Modular multi-level electric power converter

Abstract

Disclosed is an MMC converter including a three-phase iron core reactor in which three-terminal leg circuits, in which two two-terminal arm power circuits are connected in series, are provided in u, v, and w phases, positive-side and negative-side terminals of the three-terminal leg circuits are each connected in a star shape and thereby connected to a DC power supply, an intermediate terminal is connected to an AC power supply, and a circulating current due to voltage imbalance between the three legs is suppressed, wherein, in order to suppress AC inductance imbalance between the two poles and between the three phases while maintaining the linearity of inductance by avoiding magnetic saturation of each part of the iron core when an AC system accident occurs, the cross-sectional area of a yoke and side legs serving as a return path of DC magnetic flux of each phase of the three-phase iron core reactor is made larger than that of a main leg (a winding iron core with a non-magnetic gap) of each phase, two sets of windings composed of two concentrically arranged coils wound on the main leg of each phase are coaxially arranged on the main leg, and the winding direction of each of the coils and the connection between the coils are optimally selected. Furthermore, auxiliary yokes crossing the central parts of the main legs of the respective phases or side legs longitudinally crossing between the main legs of the respective phases are provided and used as paths for magnetic flux formed by an AC current.

Description

Modular Multilevel Power Converter

[0001] The present invention relates to a modular multilevel power converter (hereinafter referred to as "MMC converter" in the present invention). In particular, it relates to a modular multilevel power converter suitable for configuring a frequency conversion device by connecting the DC sides of two MMC converters in a back-to-back manner.

[0002] The circuit of the MMC converter consists of three three-terminal leg circuits provided for each phase. The three-terminal leg circuit has a positive terminal, a negative terminal, and an intermediate terminal. The three positive terminals are star-connected to the positive voltage terminal of the DC power supply, the three negative terminals are star-connected to the negative voltage terminal of the DC power supply, and the three intermediate terminals are connected to the three-phase AC voltage terminals.

[0003] The three-terminal leg circuit connects a two-terminal positive arm power circuit, a three-terminal inductive element, and a two-terminal negative arm power circuit in series in order from the positive terminal and connects them to the negative terminal, and connects the intermediate terminal of the three-terminal inductive element to the intermediate terminal of the three-terminal leg circuit.

[0004] The three-terminal inductive element has a function of suppressing the circulating current flowing between the three-terminal leg circuits when the voltage ratio between the three three-terminal leg circuits is unbalanced.

[0005] The positive and negative arm power circuits connect K unit converters that generate the required voltage by controlling the modulation ratio of a PWM converter having an energy storage element with voltage source characteristics such as a capacitor as a voltage source in series.

[0006] The control device of the MMC converter requires capacitor voltage control to keep the 6×K capacitor voltages provided in the unit converter within a predetermined range. This is different from the control devices of conventional two-level or three-level converters.

[0007] The control unit of the MMC converter allows for the separation of AC current control and DC current control. In conventional two-level or three-level converters, the current path from the DC power supply terminal to the AC power supply terminal always flows into and out of the AC power supply terminal. In contrast, with the MMC converter, AC current flows in and out only between the AC power supply terminals, and DC current flows out only between the DC power supply terminals. As a result, the control of AC current and DC current can be separated. Consequently, the active power on the AC and DC sides can be controlled independently. The difference between the active power on the AC side and the active power on the DC side becomes the charging and discharging power of 6 × K capacitors, and acts on the average voltage of 6 × K capacitors.

[0008] The functions required for capacitor voltage control specific to MMC converters can be divided into three functions: a function to maintain balance between the K capacitor voltages by mutually adjusting the modulation rates of the K PWM converters provided for each unit converter within the same arm power circuit (hereinafter referred to as "intra-arm balance control" in this invention); a function to maintain balance between the three phases of the 2 × K unit converter capacitor average voltages constituting the three-terminal leg circuit of each phase (hereinafter referred to as "inter-leg balance control" in this invention); and a function to maintain balance between the K capacitor average voltages of the positive arm power circuit and the K capacitor average voltages of the negative arm power circuit (hereinafter referred to as "inter-positive / negative balance control" in this invention).

[0009] When a capacitor is used as the energy storage element in a unit converter, the capacitor voltage fluctuates due to the charging and discharging cycle determined by the AC frequency.

[0010] Non-patent document 1 discloses a circuit configuration that aims to reduce the size of the equipment while suppressing capacitor voltage fluctuations due to charging and discharging by connecting a DC power supply isolated by a high-frequency transformer in parallel to each capacitor of the unit converter.

[0011] Patent Document 1 discloses a basic circuit configuration in which inductive elements such as reactors are provided between the positive terminal and the intermediate terminal and between the negative terminal and the intermediate terminal of a three-terminal leg circuit of an MMC converter in order to suppress the through-current between the positive terminal and the negative terminal of the three-terminal leg circuit.

[0012] Patent Document 2 discloses a method for miniaturizing the inductive element of Patent Document 1, in which a winding consisting of a coil connected to the positive arm power circuit and a coil connected to the negative arm power circuit is provided for each of the three phases, the two coils are magnetically coupled by winding the first to third winding iron cores concentrically for each of the three phases, and a three-phase five-legged reactor is provided in the MMC converter that recirculates the zero-phase magnetic flux created by the magnetic flux flowing through the first to third winding iron cores at the first and second side legs.

[0013] Patent Document 3 discloses a three-phase five-legged reactor that has a structure that allows for the manufacture and assembly of an iron core and coil, and achieves miniaturization and weight reduction by integrating three AC reactors, one for each of the three phases, while considering countermeasures against localized abnormal iron loss, induction heating, and electromagnetic vibration, and is suitable for suppressing inter-phase imbalance in inductance values.

[0014] Patent Document 4 discloses a method for controlling the balance within an arm to maintain the balance between the capacitor voltages of the unit converters that constitute an MMC converter.

[0015] Patent Document 5 discloses a basic hierarchical configuration of a control system consisting of a PWM modulator and converter current control provided for each unit converter of an MMC converter. Specifically, it discloses a first layer provided for each unit converter, a second layer that adjusts the current of each part of the MMC converter, and a third layer consisting of a higher-level control device that outputs commands to the second layer in order to adjust the active power and reactive power output of the MMC converter. It also discloses a method for adding a second harmonic circulating current command to a fundamental wave current command.

[0016] Patent Document 6 discloses a higher-level control device for two MMC converters with their DC sides connected behind each other as frequency converters, and a higher-level control device particularly suitable for suppressing DC overcurrent and continuing operation in the event of a ground fault on the AC power supply side.

[0017] Patent Document 7 discloses a positive-negative balance control configuration for an MMC converter, particularly suitable for continuing operation in the event of a ground fault on the AC power supply side.

[0018] Patent Document 8 discloses a control method for a unit converter suitable for preventing damage to the power semiconductors constituting the MMC converter. Specifically, when the current of the unit converter exceeds a threshold, it switches to the diode side and stops switching until it returns to below the threshold. To suppress the rise in the capacitor voltage of the unit converter that occurs when the switching stops, the command value of the active power component of the AC current is adjusted by comparing the average capacitor voltage value with a set value. The capacitor imbalance between the positive and negative sides caused by the switching stops is controlled by a circulating current. The output that controls the capacitor imbalance between the three phases caused by the switching stops is biased to the voltage command of the current control output and feedforward controlled. Based on the above, a configuration of an MMC converter suitable for continuing operation during grid fault propagation is disclosed, while reducing the maximum current of the self-extinguishing semiconductor power element of the unit converter by 40% and the maximum current of the antiparallel diode element by 15%.

[0019] Patent Document 9 discloses a method for realizing a variable speed device for a synchronous generator motor using two MMC converters connected behind the DC side. This method allows switching between variable speed operation via the MMC converters at low output and constant speed operation directly connected to the AC system near rated output, enabling both system efficiency and cost reduction through reduced MMC converter rated capacity.

[0020] Japanese Patent Publication No. 5189105 Japanese Patent Publication No. 7312332 European Patent Application Publication No. 0117460 Specification Japanese Patent Publication No. 4999930 Japanese Patent Publication No. 5197623 Japanese Patent Publication No. 7360559 International Publication No. 2024 / 150437 International Publication No. 2024 / 154308 Japanese Patent Publication No. 6995991

[0021] Mariam Saeed et al., "Design and construction of a DAB using SiC MOSFETs with an isolation of 24kV for PET applications", EPE 2017 ECCE Europe.

[0022] MMC converters are classified as voltage-type power converters that use self-extinguishing power semiconductor elements. Compared to conventional voltage-type power converters such as two-level and three-level converters, MMC converters have the advantage of halving power loss and eliminating the need for harmonic filters. On the other hand, they have the disadvantage of being larger due to the six air-core reactors housed in a magnetic shielding chamber.

[0023] In the case of voltage-type power converters, it is considered difficult to economically ensure sufficient overcurrent tolerance compared to power equipment such as rotating electrical machines and transformers that do not use power semiconductor elements, as well as externally excited power converters that use thyristor elements that do not have a self-extinguishing function.

[0024] The function of suppressing overcurrent in an MMC converter can be achieved by replacing the two two-terminal inductors with a single three-terminal inductor and a current control device. However, the fact that the current control device itself is configured on the premise that "the inductance value of the three-terminal inductor is constant and independent of the current" poses a practical obstacle to its implementation.

[0025] In the following explanation, for convenience, the inductance of the first two-terminal inductor will be denoted as L1. The inductance of the second two-terminal inductor will be denoted as L2.

[0026] Furthermore, let L1 and L2 be the self-inductances of the alternative three-terminal inductor, and M be the mutual inductance.

[0027] While magnetically coupling two coils via an iron core is indeed effective for miniaturizing three-terminal inductors, it has the drawback of a sharp decrease in inductance due to magnetic saturation of the iron core.

[0028] Therefore, in order to miniaturize the MMC converter while ensuring overcurrent withstand capability and allowing it to continue operating during AC system fault propagation, the most important challenge is to maintain the inductance value of the three-terminal inductive element at a predetermined value.

[0029] The following explains the challenges that depend on the current control device's functionality being dependent on "the inductance value of the three-terminal inductor being constant."

[0030] Figure 8 shows the circuit diagram of the MMC converter. Here, we ignore the branching circuits for connecting the capacitors inside the unit converter that make up the positive and negative arm power circuits. As a result, the MMC converter circuit consists of 6 nodes (P, N, u, v, w, z) and 10 branches, and the circuit degrees of freedom are 3. Two of the 3 degrees of freedom can be assigned to AC current control, and the remaining degree of freedom can be assigned to DC current control.

[0031] Considering the capacitor circuit, and assuming K is the number of unit converters in series in the arm power circuit, the circuit degrees of freedom of the MMC converter can be considered as (3 + 6 × K). When configuring the control device for the MMC converter, there is an option to consider the circuit degrees of freedom to range from a minimum of 3 to a maximum of (3 + 6 × K).

[0032] The MMC converter disclosed in Non-Patent Document 1 is an example of controlling the circuit degrees of freedom as (3 + 6 × K).

[0033] Figure 9 shows a circuit diagram where the MMC converter is considered to have 5 degrees of freedom and is divided into three current elements (1 DC current, 2 AC currents, and 2 circulating currents).

[0034] Hereinafter, in this invention, the term "alternating current" will be used to refer specifically to current passing through an AC terminal.

[0035] Since the three current elements have different frequencies, it may be possible to control the three current elements independently by adding the voltages calculated by the ARM power converter for each of the three current elements.

[0036] As mentioned above, the fundamental premise for independently controlling the three current elements is that "the inductance value of the three-terminal inductor is constant and it can be considered a linear circuit that does not depend on the current."

[0037] In the following, the configuration of the control device for the MMC converter is shown as an example of the case in which the circuit degrees of freedom of the MMC converter are considered to be 5.

[0038] Configurations of control systems based on a circuit with 5 degrees of freedom are disclosed in Patent Documents 4, 5, 6, 7, and 8.

[0039] In addition, a technique is also disclosed in which, as a circulating current, a quadruple-frequency current is controlled in addition to the double-frequency of an AC power supply, and the circuit freedom is regarded as 7. However, the present invention is not limited by the configuration of such a control system.

