族专利:
    公开号 | 公开日
    ES2733738B2|2020-09-18|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    WO2021183077A1|2020-03-13|2021-09-16|Yildiz Teknik Universitesi|A new space vector modulation method for multi-level converters|
    法律状态:
    2019-12-02| BA2A| Patent application published|Ref document number: 2733738 Country of ref document: ES Kind code of ref document: A1 Effective date: 20191202 |
    2020-09-18| FG2A| Definitive protection|Ref document number: 2733738 Country of ref document: ES Kind code of ref document: B2 Effective date: 20200918 |
    优先权:
    申请号 | 申请日 | 专利标题
    ES201930833A|ES2733738B2|2019-09-26|2019-09-26|METHOD AND PREDICTIVE CONTROL SYSTEM OF A DC / AC CONVERTER|ES201930833A| ES2733738B2|2019-09-26|2019-09-26|METHOD AND PREDICTIVE CONTROL SYSTEM OF A DC / AC CONVERTER|
    [返回顶部]