[0040] Hereinafter, in the present invention, the control configuration of the MMC converter will be described by citing Patent Document 8.

[0041] When the MMC converter shown in FIG. 9 can be regarded as a linear circuit, the three current elements (DC current, AC current, circulating current) can be regarded as linearly independent circuits because their command frequencies are different. By adding the voltages of the arm power converters calculated for each of the three current elements, the three current elements can be controlled independently. However, the voltage command to the arm power converter needs to be greater than 0 and less than 1 / 2 of the sum of the K capacitor voltages.

[0042] As described above, the major premise of the above control is the linearity of the circuit. Linearity is ensured by "the inductance value of the three-terminal inductive element being constant and independent of current" and "the average value of the K capacitor voltages constituting the arm power converter being regarded as constant".

[0043] There is a method of "installing six air-core coils as inductive elements without magnetic coupling to each other" as a method of simply and surely satisfying the above major premise. According to this method, there is an effect of preventing a rapid decrease in the inductance value due to magnetic saturation of the iron core.

[0044] On the other hand, in order to prevent magnetic saturation, there is a problem that an increase in the size of the magnetic shield chamber for shielding the leakage magnetic flux of the air-core coil and an increase in weight due to an increase in the total plate thickness of the shield steel plate are inevitable. Also, when two air-core coils connected to the same three-terminal leg circuit are installed at the positive-side DC voltage terminal and the negative-side DC voltage terminal, respectively, to avoid magnetic coupling, the withstand voltage of the coil needs to be increased, which further causes a problem of an increase in size and weight.

[0045] In order to solve the above problems, Patent Document 2 discloses a method of further reducing the size and weight using a three-phase five-leg reactor.

[0046] However, when applying the three-phase five-leg reactor of Patent Document 2 to an MMC converter equipped with the method for reducing the current capacity of a semiconductor device for power use disclosed in Patent Document 8, it is necessary to avoid magnetic saturation even for overcurrents exceeding three times the rated current during a system accident spread. For such an MMC converter, the three-phase five-leg reactor of Patent Document 2 has a first problem that "to continue operation during a system accident spread, the avoidance of magnetic saturation considering the point where the DC magnetic flux and the AC magnetic flux overlap in the yoke and the side legs is insufficient".

[0047] Further, when applying the three-phase five-leg reactor of Patent Document 2 to the variable-speed power generation motor device of Patent Document 9, it is necessary to transfer the current from the variable-speed operation by the MMC converter to the constant-speed operation by the AC system bypass connection.

[0048] The inductance for suppressing the current change during current transfer becomes only the inductance with respect to the AC current of the three-terminal inductive element.

[0049] In the case of the three-phase five-leg reactor of Patent Document 2, since the inductance of the inner coil is about 1 / 100 of the DC inductance, it becomes difficult to suppress the overcurrent during current transfer. In order to suppress this overcurrent, a three-phase AC reactor must be added between the unit transformer and the MMC converter, so there is a second problem that the merit of reducing the size and weight by the three-phase five-leg reactor of Patent Document 2 is offset.

[0050] Here, the inductance value with respect to the AC current becomes (L1 - M) and (L2 - M). The inductance with respect to the DC current becomes (L1 + L2 + 2×M).

[0051] Further, in Patent Document 2, it is disclosed that non-theoretical harmonics are generated due to the inductance imbalance between the positive and negative coils arranged concentrically for each three-phase, but it can be suppressed by current control.

[0052] However, in order to meet the strict IEEE standards for even harmonics, it is necessary to suppress not only the inductance imbalance between the positive and negative coils but also the inter-phase imbalance of the AC inductance disclosed in Patent Document 3.

[0053] Here, the inductance imbalance between the positive and negative coils refers to the imbalance between (L1-M) and (L2-M).

[0054] In this invention, in order to maintain a constant inductance value for DC current, gaps filled with multiple non-magnetic materials are installed in the first to third winding cores, while gaps filled with non-magnetic materials are not installed in the side legs and yoke. As a result, this configuration has the effect of suppressing inter-phase imbalance of AC inductance.

[0055] If the above suppression effect is insufficient, the impedance of the unit transformer installed between the AC system and the MMC converter must be increased, which negates the miniaturization and weight reduction benefits of the three-phase five-legged reactor described in Patent Document 2, presenting a third problem.

[0056] Here, the AC inductance of each phase is the sum of the AC inductances of the positive and negative coils (L1 + L2 - 2 × M). Phase imbalance refers to an imbalance in the inductance values ​​of each phase (L1 + L2 - 2 × M).

[0057] The objective of the present invention is to solve the above problems, ensure the low loss that is an advantage of MMC converters, and mitigate the disadvantages of MMC converters by achieving miniaturization of the device and improved operational continuity performance during grid fault propagation.

[0058] To solve the first problem described above and achieve the objective, contrary to the relationship of the core cross-sectional areas of the three-phase five-legged reactor disclosed in Patent Document 3, the cross-sectional areas of the first and second side legs are made larger than the cross-sectional area of ​​the first to third winding cores, which are wound with two coils arranged concentrically, and the cross-sectional area of ​​the yoke connecting the first to third winding cores and the first and second side legs is made larger than the cross-sectional area of ​​the first to third winding cores.

[0059] To solve the second problem described above and achieve the objective, the objective is achieved by installing a second winding coaxially with the first winding on the first to third winding cores of the three-phase five-legged reactor disclosed in Patent Document 2, connecting the inner coil of the first winding with the outer coil of the second winding, and connecting the outer coil of the first winding with the inner coil of the second winding, thereby eliminating the imbalance in AC inductance between the positive and negative sides of each phase.

[0060] To solve the third problem described above and achieve the objective, an auxiliary yoke is provided that starts from the midpoint between the first and second windings of the three-phase five-legged reactor equipped with the second problem-solving means described above, and is perpendicular to the first to third winding cores, connecting the first to third winding cores and the first and third winding cores with the side legs, thereby eliminating the imbalance in AC inductance between the three phases and achieving the objective.

[0061] Furthermore, in order to solve the third problem mentioned above and achieve the objective, a third side leg is provided between the first and second winding cores, and a fourth side leg is provided between the second and third winding cores, thereby eliminating the imbalance in AC inductance between the three phases and achieving the objective.

[0062] The following describes the means by which these functions are implemented.

[0063] In the column [A-1] of Figure 10, a three-terminal inductive element (22x) in which x is represented by three phases (u, v, w) is shown, consists of a first coil (22Px) and a second coil (22Nx) that are magnetically coupled, and when current control is performed by measuring the current I_xp of the first coil (22Px) that flows out to the positive side power converter via the first terminal (xp) and the current I_xn of the second coil (22Nx) that flows out to the negative side power converter via the second terminal (xn), the first terminal ( The magnetomotive force due to the through-current I_cx passing between xp) and the second terminal (xn) excites the wound iron cores of the first coil (22Px) and the second coil (22Nx), and the magnetomotive force due to the alternating current I_x flowing in from the intermediate terminal (xc) and branching out to the first terminal (xp) and the second terminal (xn) cancels out in the first coil (22Px) and the second coil (22Nx), thus representing a configuration of a three-terminal inductive element (22x) suitable for an MMC converter.

[0064] In column [A-2] of Figure 10, a configuration is shown in which the positive first coil (22Px) and the negative second coil (22Nx) are arranged coaxially and magnetically coupled via an iron core, in order to miniaturize the three-terminal inductive element (22x).

[0065] In column [A-3] of Figure 10, a configuration is shown in which the positive first coil (22Px) and the negative second coil (22Nx) are arranged concentrically and magnetically coupled via an iron core, in order to miniaturize the three-terminal inductive element (22x).

[0066] In column [B-1] of Figure 10, the steady-state waveforms of the positive first coil current I_xp and the negative second coil current I_xn are shown for one period of the AC current I_x.

[0067] Here, the positive first coil current I_xp reaches its maximum absolute value at time tp, and the negative second coil current I_xn reaches its maximum absolute value at time tn.

[0068] Column [B-2] in Figure 10 shows the magnetic flux distribution at time tp when the positive side first coil current I_xp is at its maximum in the coaxial configuration. Similarly, column [C-2] shows the magnetic flux distribution at time tn when the negative side second coil current I_xn is at its maximum.

[0069] In the coaxial configuration shown in column [A-2] of Figure 10, the magnetic flux due to the alternating current (I_x) flows horizontally from the core to the side legs between the two coils. As a result, the magnetic flux in the core portion wound around the positive first coil (22Px) is maximized by the magnetomotive force when the positive first coil current I_xp is at its maximum, and the magnetic flux in the core portion wound around the negative second coil (22Nx) is maximized by the magnetomotive force when the negative first coil current I_xn is at its maximum. The maximum magnetic flux in the core portion wound around the two coils is affected by the peak value of the alternating current I_x.

[0070] Column [B-3] in Figure 10 shows the magnetic flux distribution at time tp when the positive side first coil current I_xp is at its maximum in the concentric configuration. Similarly, column [C-3] shows the magnetic flux distribution at time tn when the negative side second coil current I_xn is at its maximum.

[0071] In the concentric configuration shown in column [A-3] of Figure 10, the magnetomotive force due to the alternating current (I_x) flowing through the two coils acts to cause a magnetic flux to flow vertically through the air gap on the cylinder surrounded by the two coils.

[0072] The return path of the magnetic flux branches into the iron core wound around the coil and the side legs. The former includes a gap filled with a non-magnetic material, while the latter does not. By setting the magnetic resistance of the former to be significantly greater than that of the latter, the majority of the magnetic flux can be directed to the side legs.

[0073] In column [B-3] of Figure 10, the AC and DC magnetic fluxes flowing through the side legs are in opposite directions, so the side leg flux appears small and the magnetic flux in the core appears to increase. However, the magnetic flux in the core can be considered the same in columns [B-3] and [C-3]. The magnetic flux in the core wound around the two coils is dominated by the through-current I_cx and is hardly affected by the peak value of the AC current I_x.

[0074] Patent Document 2 focuses on the fact that, in the event of an AC power system fault propagation, the ratio of the peak value of the AC current (I_x) of the MMC converter to its rated value is greater than the ratio of the peak value of the through-current (I_cx) to its rated value, and discloses a three-phase five-legged reactor for an MMC converter in which two coils are concentrically arranged for every three phases.

[0075] However, the DC overcurrent suppression function of the higher-level control device disclosed in Patent Document 6 makes it possible to further suppress the peak multiplier during AC system fault propagation. As a result, it is now possible to increase the peak value of the AC current while maintaining the through-current peak value that is constrained by the magnetic saturation of the winding core.

[0076] Furthermore, the control method for the unit converter disclosed in Patent Document 8 reduces the peak current value of the self-extinguishing semiconductor element that determines the size of the positive and negative arm power converters by 45%, and reduces the peak value of the arm current (I_xp, I_xn) by 15%. This allows for a larger current rating of the antiparallel diode, which has a lower cost per current rating compared to the self-extinguishing semiconductor element, and thus increases the peak value of the AC current.

[0077] As mentioned above, technological advancements in the control devices for MMC converters have made it possible to increase the rated value of the AC current (I_x) while maintaining the same size as the ARM power converter. As a result, if magnetic saturation of the three-terminal inductor (22x) can be suppressed, the AC overcurrent withstand capability, which is determined by the peak value of the AC current (I_x) during a grid fault, can be increased. Conversely, it has become important to suppress magnetic saturation due to AC overcurrent in any part of the three-terminal inductor (22x).

[0078] For example, if AC magnetic flux accounts for 10% of the magnetic flux in the wound core portion wound by two coils, and the peak AC current value during a system fault propagation becomes three times the rated value, the AC magnetic flux will increase to the equivalent of 30% unless the wound core saturates. If magnetic saturation occurs in the side legs, yoke, or wound core wound by the coils, the AC inductance value will drop sharply, the peak AC current value will increase cumulatively, and the MMC converter will no longer be able to continue operating.

[0079] As shown in column [C-3] of Figure 10, in the case of a concentric arrangement, at time tn when the absolute value of the current in the negative second coil peaks, the through-current and the alternating current magnetomotive force are added together, and the magnetic flux of the core wound around the coil hardly increases. However, at the side legs and yoke, the magnetic flux of the wound core and the alternating magnetic flux merge, causing an increase in magnetic flux.

[0080] As disclosed in Patent Document 3, in conventional three-phase five-legged reactors, the cross-sectional area of ​​the side legs and yoke was made smaller than the cross-sectional area of ​​the wound core.

[0081] In the three-phase five-legged reactor for MMC converters disclosed in Patent Document 2, the setting of the core cross-sectional area was not clear, leaving challenges in continuing operation during AC system fault propagation.

[0082] When the winding core, side legs, and yoke are constructed from laminated cores of the same material, it is desirable that the sum of the cross-sectional areas of the first and second side legs be greater than the sum of the cross-sectional areas of the first, second, and third winding cores. In other words, it is desirable that the cross-sectional area of ​​the side legs be greater than 1.5 times the cross-sectional area of ​​the winding core.

[0083] Therefore, contrary to the relationship of the relative sizes of the core cross-sectional areas of each part of the three-phase five-legged reactor disclosed in Patent Document 3, the objective is achieved by making the cross-sectional areas of the first and second side legs larger than the cross-sectional area of ​​the first to third winding cores, which are wound with two coils arranged concentrically, and by making the cross-sectional area of ​​the yoke connecting the first to third winding cores and the first and second side legs larger than the cross-sectional area of ​​the first to third winding cores.

[0084] The following describes the means to solve the second problem and achieve the objective.

[0085] Figure 11 shows an example of a frequency converter configured by connecting the DC terminals of two MMC converters 11_1 and 11_2, with their DC sides connected at the rear. The AC terminal of the first MMC converter 11_1 is connected to the AC system 11_3 via a unit transformer 11_4, and the AC terminal of the second MMC converter 11_2 is connected to the armature winding of a synchronous generator motor 11_5. The rotational phase θr of the synchronous generator motor 11_5 is detected by a rotational phase detector 11_7 and input to the control device 11_8 to configure a variable speed generator motor.

[0086] Patent Document 9 discloses a variable-speed pumped-storage power generation system in which a synchronous generator motor 11_5 and a pump turbine 11_6 are directly connected by a rotating shaft, and both ends of a load switch 11_9 are branched to the AC terminals of two MMC converters. At low output, the load switch 11_9 is opened to select variable-speed operation via the MMC converter, and near rated output, the load switch 11_9 is closed to connect directly to the AC system and select constant-speed operation.

[0087] A method is disclosed that allows for bidirectional switching between variable-speed operation and constant-speed operation during operation, thereby achieving both system efficiency and cost reduction through a reduction in the rated capacity of the MMC converter.

[0088] In the variable-speed pumped-storage power generation system described above, during constant-speed operation, the current Ig shown in Figure 11 flows from the u-phase terminal of the synchronous generator motor 11_5 through the load switch 11_9 to the u-phase terminal of the unit transformer 11_4.

[0089] When switching from constant-speed operation to variable-speed operation, the operation of the two MMC converters 11_1 and 11_2 is started, a commutation current It_P is started to flow from the positive power converter 21Pu of the first MMC converter 11_1 to the positive power converter 21Pu of the second MMC converter 11_2 via the load switch 11_9, and a commutation current It_N is started to flow from the negative power converter 21Nu of the first MMC converter 11_1 to the negative power converter 21Nu of the second MMC converter 11_2 via the load switch 11_9.

[0090] The control device 11_8 adjusts the command values ​​of the positive and negative commutation currents (It_P, It_N) to half of the generator current Ig so that the sum of the positive and negative commutation currents (It_P + It_N) matches the generator current Ig for a short time (within one cycle of the AC frequency). As a result, when the current in the load switch 11_9 approaches zero, the load switch 11_9 is opened, the transition to variable speed operation is completed, and variable speed operation by the two MMC converters 11_1 and 11_2 begins.

[0091] As shown in the AC current path in Figure 11 above, the inductive element that suppresses changes in the AC current is the sum of the inductance of the unit transformer 11_4 and the AC inductance of the three-terminal inductive element (22x). However, in the case of commutation current (It_P, It_N), the inductive element that suppresses changes is only the AC inductance of the three-terminal inductive element (22x).

[0092] If the self-inductance of the positive coil Px is L1, the self-inductance of the negative coil Nx is L2, and the mutual inductance is M, then when the commutation current is equally divided between the positive side (It_P) and the negative side (It_N), the AC inductance of this three-terminal inductive element (22x) will be (L1-M) on the positive side and (L2-M) on the negative side. When the positive coil Px is the inner coil, as disclosed in Patent Document 3, the AC inductance of the positive coil Px is only about 1 / 100 of the inductance (L1 + L2 + 2 × M) for the through-current (sum of DC current and circulating current). As a result, the circuit time constant of the positive commutation current becomes less than 1 / 10 of the circuit time constant of the AC current other than during commutation, which presents a problem as it becomes unadjustable.

[0093] To solve this problem, one possible method is to allocate the inner coil of the three-terminal inductor element of the MMC converter on the AC system 11_3 side to the positive side and the inner coil of the three-terminal inductor element of the MMC converter on the synchronous generator motor 11_5 side to the negative side, thereby ensuring the inductance of at least one outer coil in the commutation current paths on both the positive and negative sides. However, this method suffers from a problem in that the inductance balance condition between the positive and negative sides, which is a prerequisite for equally dividing the AC current control into positive and negative sides, is greatly disrupted because there is no inductance in the unit transformer 11_4. As a result, the sum of the positive and negative commutation currents (It_P + It_N) cannot be adjusted to follow the generator current Ig.

[0094] Alternatively, to solve this problem, one method is to add a three-phase AC two-terminal reactor between one of the AC terminals of the two MMC converters and the branching point to the load switch 11_9 to secure the inductance of the commutation current. However, adding AC inductance not only increases the installation space, but also necessitates compensating for the voltage drop due to the added AC inductance during normal variable-speed operation by increasing the voltage of the MMC converter. As a result, this leads to the problem of increasing the size of the MMC converter.

[0095] To solve this second problem, as shown in Figure 12, one method to solve the above-mentioned second problem is to divide the number of turns of the first winding wound around the first to third winding cores of a three-phase five-legged reactor equipped with first and second side legs into two equal parts, add a second winding newly arranged coaxially with the first winding, connect the inner coil of the first winding to the outer coil of the second winding, and connect the outer coil of the first winding to the inner coil of the second winding, thereby eliminating the imbalance in AC inductance between the positive and negative sides of each phase.

[0096] In this case, the magnetomotive force caused by the current flowing through the inner coil of the first winding and the outer coil of the second winding excites the iron core in the same direction, and the windings must be wound such that the magnetomotive force caused by the current flowing through the outer coil of the first winding and the inner coil of the second winding excites the iron core in the same direction.

[0097] As shown in Figure 13, the options for the inner and outer coils that make up a concentric winding can be classified into four quadrants based on the winding diameter and winding direction.

[0098] In the present invention, for the sake of explanation, right-handed winding will be shown as the positive winding direction on the right half of the surface, left-handed winding as the reverse winding direction on the left half of the surface, the inner coil on the upper half of the surface, and the outer coil on the lower half of the surface. Coils arranged in the first to fourth quadrants will be referred to as the first quadrant coil to the fourth quadrant coil, respectively.

[0099] From the perspective of reducing the number of coil types used, it is desirable that both the first and second windings consist of only two types: first quadrant coils and fourth quadrant coils, or only two types: second quadrant coils and third quadrant coils.

[0100] As shown in Figure 24, for example, if the fourth quadrant coil is the outer coil of the first and second windings, and the first quadrant coil is the inner coil of the first and second windings, then terminal a4 of the second winding is set to terminal xp of the three-terminal inductor element 22x, terminal b4 of the second winding is connected to terminal a1 of the first winding, and terminal b1 of the first winding is set to terminal xc.

[0101] Since it is necessary to excite the iron core in the same direction with a through-current from terminal xp through terminal xc towards terminal xn, a conductor is required that passes outside the first and second windings in order to connect terminal b1 of the first winding, which is terminal xc, to terminal a1 of the inner coil of the second winding.

[0102] Then, the b1 terminal of the first quadrant coil, which will be the inner coil of the second winding, is connected to the a4 terminal of the fourth quadrant coil, which will be the outer coil of the first winding, and this b4 terminal is designated as the xn terminal.

[0103] Since the xc terminal is divided into the first winding and the second winding, an external conductor is required to connect the two, just as it is when combining a second quadrant coil and a third quadrant coil.

[0104] On the other hand, although it has the drawback of requiring four types of coils per phase, by using the first quadrant coil and the third quadrant coil to form the first or second winding, and the second quadrant coil and the fourth quadrant coil to form the second or first winding, the xc terminals, which are separated into the first and second windings, can be connected over the shortest distance.

[0105] For example, as shown in Figure 25, if the fourth quadrant coil is the outer coil of the second winding and the first quadrant coil is the inner coil of the first winding, then terminal a4 is set to terminal xp of the three-terminal inductor element 22x, terminal b4 and terminal a1 are connected, and terminal b1 is set to terminal xc.

[0106] In this case, the through-current from terminal xp through terminal xc towards terminal xn requires the core to be excited in the same direction, so the third quadrant coil is selected as the outer coil of the first winding, and the second quadrant coil is selected as the inner coil of the second winding.

[0107] As a result, terminal b3, which is adjacent to terminal b1 (which becomes the xc terminal) and is on the same side as the terminal of the outer coil, becomes another xc terminal, thus eliminating the need for a long outer conductor.

[0108] Then, connect terminal a3 of the first winding to terminal b2 of the second winding, and make terminal a2 terminal xn.

[0109] As described above, by configuring two windings with four types of coils from the first to the fourth quadrants for each of the three AC phases, and arranging the two windings coaxially on the core of each phase, the second problem is solved by balancing the positive and negative AC inductances with a minimum number of connecting conductors, and the objective can be achieved while eliminating induction heating and electromagnetic vibration of the surrounding area caused by the external conductor.

[0110] The following describes the means to solve the third problem and achieve the objective.

[0111] Figure 14-1 shows the magnetic flux due to the alternating current flowing through the positive coil 22Pv and the negative coil 22Nv that wind around the second leg of the three-phase five-legged reactor in Figure 12.

[0112] The magnetic flux distributions of the first and second windings are symmetrical vertically, indicating that the second problem has been solved.

[0113] Since the magnetic flux is symmetrical with respect to the second leg, focusing on the left half, the vertical magnetic flux passing through the cylindrical air gap between the inner and outer coils of the first winding is divided into three parts: the flux that recirculates through the first side leg in the path shown by the solid line, the flux that recirculates through the second winding core shown by the dashed line, and the flux that recirculates through the first winding core shown by the dashed line. The ratio of flux recirculation is proportional to the reciprocal of the magnetic resistance of each recirculation path.

[0114] To take advantage of the concentric arrangement shown in column [A-3] of Figure 10, it is desirable that magnetic flux flows only through the paths indicated by solid lines and not through the paths indicated by dashed lines.

[0115] Figure 14-2 shows the magnetic flux due to the alternating current flowing through the positive coil 22Pu and the negative coil 22Nu that wind the first winding core of the three-phase five-legged reactor in Figure 12.

[0116] Similar to Figure 14-1, the magnetic flux distributions of the first and second windings are symmetrical vertically, indicating that the second problem has been solved.

[0117] Focusing on the magnetic flux flowing through the left half of the first leg, the vertical magnetic flux passing through the cylindrical air gap sandwiched between the inner and outer coils of the first winding is divided into two: the magnetic flux that recirculates through the first side leg (shown by the solid line) and the magnetic flux that recirculates through the first winding core (shown by the dashed line).

[0118] Focusing on the magnetic flux flowing through the right half of the first leg, the vertical magnetic flux passing through the cylindrical air gap sandwiched between the inner and outer coils of the first winding is divided into four parts: the magnetic flux that recirculates through the second side leg (shown by the solid line), the magnetic flux that recirculates through the first winding core (shown by the dashed line), the magnetic flux that recirculates through the second winding core (shown by the dashed line), and the magnetic flux that recirculates through the third winding core (shown by the dashed line).

[0119] The ratio of magnetic flux division is proportional to the reciprocal of the magnetic resistance of each return channel.

[0120] To take advantage of the concentric arrangement shown in column [A-3] of Figure 10, it is desirable that magnetic flux flows only through the paths indicated by solid lines and not through the paths indicated by dashed lines.

[0121] Figures 14-1 and 14-2 show different AC flux paths. The AC inductance of the coil winding the second leg is smaller than the AC inductance of the coils winding the first and third legs, resulting in a third problem: increased non-theoretical harmonics due to phase imbalance in AC inductance.

[0122] Furthermore, the proportion of magnetic flux returning from the winding core increases beyond the side legs, which are the intended return paths for the magnetic flux. This reduces the advantages of adopting a concentric arrangement and creates a problem where the winding core is more susceptible to magnetic saturation due to alternating current.

[0123] As shown in Figure 15-1, the third problem can be solved by providing auxiliary yokes between the first and second windings of each phase. The first side leg and the first winding core are fastened with auxiliary yoke 15_1, the first winding core and the second winding core with auxiliary yoke 15_2, the second winding core and the third winding core with auxiliary yoke 15_3, and the third winding core and the second side leg with auxiliary yoke 15_4.

[0124] Alternatively, as shown in Figure 15-2, the third problem can be solved by dividing the first winding core, the second winding core, and the third winding core into two equal parts between the first and second windings, and providing an auxiliary yoke 15_5 that fastens the first and second side legs together between the dividing surfaces.

[0125] Alternatively, as shown in Figure 15-3, the third problem can be solved by providing a third side leg between the first winding core and the second winding core, and a fourth side leg between the second winding core and the third winding core, and fastening the upper and lower yokes with the third and fourth side legs.

[0126] Figure 16-1 shows the magnetic flux due to the alternating current flowing through the positive coil 22Pv and the negative coil 22Nv that wind around the second leg of the three-phase five-legged reactor shown in Figure 15-1.

[0127] Similar to Figure 15-1, the magnetic flux is symmetrical with respect to the second leg, so focusing on the left half, the vertical magnetic flux passing through the cylindrical air gap between the inner and outer coils of the first winding is divided into three parts: the flux that recirculates through the first side leg (shown by the solid line), the flux that recirculates through the second winding core (shown by the dashed line), and the flux that recirculates through the first winding core (shown by the dashed line). The ratio of flux recirculation is proportional to the reciprocal of the magnetic resistance of each recirculation channel.

[0128] If the cross-sectional area of ​​the magnetic path is set so that the magnetic resistance of auxiliary yokes 15-1 and 15-2 is negligible, the current division ratio is proportional to the reciprocal of the magnetic resistance of the side leg, the first winding core, and the second winding core, so most of the magnetic flux flows to the first side leg. As a result, the AC magnetic flux is divided into two, flowing to the first side leg and the second side leg.

[0129] Figure 16-2 shows the magnetic flux due to the alternating current flowing through the positive coil 22Pu and the negative coil 22Nu that wind around the first leg of the three-phase five-legged reactor shown in Figure 15-2.

[0130] Focusing on the magnetic flux flowing through the left half of the first leg, the vertical magnetic flux passing through the cylindrical air gap sandwiched between the inner and outer coils of the first winding is divided into two: the magnetic flux that recirculates through the first side leg (shown by the solid line) and the magnetic flux that recirculates through the first winding core (shown by the dashed line).

[0131] If the cross-sectional area of ​​the magnetic path is set so that the magnetic resistance of the auxiliary yoke 15-1 is negligible, the current division ratio is proportional to the reciprocal of the magnetic resistance of the side leg and the first winding core, so most of the magnetic flux flows to the first side leg.

[0132] Focusing on the magnetic flux flowing through the right half of the first leg, the vertical magnetic flux passing through the cylindrical air gap sandwiched between the inner and outer coils of the first winding is divided into four parts: the magnetic flux that recirculates through the second side leg (shown by the solid line), the magnetic flux that recirculates through the first winding core (shown by the dashed line), the magnetic flux that recirculates through the second winding core (shown by the dashed line), and the magnetic flux that recirculates through the third winding core (shown by the dashed line).

[0133] The ratio of magnetic flux division is proportional to the reciprocal of the magnetic resistance of each return channel.

[0134] If the cross-sectional area of ​​the magnetic path is set so that the magnetic resistance of auxiliary yokes 15-2, 15-3, and 15-4 is negligible, the current division ratio is proportional to the reciprocal of the magnetic resistance of the side leg, the second winding core, and the third winding core, so most of the magnetic flux flows to the second side leg. As a result, the AC magnetic flux is divided into two, flowing to the first side leg and the second side leg.

[0135] A comparison of Figures 16-1 and 16-2 shows that the three-phase five-legged reactor, with the addition of auxiliary yokes 15-1, 15-2, 15-3, and 15-4 shown in Figure 15-1, can eliminate the imbalance in the three-phase AC inductance and solve the third problem by virtually eliminating the AC magnetic flux flowing through the first, second, and third winding cores.

[0136] Alternatively, the third problem can be solved by eliminating the imbalance in the three-phase AC inductance and virtually eliminating the AC magnetic flux flowing through the first, second, and third winding cores by using a three-phase five-legged reactor with an auxiliary yoke 15-5 added as shown in Figure 15-2.

[0137] Figure 17-1 shows the magnetic flux due to the alternating current flowing through the positive coil 22Pv and the negative coil 22Nv that wind around the second leg of the three-phase reactor in Figure 15-3.

[0138] Similar to Figure 15-1, the magnetic flux is symmetrical with respect to the second leg, so focusing on the left half, the vertical magnetic flux passing through the cylindrical air gap between the inner and outer coils of the first winding is divided into two: the magnetic flux that recirculates through the side leg 15_6 (shown by the solid line) and the magnetic flux that recirculates through the second winding core (shown by the dashed line). The ratio of magnetic flux recirculation is proportional to the reciprocal of the magnetic resistance of each recirculation channel.

[0139] If the cross-sectional area of ​​the magnetic path is set so that the magnetic resistance of the side leg 15_6 is negligible, the current division ratio is proportional to the reciprocal of the air gap magnetic resistance from the side leg 15_6 to the outer coil, and the reciprocal of the sum of the air gap magnetic resistance of the second winding core and the air gap magnetic resistance up to the inner coil. As a result, most of the magnetic flux flows to the side leg 15_6. However, the magnetic flux flowing through the second winding core is greater than that shown in Figure 16-1.

[0140] Figure 17-2 shows the magnetic flux due to the alternating current flowing through the positive coil 22Pu and the negative coil 22Nu that wind around the first leg of the three-phase reactor in Figure 15-3.

[0141] Focusing on the magnetic flux flowing through the left half of the first leg, the vertical magnetic flux passing through the cylindrical air gap sandwiched between the inner and outer coils of the first winding is divided into two: the magnetic flux that recirculates through the first side leg (shown by the solid line) and the magnetic flux that recirculates through the first winding core (shown by the dashed line).

[0142] If the cross-sectional area of ​​the magnetic path is set so that the magnetic resistance of the first leg is negligible, the current division ratio is proportional to the reciprocal of the air gap magnetic resistance from the first leg to the outer coil, and the reciprocal of the sum of the air gap magnetic resistance of the first winding core and the air gap magnetic resistance to the inner coil. As a result, almost all of the magnetic flux flows to the first leg.

[0143] Focusing on the magnetic flux flowing through the right half of the first leg, the vertical magnetic flux passing through the cylindrical air gap sandwiched between the inner and outer coils of the first winding is divided into two: the magnetic flux that recirculates through the side leg 15_6, shown by the solid line, and the magnetic flux that recirculates through the first winding core, shown by the dashed line.

[0144] If the cross-sectional area of ​​the magnetic path is set so that the magnetic resistance of the side leg 15_6 is negligible, the current division ratio is proportional to the reciprocal of the air gap magnetic resistance from the side leg 15_6 to the outer coil, and the reciprocal of the sum of the series magnetic resistances of the air gap magnetic resistance of the first winding core and the air gap magnetic resistance to the inner coil. As a result, almost all of the magnetic flux flows to the side leg 15_6.

[0145] A comparison of Figures 17-1 and 17-2 shows that a three-phase reactor with the addition of side legs 15_6 and 15_7 shown in Figure 15-3 can eliminate the imbalance in the three-phase AC inductance and suppress the AC magnetic flux flowing through the first, second, and third winding cores.

[0146] However, although the magnetic flux flowing through the first and second winding cores is greater than that shown in Figures 16-1 and 16-2, the third problem can be solved by adjusting the magnetic resistance of each part.

[0147] On the other hand, Figure 15-3 has the advantage of being easier to assemble compared to Figures 15-1 and 15-2.

[0148] The above configuration proved to be suitable for solving the problem.

[0149] The MMC converter according to the present invention can achieve both miniaturization of the device and the assurance of continued operation performance during the propagation of a grid fault.

[0150] Figure 1 shows the configuration of an MMC converter according to the present invention. Figure 2 shows the configuration of a leg circuit according to the present invention. Figure 3 shows the configuration of a unit converter according to the present invention. Figure 4 shows the configuration of a power converter control device according to the present invention. Figure 5 shows the configuration of a DC current control device according to the present invention. Figure 6 shows the configuration of a positive / negative balance control device according to the present invention. Figure 7 shows the configuration of a unit converter control device according to the present invention. Figure 8 is a circuit diagram of an MMC converter when branch circuits for connecting capacitors inside the unit converter are ignored. Figure 9 shows the current path when the circuit degrees of freedom of the MMC converter are considered to be 5 and divided into three current elements (DC current × 1, AC current × 2, circulating current × 2). Figure 10 shows the circuit configuration of a three-terminal inductive element according to the present invention, current waveform, and a comparison of the case where two coils are magnetically coupled in a coaxial arrangement and the case where they are magnetically coupled in a concentric arrangement. Figure 11 shows the commutation current path when switching from constant speed operation to variable speed operation in a variable speed pumped-storage power generation system in which the DC side of the MMC converter according to the present invention is connected to the back, the AC side of one MMC converter is connected to a synchronous generator motor, and a bypass circuit directly connected to an AC power source is provided. Figure 12 shows the winding and core configuration of Embodiment 2 of the 3-phase 5-legged reactor according to the present invention. Figure 13 shows a comparison of four types of inner and outer coils that constitute the winding of the 3-phase 5-legged reactor according to the present invention. Figure 14-1 shows the magnetic flux distribution when the second winding core is excited with AC current in Embodiment 2 of the 3-phase 5-legged reactor according to the present invention. Figure 14-2 shows the magnetic flux distribution when the first winding core is excited with AC current in Embodiment 2 of the 3-phase 5-legged reactor according to the present invention. Figure 15-1 shows the winding and core configuration of Embodiment 3 of the 3-phase 5-legged reactor according to the present invention. Figure 15-2 shows the winding and core configuration of Embodiment 5 of the three-phase five-legged reactor according to the present invention. Figure 15-3 shows the winding and core configuration of Embodiment 6 of the three-phase five-legged reactor according to the present invention. Figure 16-1 shows the magnetic flux distribution when the second winding core is excited with alternating current in Embodiment 3 of the three-phase five-legged reactor according to the present invention.Figure 16-2 shows the magnetic flux distribution when the first winding core is excited with alternating current in Embodiment 3 of the three-phase five-legged reactor according to the present invention. Figure 17-1 shows the magnetic flux distribution when the second winding core is excited with alternating current in Embodiment 6 of the three-phase five-legged reactor according to the present invention. Figure 17-2 shows the magnetic flux distribution when the first winding core is excited with alternating current in Embodiment 6 of the three-phase five-legged reactor according to the present invention. Figure 18 is a diagram showing the structure of Embodiment 1 of the three-phase five-legged reactor according to the present invention in front view, left side view, rear view, and top view, as shown in third-angle projection. Figure 19 is a diagram showing the structure of Embodiment 2 of the three-phase five-legged reactor according to the present invention in front view, left side view, rear view, and top view, as shown in third-angle projection. Figure 20 shows the structure of Embodiment 3 of the three-phase five-legged reactor according to the present invention, as depicted in third-angle projection, in front view, left side view, rear view, and top view. Figure 21 shows the structure of Embodiment 4, a derivative structure of Embodiment 3 of the three-phase five-legged reactor according to the present invention, as depicted in third-angle projection, in front view, left side view, rear view, and top view. Figure 22 shows the structure of Embodiment 5 of the three-phase five-legged reactor according to the present invention, as depicted in third-angle projection, in front view, left side view, rear view, and top view. Figure 23 shows the structure of Embodiment 6 of the three-phase five-legged reactor according to the present invention, as depicted in third-angle projection, in front view, left side view, rear view, and top view. Figure 24 shows an example of coil configuration applied to the MMC converter according to the present invention. Figure 25 shows an example of coil configuration applied to the MMC converter according to the present invention.

[0151] The following describes in detail an embodiment of the MMC converter according to the present invention, based on the drawings. However, this embodiment does not limit the present invention.

[0152] Figure 1 shows the configuration of the MMC converter according to the present invention.

[0153] 1 is an MMC converter, which is connected between the positive terminal (P) and negative terminal (N) of the DC power supply 1_1 and the three terminals (u, v, w) of the three-phase AC power supply 1_2. The MMC converter 1 is equipped with three leg circuits 2u, 2v, and 2w.

[0154] The positive terminals (Pu, Pv, Pw) of the three leg circuits 2u, 2v, and 2w are connected in a star configuration to the positive terminal (P) of the DC power supply 1_1, the negative terminals (Nu, Nv, Nw) are connected in a star configuration to the negative terminal (N) of the DC power supply 1_1, and the intermediate terminals (ACu, ACv, ACw) are connected to the three terminals (u, v, w) of the AC power supply 1_2. 1_3 is an AC sensor that outputs the voltage (Vac), reactive power (Qac), and reference phase (θ) of the AC power supply 1_2.

[0155] Figure 2 shows the configuration of Example 1 of the leg circuit 2 (2u, 2v, 2w) according to the present invention.

[0156] For simplicity, x will be used below to represent the three phases (u, v, w) of AC power supply 1_2.

[0157] The leg circuit 2 comprises a two-terminal positive arm power circuit 21Px, a two-terminal negative arm power circuit 21Nx, a three-terminal inductive element 22x consisting of a positive inductive element 22Px and a negative inductive element 22Nx such as a reactor, and two current transformers (23Px, 23Nx).

[0158] The positive terminal (APx) of the positive arm power circuit 21Px is connected to the positive terminal (Px) of the leg circuit 2, its negative terminal (BPx) is connected to the positive terminal (xp) of the three-terminal inductor element 22x, and the intermediate terminal (xc) of the three-terminal inductor element 22x is connected to the intermediate terminal (ACx) of the leg circuit 2.

[0159] The negative terminal (xn) of the three-terminal inductor element 22x is connected to the positive terminal (ANx) of the negative arm power circuit 21Nx, and its negative terminal (BNx) is connected to the negative terminal (Nx) of the leg circuit 2.

[0160] The positive-side current transformer 23Px detects the arm current (I_xp) of the positive-side arm power circuit 21Px, with the direction from the negative terminal (BPx) to the positive terminal (APx) being positive, and distributes the output to the positive-side arm control device 24Px and the power converter control device 4.

[0161] The negative current transformer 23Nx detects the arm current (I_xn) of the negative arm power circuit 21Nx, with the direction from the positive terminal (ANx) to the negative terminal (BNx) being positive, and distributes the output to the negative arm control device 24Nx and the power converter control device 4.

[0162] The positive arm power circuit 21Px and the negative arm power circuit 21Nx each have K (where K is a natural number of 2 or more) two-terminal unit converters 3 connected in series between their respective positive terminals (APx, ANx) and negative terminals (BPx, BNx).

[0163] 24 is an arm control device, where the positive arm control device 24Px is connected to the positive arm power circuit 21Px, and the negative arm control device 24Nx is connected to the negative arm power circuit 21Nx.

[0164] The arm control device 24 consists of K unit converter control devices 8 and an average value calculator 2_1 that outputs the average value of K input signals.

[0165] For simplicity, y will be used below to represent the subscripts (P, N) of the two poles of DC power supply 1_1. k will be used to represent the subscripts (1, 2, ..., K) of the K unit converters.

[0166] The positive arm control device 24Px receives the arm voltage command (Vrf_xp) from the power converter control device 4 and branches the reference phase (θ_x) into input. It also receives the arm current (I_xp) from the positive current transformer 23Px as the unit converter current (Ib_xp=I_xp) via the gain 2_2.

[0167] The negative arm control device 24Nx receives the arm voltage command (Vrf_xn) from the power converter control device 4 and branches the reference phase (θ_x) into input. It also receives the arm current (I_xn) from the negative current transformer 23Nx as the unit converter current (Ib_xn = -I_xn) with its sign inverted via the gain 2_3.

[0168] The arm control device 24 distributes and outputs the unit converter current (Ib_xy) to the unit converter control devices 8, which are provided for each of the K unit converters 3. It also inputs a reference phase (θ_x) and distributes and outputs the arm capacitor average voltage (Vc_xy_t), which is moved averaged over the AC system period, to the unit converter control devices 8.

[0169] Furthermore, the average voltage (Vc_xy) of the arm capacitor from the average value calculator 2_1 is output to the power converter control device 4.

[0170] Figure 3 shows the configuration of the unit converter 3 according to the present invention.

[0171] The unit converter 3 comprises a unit converter power circuit and a unit converter auxiliary circuit, both configured as half-bridge power circuits.

[0172] The unit converter power circuit is configured by connecting the positive terminal of the upper self-extinguishing element 31H (the collector of the IGBT or the anode of the IGCT) to the positive terminal of the capacitor 32, and branching the negative terminal of the upper self-extinguishing element 31H (the emitter of the IGBT or the cathode of the IGCT) to the positive terminal (C) of the unit converter 3 and the positive terminal of the lower self-extinguishing element 31L. The negative terminal of the lower self-extinguishing element 31L is branched to the negative terminal of the capacitor 32 and the negative terminal (D) of the unit converter 3. The upper diode 33H is connected in antiparallel to the upper self-extinguishing element 31H. The lower diode 33L is connected in antiparallel to the lower self-extinguishing element 31L to form a half-bridge power circuit.

[0173] The unit converter auxiliary circuit comprises a gate drive device 36, a capacitor voltage detector 34, and a voltage signal converter 35.

[0174] The gate drive device 36 receives gate control signals (GH, GL) from the unit converter control device 8, which constitutes the positive arm control device 24Px and the negative arm control device 24Nx, performs level conversion, and outputs biased gate pulses to the gate circuit of the upper self-extinguishing element 31H and the gate circuit of the lower self-extinguishing element 31L.

[0175] The gate control signals (GH, GL) adjust the on / off period to adjust the terminal voltage (Vb) of the unit converter 3 to the required value through PWM modulation.

[0176] The capacitor voltage detector 34 outputs the voltage of the capacitor 32 to the voltage signal converter 35, which converts this to a signal level to generate a voltage signal Vc. The unit converter 3 outputs the capacitor voltage signal Vc generated by the voltage signal converter 35 to the positive arm control device 24Px and the negative arm control device 24Nx as a capacitor voltage signal (Vc_xy_k).

[0177] Figure 7 shows the configuration of the unit converter control device 8 according to the present invention.

[0178] For each arm power circuit 21, the arm voltage command (Vrf_xy) output from the power converter control device 4 is multiplied by 1 / K to calculate the unit converter voltage command (Vcrf_xy).

[0179] The capacitor voltage deviation (Vcd_xy_k) is calculated by comparing the capacitor arm moving average voltage signal (Vc_xy_t) with the capacitor voltage signal (Vc_xy_k).

[0180] The current sign detector 802 detects the sign of the unit converter current (Ib_xy). When the unit converter current (Ib_xy) has a positive sign, it switches the Gain of the gain 803 to positive; when the unit converter current (Ib_xy) has a negative sign, it switches the Gain of the gain 803 to negative. The output of this gain 803 is used to energize the unit converter voltage command (Vcrf_xy) and input to the divider 801.

[0181] By biasing the output of this gain 803 to the unit converter voltage command (Vcrf_xy), the in-arm balancing control of the capacitor voltage is achieved.

[0182] The divider 801 outputs a modulation rate command (αrf_xy_k) obtained by normalizing the unit converter voltage command (Vcrf_xy) after the gain 803 has been energized by the capacitor arm moving average voltage signal (Vc_xy_t).

[0183] 804 is a limiter that generates a modulation rate signal (αrf) that limits the input modulation rate command (αrf_xy_k) to between a maximum value αmax less than 1 and a minimum value αmin greater than 0, and outputs it to the comparator 805.

[0184] 806 is a carrier output device that outputs a triangular wave with a maximum value of 1 and a minimum value of 0 as the carrier wave to comparator 805.

[0185] The comparator 805 outputs the output signal (GH) at level 1 and the output signal (GL) at level 0 when the modulation index signal (αrf) is greater than the carrier wave. When the modulation index signal (αrf) is less than the carrier wave, it outputs the output signal (GH) at level 0 and the output signal (GL) at level 1, and alternately outputs the gate control command (GH_xy_k) to the upper self-extinguishing element (31H) and the gate control command (GL_xy_k) to the lower self-extinguishing element (31L) on and off.

[0186] Figure 4 shows the configuration of Embodiment 1 of the power converter control device 4 according to the present invention. Figure 4 references the configuration disclosed in Patent Documents 7 and 8.

[0187] 401 is an arm current calculator that takes the arm currents (I_up, I_vp, I_wp, I_un, I_vn, I_wn) from the current transformer 23 as input and calculates and outputs the AC currents (I_u, I_v, I_w) that enter the intermediate terminals (ACu, ACv, ACw) of the MMC converter 1 from the AC power supply 1_2, and the through-currents (I_cu, I_cv, I_cw) that enter the positive terminal (Px) from the negative terminal (Nx) of the leg circuit (2u, 2v, 2w). It also calculates and outputs the DC current (I_dc) from the positive terminal (P) of the MMC converter 1.

[0188] The alternating current (I_u, I_v, I_w) is calculated using equation (1).

[0189]

[0190] The through-currents (I_cu, I_cv, I_cw) are calculated using equation (2).

[0191]

[0192] The direct current (I_dc) is calculated using equation (3).

[0193]

[0194] 402 is a capacitor voltage calculator that receives the average capacitor arm voltages (Vc_up, Vc_vp, Vc_wp, Vc_un, Vc_vn, Vc_wn) from the arm control device 24 as input, and calculates and outputs the average capacitor leg voltages (Vc_u, Vc_v, Vc_w) of the leg circuit 2 and the positive-negative capacitor difference voltages (Vc_pnu, Vc_pnv, Vc_pnw). It also calculates and outputs the average voltage of all capacitors (Vc_ave).

[0195] The capacitor leg average voltages (Vc_u, Vc_v, Vc_w) are calculated using equation (4).

[0196]

[0197] The voltage difference between the positive and negative terminals of the capacitor (Vc_pnu, Vc_pnv, Vc_pnw) is calculated using equation (5).

[0198]

[0199] The average voltage of all capacitors (Vc_ave) is calculated using equation (6).

[0200]

[0201] The power converter control device 4 outputs a DC current (I_dc) to the DC current control device 5, and the DC current control device 5 inputs a DC voltage command (Vrf_dc).

[0202] The power converter control device 4 comprises an AC current control device 6, a circulating current control device 7, a positive / negative balance control device 9, a leg balance control device 901, and adders 902 and 903.

[0203] 403 is a capacitor voltage regulator that compares the average voltage of all capacitors (Vc_ave) from the capacitor voltage calculator 402 with the command value (Vcrf) and inputs it to output the active power component (Irf_aq) of the AC current command to the AC current control device 6 and the positive / negative balance control device 9.

[0204] 404 is a reactive power regulator that matches the reactive power (Qac) and command value (Qacrf) from the AC sensor 1_3 and outputs it to the q-side input of the output switch 405.

[0205] 406 is an AC voltage regulator that matches the voltage (Vac) from AC sensor 1_3 with the command value (Vacrf) and outputs it to the v-side input of output switch 405.

[0206] The output switch 405 selects the input signal from the reactive power regulator 404 connected to the q-side input or the input signal from the AC voltage regulator 406 connected to the v-side input, and outputs it to the AC current control device 6 and the positive / negative balance control device 9 as the reactive power component (Irf_ad) of the AC current command.

[0207] The AC current control device 6 inputs the AC current (I_u, I_v, I_w) from the arm current calculator 401 to a three-phase two-phase converter (not shown) that uses the positive-sequence voltage θ of the AC power supply 1_2 as the reference phase, and outputs two-phase current (I_aq, I_ad). This two-phase current (I_aq, I_ad) is then compared with a two-phase current command (Irf_aq, Irf_ad) and input to a proportional-integral calculator (not shown). The output of this proportional-integral calculator, a two-phase voltage command (Vrf_aq, Vrf_ad), is input to a two-phase three-phase inverse converter (not shown) and outputs the three-phase arm voltage command AC component (Vrf_au, Vrf_av, Vrf_aw).

[0208] The three-phase two-phase converter that constitutes the AC current control device 6 calculates the relationship between the input and output in equation (7).

[0209]

[0210] The two-phase three-phase inverse converter that constitutes the AC current control device 6 calculates the relationship between the input and output in equation (8).

[0211]

[0212] The circulating current control device 7 inputs the through-current (I_cu, I_cv, I_cw) and the three-phase circulating current commands (Irf_cu, Irf_cv, Irf_cw) from the arm current calculator 401 to a three-phase two-phase converter (not shown) with a reference phase twice the angle (2θ). The output of this three-phase two-phase converter, the two-phase circulating current commands (Irf_cq, Irf_cd) and circulating currents (I_cq, I_cd), are compared for each phase and input to a proportional-integral calculator (not shown). The output of this proportional-integral calculator, the two-phase voltage commands (Vrf_cq, Vrf_cd), is input to a two-phase three-phase inverse converter (not shown) to output the three-phase arm voltage command circulating component (Vrf_cu, Vrf_cv, Vrf_cw).

[0213] The three-phase two-phase converter that constitutes the circulating current control device 7 calculates the relationship between the input and output in equations (9) and (10).

[0214]

[0215]

[0216] The two-phase three-phase inverse converter that constitutes the circulating current control device 7 calculates the relationship between the input and output in equation (11).

[0217]

[0218] The positive-negative balance control device 9 receives the positive-negative differential voltage (Vc_pnu, Vc_pnv, Vc_pnw), two-phase AC current command values ​​(Irf_aq, Irf_ad), and reference phase (θ) from the capacitor voltage calculator 402, and outputs three-phase circulating current commands (Irf_cu, Irf_cv, Irf_cw).

[0219] The inter-leg balancing control device 901 performs proportional and integral calculations to make the leg voltage deviations (Vc_bu, Vc_bv, Vc_bw) zero and outputs the three-phase arm voltage command leg balancing components (Vrf_bu, Vrf_bv, Vrf_bw).

[0220] The arm voltage command calculator 410, which consists of an adder and a sign inverter, and adders 902 and 903 output arm voltage commands (Vrf_up, Vrf_un, Vrf_vp, Vrf_vn, Vrf_wp, Vrf_wn).

[0221] The arm voltage commands (Vrf_up, Vrf_un, Vrf_vp, Vrf_vn, Vrf_wp, Vrf_wn) are calculated using equation (12).

[0222]

[0223] Figure 5 shows the configuration of the DC current control device 5 according to the present invention.

[0224] The divider 501 takes an active power command (Prf) and a DC voltage command (Vdcrf) as inputs and outputs a DC current command (Irf_dc).

[0225] The DC current command (Irf_dc) output by the divider 501 and the DC current (I_dc) are compared and input to the DC current regulator 503.

[0226] The command value from the DC current regulator 503 is tuned with a DC voltage command (Vdcrf) and output as the DC component of the arm voltage command (Vrf_dc).

[0227] 504 is an adder that inverts the sign of the output of the DC current regulator 503 to generate a DC voltage command (Vdcrf), which is then output as the DC component of the arm voltage command (Vrf_dc_fc) to other MMC converters (not shown in Figure 1) that are connected behind the DC terminals (P, N) of the MMC converter 1.

[0228] Figure 6 shows the configuration of the positive / negative balance control device 9 according to the present invention.

[0229] The positive-negative voltage difference (Vc_pnu, Vc_pnv, Vc_pnw) from the capacitor voltage calculator 402 is input to the gains 1101u, 1101v, and 1101w. The reference phase (θ) is shifted in the phase order of each phase and input to the cosine wave generators 1102u, 1102v, and 1102w.

[0230] The outputs of gains 1101u, 1101v, and 1101w, and the outputs of cosine wave generators 1102u, 1102v, and 1102w are input to the first input terminal (a) and second input terminal (b) of multipliers 1103u, 1103v, and 1103w, respectively, for each phase.

[0231] 1105 is an absolute value calculator that outputs the vector absolute value of the two-phase AC current command values ​​(Irf_aq, Irf_ad), inputs it to the function generator 1106, and inputs the output of the function generator 1106 to the first terminal of the switch 1104.

[0232] The second terminal of switch 1104 is branched and input to the third input terminal (c) of multipliers 1103u, 1103v, and 1103w. In addition, the switching signals of switch 1104 are input to multipliers 1103u, 1103v, and 1103w.

[0233] The multipliers 1103u, 1103v, and 1103w output the multiplication result (a × b × c) of the first to third inputs when the switch 1104 is closed, and output the multiplication result (a × b) of the first and second inputs when the switch 1104 is open.

[0234] The outputs of the multipliers 1103u, 1103v, and 1103w are output as three-phase current commands (Irf_cu, Irf_cv, Irf_cw).

[0235] Figure 18 shows the structure of Embodiment 1 of the three-phase five-legged reactor according to the present invention, as represented in third-angle projection, in front view, left side view, rear view, and top view.

[0236] 101u, 101v, and 101w are wound cores consisting of three iron cores containing multiple voids filled with a non-magnetic material, and are wound around winding 106x, in which x is represented by three phases (u, v, w).

[0237] The winding 106x consists of a concentrically arranged inner coil 104x and an outer coil 105x, which are magnetically coupled by the winding core 101x.

[0238] In the example shown in Figure 18, the inner coil 104x is right-handed (first quadrant coil) and the outer coil 105x is left-handed (third quadrant coil). However, it is also possible to select a left-handed inner coil 104x (second quadrant coil) and a right-handed outer coil 105 (fourth quadrant coil).

[0239] The a1 terminal of the inner coil 104x is brought out to the positive terminal (xn terminal), the a3 terminal of the outer coil 105x is brought out to the negative terminal (xp terminal), the b1 terminal of the inner coil 104x and the b3 terminal of the outer coil 105x are connected to the branch conductor 110x, and the conductor connection part is brought out to the intermediate terminal (xc).

[0240] 102r and 102s are side legs consisting of two iron cores filled with a non-magnetic material and containing no voids, and their core axes are positioned on the same plane as the axes of the three wound iron cores (101u, 101v, and 101w).

[0241] 103t and 103b are yokes consisting of two void-free iron cores filled with a non-magnetic material, and their core axes are arranged on the same plane as the axes of the three wound cores (101u, 101v, 101w) and the two side legs (102r, 102s).

[0242] When the core materials of the winding core (101u, 101v, 101w), side legs (102r, 102s), and yoke (103t, 103b) are the same, the sum of the cross-sectional areas of the side legs should be greater than the sum of the magnetic path cross-sectional areas of the winding core, and the sum of the cross-sectional areas of the yoke should be greater than the sum of the magnetic path cross-sectional areas of the winding core. Alternatively, by making the cross-sectional areas of the side legs and yoke greater than 1.5 times the cross-sectional area of ​​the winding core, it is possible to prevent magnetic saturation of the side legs and yoke from starting before the winding core when an AC system fault propagates, thereby preventing the phenomenon in which the inductance values ​​of AC current and through current drop sharply and operation becomes impossible.

[0243] Furthermore, by arranging the terminal lead-out positions of the inner coil 104x and the outer coil 105x symmetrically with respect to the plane stretched by the iron core axis, it is possible to minimize the length of the connecting conductor between the coils while ensuring the insulation distance between the positive terminal (xp) and the negative terminal (xn).

[0244] Figure 19 shows the structure of Embodiment 2 of the three-phase five-legged reactor according to the present invention, as represented in third-angle projection, in front view, left side view, rear view, and top view.

[0245] The components numbered the same as in Figure 18 have the same content, and their explanations will be omitted to avoid duplication.

[0246] Winding 109x, in which x is represented by three phases (u, v, w), is arranged coaxially with winding 106x and winding core 101x.

[0247] The winding 109x consists of a concentrically arranged inner coil 107x and an outer coil 108x, which are magnetically coupled by the winding core 101x.

[0248] In the example in Figure 19, left-handed winding (second quadrant coil) is selected for the inner coil 107x and right-handed winding (fourth quadrant coil) is selected for the outer coil 108x. However, if the inner coil (104x) of winding 106x is left-handed winding (second quadrant coil) and the outer coil (105x) is right-handed winding (fourth quadrant coil), then right-handed winding (first quadrant coil) must be selected for the inner coil 107x and left-handed winding (third quadrant coil) for the outer coil 108x.

[0249] In the example shown in Figure 19, the winding 109x is positioned on the yoke 103b side, but the winding 109x may also be positioned on the yoke 103t side.

[0250] The a2 terminal of the inner coil 107x is brought out to the positive terminal (xp terminal), the a4 terminal of the outer coil 108x is brought out to the negative terminal (xn terminal), the b2 terminal of the inner coil 107x and the a3 terminal of the outer coil 105x are connected by a connecting conductor 111x, and the b4 terminal of the outer coil 108x and the a1 terminal of the inner coil 104x are connected by a connecting conductor 112x.

[0251] By magnetically coupling the windings 106x and 109x described above with the winding core 101x, the positive inductance (L1-M) for the AC current between the positive terminal (xp terminal) and the intermediate terminal (xc terminal) and the negative inductance (L2-M) for the AC current between the negative terminal (xn terminal) and the intermediate terminal (xc terminal) can be made to the same value.

[0252] By arranging the a3 terminal of the outer coil 105x and the b2 terminal of the inner coil 107x on the same side with respect to the plane stretched by the core axis of the wound iron core (101u, 101v, 101w), and arranging the a1 terminal of the inner coil 104x and the b4 terminal of the outer coil 108x on the opposite side from the a3 and b2 terminals, the connecting conductors 111x and 112x are made as short as possible, thus avoiding crossing between them.

[0253] Figure 20 shows the structure of Embodiment 3 of the three-phase five-legged reactor according to the present invention, as represented in third-angle projection, in front view, left side view, rear view, and top view.

[0254] The components numbered the same as those in Figures 18 and 19 have the same content, and their explanations are omitted to avoid duplication.

[0255] 113r and 113s are auxiliary yokes made of an iron core filled with a non-magnetic material and containing no voids. 113r fastens from the space between windings 106u and 109u of the wound core 101u to the space between windings 106v and 109v of the wound core 101v, and 113s fastens from the space between windings 106w and 109w of the wound core 101w to the space between windings 106v and 109v of the wound core 101v.

[0256] The core cross-sectional areas of the auxiliary yokes 113r and 113s are made smaller than the core cross-sectional areas of the yokes 103t and 103b.

[0257] 114r and 114s are auxiliary yokes made of an iron core filled with a non-magnetic material and containing no voids. 114r fastens from the winding 106u and winding 109u of the wound core 101u to the side leg 102r, and 114s fastens from the winding 106w and winding 109w of the wound core 101w to the side leg 102s.

[0258] The core cross-sectional areas of the auxiliary yokes 114r and 114s are made smaller than the core cross-sectional areas of the yokes 103t and 103b.

[0259] By setting the cross-sectional areas of the above auxiliary yokes (113r, 113s, 114r, 114s) so that magnetic saturation does not occur even when magnetic flux flows due to AC overcurrent during AC system fault propagation, it is possible to suppress the three-phase imbalance of AC inductance values ​​and maintain stable operation even when an AC system fault propagation occurs.

[0260] Furthermore, by making the cross-sectional area of ​​the auxiliary yokes (113r, 113s, 114r, 114s) smaller than the core cross-sectional area of ​​yokes 103t and 103b, the elongation of the wound core (101u, 101v, 101w) is minimized, resulting in a smaller and lighter design.

[0261] Furthermore, since the division of the winding core (101u, 101v, 101w) can be avoided, there is an effect of suppressing the decrease in DC inductance during steady-state operation and magnetic saturation during AC system fault propagation.

[0262] Figure 21 shows the structure of Embodiment 4 of the three-phase five-legged reactor according to the present invention, as represented in third-angle projection, in front view, left side view, rear view, and top view.

[0263] The components numbered the same as those in Figures 18, 19, and 20 are of the same nature, and their explanations are omitted to avoid duplication.

[0264] 115r is a non-magnetic elastic material sandwiched between the connecting surfaces of the auxiliary yoke 114r and the side leg 102r, and 115s is a non-magnetic elastic material sandwiched between the connecting surfaces of the auxiliary yoke 114s and the side leg 102s.

[0265] This has the effect of suppressing vibrations and noise generated in the auxiliary yoke 114r and auxiliary yoke 115s.

[0266] On the other hand, sandwiching these non-magnetic elastic materials 115r and 115s increases the alternating magnetic flux flowing through the wound iron core (101u, 101v, 101w), but the increase in alternating magnetic flux can be suppressed by reducing the thickness.

[0267] Figure 22 shows the structure of Embodiment 5 of the three-phase five-legged reactor according to the present invention, as represented in third-angle projection, in front view, left side view, rear view, and top view.

[0268] The components numbered the same as in Figure 18 have the same content, and their explanations will be omitted to avoid duplication.

[0269] Yoke 103b is a yoke made of an iron core filled with a non-magnetic material and containing no voids, similar to yoke 103b in Figure 18. However, as described later, the cross-sectional area of ​​the iron core is reduced to lower the dimensions and weight.

[0270] Yoke 118t is a yoke made of the same material and with the same dimensions as yoke 103b, and they are in close contact with each other on their backs.

[0271] 116u, 116v, and 116w are wound cores consisting of three iron cores containing multiple voids filled with a non-magnetic material, and are wound with winding 109x, in which x is represented by three phases (u, v, w).

[0272] 117r and 117s are side legs consisting of two iron cores filled with non-magnetic material and containing no voids, and their core axes are positioned on the same plane as the axes of the three wound iron cores (116u, 116v, and 116w).

[0273] 118t and 118b are yokes consisting of two void-free iron cores filled with a non-magnetic material, and their core axes are arranged on the same plane as the axes of the three wound iron cores (116u, 116v, 116w) and the two side legs (117r, 117s).

[0274] The above-mentioned wound cores (116u, 116v, 116w), side legs (117r, 117s), and yokes (118t, 118b) are configured to be symmetrical with the wound cores (101u, 101v, 101w), side legs (102r, 102s), and yokes (103b, 103t) at the joint surface between yoke 118t and yoke 103b.

[0275] The winding 109x consists of a concentrically arranged inner coil 107x and an outer coil 108x, which are magnetically coupled by the winding core 116x.

[0276] In the example in Figure 22, left-handed winding (second quadrant coil) is selected for the inner coil 107x and right-handed winding (fourth quadrant coil) is selected for the outer coil 108x. However, if the inner coil (104x) of winding 106x is left-handed winding (second quadrant coil) and the outer coil (105x) is right-handed winding (fourth quadrant coil), then right-handed winding (first quadrant coil) must be selected for the inner coil 107x and left-handed winding (third quadrant coil) for the outer coil 108x.

[0277] The a2 terminal of the inner coil 107x is brought out to the positive terminal (xp terminal), the a4 terminal of the outer coil 108x is brought out to the negative terminal (xn terminal), the b2 terminal of the inner coil 107x and the a3 terminal of the outer coil 105x are connected by a connecting conductor 111x, and the b4 terminal of the outer coil 108x and the a1 terminal of the inner coil 104x are connected by a connecting conductor 112x.

[0278] By magnetically coupling the windings 106x and 109x described above with the winding core 101x and winding core 116x, the positive inductance (L1-M) for the AC current between the positive terminal (xp terminal) and the intermediate terminal (xc terminal) and the negative inductance (L2-M) for the AC current between the negative terminal (xn terminal) and the intermediate terminal (xc terminal) can be made to the same value.

[0279] Even if the sum of the cross-sectional areas of yoke 118t and yoke 103b is smaller than the cross-sectional areas of yoke 103t and yoke 118b, it is still possible to set it so that yoke 118t and yoke 103b do not become magnetically saturated by the magnetic flux caused by AC overcurrent during the propagation of an AC system fault.

[0280] This has the effect of suppressing three-phase imbalance in AC inductance values ​​and enabling stable operation even when an AC system fault propagates.

[0281] Furthermore, by using non-oriented electrical steel sheets for the yoke irons 118t and 103b, the magnetic resistance to the magnetic flux penetrating the wound core (101x) and wound core (106x) can be reduced. The reduction in magnetic resistance can be offset by increasing the thickness of the yoke irons 118t and 103b, which has the effect of ensuring the strength of the core structure that is divided into two symmetrically.

[0282] Figure 23 shows the structure of Embodiment 6 of the three-phase five-legged reactor according to the present invention, as represented in third-angle projection, in front view, left side view, rear view, and top view.

[0283] The components numbered the same as those in Figures 18 and 19 have the same content, and their explanations are omitted to avoid duplication.

[0284] 119r and 119s are side legs made of an iron core filled with a non-magnetic material and containing no voids, and they fasten the yoke iron 1103t and the yoke iron 103b together.

[0285] The side leg 119r passes through the midpoint between the wound core 101u and the wound core 101v, and the side leg 119s passes through the midpoint between the wound core 101v and the wound core 101w.

[0286] Even if the cross-sectional areas of the side legs 119r and 119s are made smaller than those of the side legs 102r and 102s, it is still possible to set it so that the side legs 119r and 119s do not become magnetically saturated by the magnetic flux caused by the AC overcurrent when an AC system fault propagates.

[0287] This has the effect of suppressing three-phase imbalance in AC inductance values ​​and enabling stable operation even when an AC system fault propagates.

[0288] In the case of the three-phase reactor shown in Figure 23, the wound core (101u, 101v, 101w) and side legs (102r, 102s) are erected on the yoke 103b, and at the same time, the side legs (119r, 119s) are erected, and the windings (106u, 106v, 106w) and windings (109u, 109v, 109w) are fixed to the wound core (101u, 101v, 101w) or the yoke 103b, after which the yoke 103t can be fastened. This has the effect of simplifying the use of jigs for assembling auxiliary yokes, vibration damping measures using non-magnetic elastic materials, and the fastening mechanism of the core structure divided into upper and lower halves.

[0289] 1, 11_1, 11_2 MMC converter 1_1 DC power supply 1_2, 11_3 3-phase AC power supply 11_4 Unit transformer 11_5 Synchronous generator motor 11_6 Pump turbine 11_7 Rotational phase detector 11_8 Control device 11_9 Load switch 17_1, 17_2, 17_3, 17_4, 17_5 Auxiliary yoke 17_6, 17_7 Side legs 1_3 AC sensor 2, 2u, 2v, 2w, 28 Leg circuit 21, 21Px, 21Nx Arm power circuit 22Px, 22Nx Induction element 23, 23Px, 23Nx Current transformer 24, 24Px, 24Nx Arm control device 2_1 Moving average value calculator 2_2, 2_3, 1101u, 1101v, 1101w Gain 3 Unit converter 31, 31H, 31L Self-extinguishing element 32 Capacitor 33, 33H, 33L Diode 34 Capacitor voltage detector 35 Voltage signal converter 36 Gate drive device 4 Power converter control device 401 Arm current calculator 402 Capacitor voltage calculator 403 Capacitor voltage regulator 404 Reactive power regulator 405 Output switch 406 AC voltage regulator 410 Arm voltage command calculator 5 DC current control device 501, 801 Divider 503 DC current regulator 504 Adder 6 AC current control device 7 Circulating current control device 8 Unit converter control device 802 Current code detector 804 Limiter 805 Comparator 806 Carrier output device 9 Positive / Negative Balance Control Device 901 Leg-to-Leg Balance Control Device 101u, 101v, 101w, 116u, 116v, 116w Winding Core 102r, 102s, 117r, 117s, 119r, 119s Side Legs 103t, 103b, 118t, 118b Yoke 104u, 104v, 104w, 107u, 107v, 107w Inner Coil 105u, 105v, 105w, 108u, 108v, 108w Outer Coil 106u, 106v, 106w, 109u, 109v, 109w Winding 110u, 110v, 110w Branch Conductor 111u, 111v, 111w, 112u, 112v, 112w: Connecting conductors; 113r, 113s, 114r, 114s: Auxiliary yokes; 115r, 115s: Non-magnetic material

Claims

1. A modular multilevel power converter (1) connected between the positive voltage terminal (P terminal) and negative voltage terminal (N terminal) and three-phase AC voltage terminals (u terminal, v terminal, w terminal) of a DC power supply, wherein the modular multilevel power converter (1) comprises a power converter control device (4) and three three-terminal leg circuits (2u, 2v, 2w), wherein the positive terminals (Pu, Pv, Pw) of the three-terminal leg circuits (2u, 2v, 2w) are connected in a star configuration to the positive voltage terminal (P terminal) of the DC power supply, the negative terminals (Nu, Nv, Nw) are connected in a star configuration to the negative voltage terminal (N terminal) of the DC power supply, and the intermediate terminals (ACu, ACv, ACw) of the three-terminal leg circuits (2u, 2v, 2w) are connected to the three-phase AC voltage terminals (u terminal, v terminal, w terminal), The three-terminal leg circuit (2x), in which x is represented as a three-phase (u, v, w), comprises a two-terminal positive arm power circuit (21Px), a two-terminal negative arm power circuit (21Nx), a three-terminal inductive element (22x), a positive current transformer (23Px), and a negative current transformer (23Nx), with the positive terminal (APx) of the positive arm power circuit (21Px) connected to the positive terminal (Px) of the three-terminal leg circuit (2x), and the negative terminal (BPx) of the positive arm power circuit (21Px) connected to the first terminal of the three-terminal inductive element (22x). A three-terminal leg circuit (2x) is provided with an energy storage average voltage calculation circuit (2_1), and it connects terminal (xp) to the intermediate terminal (xc) of the three-terminal inductive element (22x) to the intermediate terminal (ACx) of the three-terminal leg circuit (2x), connects the second terminal (xn) of the three-terminal inductive element (22x) to the positive terminal (ANx) of the negative arm power circuit (21Nx), connects the negative terminal (BNx) of the negative arm power circuit (21Nx) to the negative terminal (Nx) of the three-terminal leg circuit (2x), and connects terminal (xp) to the intermediate terminal (xc) of the three-terminal inductive element (22x) to the intermediate terminal (ACx) of the three-terminal leg circuit (2x), and connects terminal (xn) of the three-terminal inductive element (22x) to the positive terminal (ANx) of the negative arm power circuit (21Nx), and connects the negative terminal (BNx) of the negative arm power circuit (21Nx) to the negative terminal (Nx) of the three-terminal leg circuit (2x), The positive current transformers (23Pu, 23Pv, 23Pw) measure and output the current (I_up, I_vp, I_wp) of the positive arm power circuit (21Pu, 21Pv, 21Pw) to the power converter control device (4), and the negative current transformers (23Nu, 23Nv, 23Nw) measure and output the current (I_un, I_vn, I_wn) of the negative arm power circuit (21Nu, 21Nv, 21Nw) to the power converter control device (4).The positive arm power circuit (21Px) and the negative arm power circuit (21Nx), in which x is represented by a three-phase (u, v, w) notation, are series circuits consisting of K (K is a natural number of 2 or more) unit converters (3) each having one energy storage device (32) and at least two self-extinguishing power semiconductor elements (31H, 31L), wherein the unit converters (3) receive PWM-modulated gate (GH, GL) commands to the self-extinguishing power semiconductor elements (31H, 31L) output from the power converter control device (4), output a required terminal voltage with zero as the lower limit and the voltage of the energy storage device (32) as the upper limit, and output a voltage measurement signal (Vc) of the energy storage device (32) to the energy storage device average voltage calculation circuit (2_1), The energy storage average voltage calculation circuit (2_1) is an energy storage average voltage calculation circuit (2_1) that outputs the average voltage (Vc_xp) of the K energy storage units (32) of the positive arm power circuit (21Px) and the average voltage (Vc_xn) of the K energy storage units (32) of the negative arm power circuit (21Nx) to the power converter control device (4), and the three-terminal inductive elements (22u, 22v, 22w) are three-terminal inductive elements (22u, 22v, 22w) that have a function to suppress circulating current flowing between the three-terminal leg circuits (2u, 2v, 2w) when the voltage ratio between the three-terminal leg circuits (2u, 2v, 2w) is unbalanced, The power converter control device (4) receives a DC voltage command (Vrf_dc) from the DC current control device (5), outputs a DC current calculation value (I_dc) to the DC current control device (5), and includes a function to suppress circulating current flowing between the three terminal leg circuits (2u, 2v, 2w) when the voltage ratio between the three terminal leg circuits (2u, 2v, 2w) is unbalanced, and the DC current control device (5) is a DC current control device (5) that outputs a DC voltage command (Vrf_dcI_dc) such that the DC current command value and the DC current calculation value (I_dc) match,The three-terminal inductive element (22u, 22v, 22w) comprises a first winding core (101u), a second winding core (101v), and a third winding core (101w) arranged in a permutation order for each phase. These three winding cores (101u, 101v, 101w) consist of a core containing multiple gaps filled with a non-magnetic material. A first side leg (102r) is arranged in a permutation order on the first winding core (101u) side and a third side leg (102r) is arranged in a permutation order on the third winding core (101w) side. It comprises a second side leg (102s) positioned thereon, and these two side legs (102r, 102s) consist of a gapless core filled with a non-magnetic material, and an upper yoke (103t) and a lower yoke (103b) that magnetically connect the winding iron core (101u, 101v, 101w) and the side legs (102r, 102s), and these two yokes (103t, 103b) consist of a gapless core filled with a non-magnetic material, The aforementioned winding core (101u, 101v, 101w) is provided with two coils around the winding core for each phase, and the first coil (104u, 104v, 104w) and the second coil (105u, 105v, 105w) are wound concentrically to form a winding (106u, 106v, 106w), and the winding (106x), in which x is represented by the three phases (u, v, w), is a winding (106x) that is wound in a direction such that the magnetomotive force on the winding core (101x) cancels out due to the through-current that passes through the first terminal (xp) and the second terminal (xn) of the three-terminal inductive element (22x), and the branch current that splits from the intermediate terminal (xc) to the first terminal (xp) and the second terminal (xn). A modular multilevel power converter (1) is characterized by having a three-terminal inductive element (22u, 22v, 22w) in which the core cross-sectional areas of the first and second side legs (102r, 102s) are larger than the core cross-sectional areas of the three wound cores (101u, 101v, 101w), and the core cross-sectional area of ​​the yoke (103t, 103b) is larger than the core cross-sectional areas of the three wound cores (101u, 101v, 101w).

2. In the modular multilevel power converter (1) according to claim 1, the three wound cores (101u, 101v, 101w) of the three terminal inductive elements (22u, 22v, 22w) are provided with a third coil (107u, 107v, 107w) consisting of an inner diameter opposite direction screw winding with the same number of turns as the first coil (104u, 104v, 104w), and a fourth coil (108u The third coil (107u, 107v, 107w) and the fourth coil (108u, 108v, 108w) are wound concentrically to form a second winding (109u, 109v, 109w), and the winding (106u, 106v, 106w) and the second winding (109u, 109v, 109w) are coaxially arranged with respect to the winding core (101u, 101v, 101w). The first coil (104u, 104v, 104w) is wound in the forward direction on the inner diameter side, the second coil (105u, 105v, 105w) is wound in the opposite direction on the outer diameter side, and the device includes branch conductors (110u, 110v, 110w) that connect the first terminal of the first coil (104u, 104v, 104w) to the first terminal of the second coil (105u, 105v, 105w) individually to each phase, and the device also includes first pole connecting conductors (111u, 111v, 111w) that connect the first terminal of the third coil (107u, 107v, 107w) to the second terminal of the second coil (105u, 105v, 105w) individually to each phase, The device includes a second pole connecting conductor (112u, 112v, 112w) that connects the first terminal of the fourth coil (108u, 108v, 108w) and the second terminal of the first coil (104u, 104v, 104w) individually for each phase, The second terminal of the third coil (107u, 107v, 107w) is connected to the first terminals (up, vp, wp) of the three-terminal inductor element (22u, 22v, 22w) phase by phase, the second terminal of the fourth coil (108u, 108v, 108w) is connected to the second terminals (un, vn, wn) of the three-terminal inductor element (22u, 22v, 22w) phase by phase, and the branch conductors (110u, 110v, 110w) are connected to the intermediate terminals (uc, vc, wc) of the three-terminal inductor element (22u, 22v, 22w) phase by phase.A modular multilevel power converter (1) is characterized by comprising a three-terminal inductive element (22u, 22v, 22w) consisting of a winding (106x) and a second winding (109x) wound in a direction in which the magnetomotive force due to a through-current passing through the first terminal (xp) and second terminal (xn) of the three-terminal inductive element (22x), in which x is represented as three phases (u, v, w), excites the wound core (101x), and the magnetomotive force due to branch currents that split from the intermediate terminal (xc) to the first terminal (xp) and the second terminal (xn) demagnetizes the wound core (101x).

3. A modular multilevel power converter (1) according to claim 1 or claim 2, wherein the first auxiliary yoke (113r) connects the first winding core (101u) and the second winding core (101v), the first auxiliary yoke (113r) starts between the winding (106u) and the second winding (109u) that winds the first winding core (101u), and ends between the winding (106v) and the second winding (109v) that winds the second winding core (101v), A second auxiliary yoke (113s) connecting the third winding core (101w) and the second winding core (101v), comprising a second auxiliary yoke (113s) that starts between the winding (106w) and the second winding (109w) around which the third winding core (101w) is wound, and ends between the winding (106v) and the second winding (109v) around which the second winding core (101v) is wound, wherein the first and second auxiliary yokes (113r, 113s) consist of a core filled with a non-magnetic material and containing no gaps. A third auxiliary yoke (114r) connecting the first winding core (101u) and the first side leg (102r), the third auxiliary yoke (114r) starting from the point between the winding (106u) and the second winding (109u) that winds around the first winding core (101u) and ending at the first side leg (102r), A modular multilevel power converter (1) comprising a fourth auxiliary yoke (114s) connecting the third winding core (101w) and the second side leg (102s), the fourth auxiliary yoke (114s) starting between the winding (106w) that winds the third winding core (101w) and the second winding (109w), and ending at the second side leg (102s), wherein the third and fourth auxiliary yokes (114r, 114s) are equipped with three-terminal inductive elements (22u, 22v, 22w) made of a gap-free core filled with a non-magnetic material.

4. The modular multilevel power converter (1) according to claim 1, comprising a fourth winding core (116u), a fifth winding core (116v), and a sixth winding core (116w) arranged sequentially for each phase, wherein these three winding cores (116u, 116v, 116w) consist of a core containing a plurality of gaps filled with a non-magnetic material, and a third side leg (117r) arranged sequentially on the fourth winding core (116u) side and a fourth side leg (117s) arranged sequentially on the sixth winding core (116w) side The structure is provided with the following: these two side legs (117r, 117s) consist of a gapless core filled with a non-magnetic material, and the fourth to sixth winding core (116u, 116v, 116w) and the third and fourth side legs (117r, 117s) are magnetically connected by an upper yoke (118t) and a lower yoke (118b), these two yokes (118t, 118b) consist of a gapless core filled with a non-magnetic material. The first winding core (101u) and the fourth winding core (116u) are arranged coaxially, the second winding core (101v) and the fifth winding core (116v) are arranged coaxially, the third winding core (101w) and the sixth winding core (116w) are arranged coaxially, the first side leg (102r) and the third side leg (117r) are arranged coaxially, the second side leg (102s) and the fourth side leg (117s) are arranged coaxially, and the bottom surface of the lower yoke (103b) and the top surface of the upper yoke (118t) are in close contact with each other. The device comprises a third coil (107u, 107v, 107w) consisting of an inner diameter side opposite direction screw winding with the same number of turns as the first coil (104u, 104v, 104w), and a fourth coil (108u, 108v, 108w) consisting of an outer diameter side forward direction screw winding with the same number of turns as the second coil (105u, 105v, 105w), and the third coil (107u, 107v, 107w) and the fourth coil (108u, 108v, 108w) are wound concentrically to form a second winding (109u, 109v, 109w), and the winding (106u, 106v, 106w) and the second winding (109u, 109v, 109w) are coaxially arranged with respect to the winding core (101u, 101v, 101w).The first coil (104u, 104v, 104w) is wound in the forward direction on the inner diameter side, the second coil (105u, 105v, 105w) is wound in the opposite direction on the outer diameter side, and the device includes branch conductors (110u, 110v, 110w) that connect the first terminal of the first coil (104u, 104v, 104w) to the first terminal of the second coil (105u, 105v, 105w) individually to each phase, and the device also includes first pole connecting conductors (111u, 111v, 111w) that connect the first terminal of the third coil (107u, 107v, 107w) to the second terminal of the second coil (105u, 105v, 105w) individually to each phase, The device includes a second pole connecting conductor (112u, 112v, 112w) that connects the first terminal of the fourth coil (108u, 108v, 108w) and the second terminal of the first coil (104u, 104v, 104w) individually for each phase, The second terminal of the third coil (107u, 107v, 107w) is connected to the first terminals (up, vp, wp) of the three-terminal inductor element (22u, 22v, 22w) phase by phase, the second terminal of the fourth coil (108u, 108v, 108w) is connected to the second terminals (un, vn, wn) of the three-terminal inductor element (22u, 22v, 22w) phase by phase, and the branch conductors (110u, 110v, 110w) are connected to the intermediate terminals (uc, vc, wc) of the three-terminal inductor element (22u, 22v, 22w) phase by phase. The winding (106x) and the second winding (109x), in which x is represented by three phases (u, v, w), are windings (106x) and the second winding (109x) that are wound in a direction such that the magnetomotive forces on the winding core (101x) and the winding core (116x) cancel each other out due to the branch currents that split from the intermediate terminal (xc) to the first terminal (xp) and the second terminal (xn), respectively, where x is represented by three phases (u, v, w). A modular multilevel power converter (1) is characterized by having a three-terminal inductive element (22u, 22v, 22w) in which the combined core cross-sectional area of ​​the lower yoke (103b) and the upper yoke (118t) is smaller than the core cross-sectional area of ​​the upper yoke (103t) and the lower yoke (118b).

5. A modular multilevel power converter (1) according to claim 1 or claim 2, comprising: a fifth side leg (119r) starting from the midpoint between the winding core (101u) and the winding core (101v) that are surface-connected to the upper yoke (103t), and ending at the midpoint between the winding core (101u) and the winding core (101v) that are surface-connected to the lower yoke (103b); and a sixth side leg (119s) starting from the midpoint between the winding core (101v) and the winding core (101w) that are surface-connected to the upper yoke (103t), and ending at the midpoint between the winding core (101v) and the winding core (101w) that are surface-connected to the lower yoke (103b), A modular multilevel power converter (1) is characterized by having a three-terminal inductive element (22u, 22v, 22w) in which the core cross-sectional areas of the fifth leg (119r) and the sixth leg (119s) are smaller than the core cross-sectional area of ​​the wound iron core (101u, 101v, 101w).

Images