A matrix topology based hvdc power supply system and a selective protection method thereof

Abstract

The application relates to the technical field of power supply, in particular to an HVDC power supply system based on a matrix topology and a selective protection method and system thereof, the system comprising M AC / DC power units, M gating switches, N distributed DC sub-buses and a controller, and forming an M*N matrix topology. The gating switch has four positions of pre-charging, switching on, discharging and grounding, and N groups of load static contacts in the switching-on position are connected to the N sub-buses respectively. When a fault occurs, the power electronic switch in the power unit cuts off the fault current in microseconds, then the switch is switched to the discharging position to discharge residual charge and switched to the grounding position to be physically isolated. The application realizes accurate fault isolation, dynamic recombination, multi-module parallel connection and N+X redundancy, greatly reduces the cost of downstream equipment, and is suitable for high-power-density scenes such as intelligent computing centers.

Description

Technical Field

[0001] This invention relates to the field of power supply technology, specifically to an HVDC power supply system based on matrix topology and its selective protection method and system. Background Technology

[0002] In the field of AIDC (AI Data Center), with the continuous increase in the power consumption of computing chips, the power of a single rack has jumped from the traditional tens of kilowatts to 100kW or even over 130kW, placing higher demands on the efficiency and reliability of the power supply system. In the data center power supply field, HVDC (High-Voltage Direct Current) specifically refers to DC power supply systems with an output voltage of 240V and above. As a new generation power supply solution to replace traditional AC UPS, it significantly improves energy efficiency by reducing AC-DC conversion links and has become the mainstream choice for AI data centers. Currently, the mainstream HVDC voltage levels include 240V / 336V (widely deployed in domestic data centers) and ±400V / 800V (the third-generation solution supporting high AI computing power). Among them, the 800V system is considered an inevitable future trend due to its high efficiency and low copper consumption advantages.

[0003] In existing HVDC power supply systems, multiple AC / DC power units are typically connected to a common DC bus in parallel or series, and then branched from the bus into several branches. Each branch provides independent protection for its load through a circuit breaker or fuse. Taking Huawei Technologies Co., Ltd.'s invention patent US10177561B2 as an example, this solution provides a selective protection circuit for HVDC power supply systems. This solution represents the current technical content of single bus and branch protection in HVDC systems, and therefore still has the following shortcomings: (a) Inherent defects of DC systems. The current in a DC system is a flat and constant value, unlike that in an AC system which naturally crosses zero every half cycle. When the circuit breaker contacts open, the arc continues to lengthen and is difficult to extinguish, making arc extinguishing difficult. The requirements for the withstand voltage of the break and the arc-extinguishing chamber are much higher than those for AC switches. This presents a natural technical obstacle to fault protection in DC systems.

[0004] (ii) Protection devices suffer from performance bottlenecks or excessively high costs. Currently, the main implementation methods are as follows: DC circuit breaker design: Contact opening time is typically 30~40 seconds. ms This requires the upstream power supply to have a stronger fault current withstand capability, which leads to a sharp increase in the cost of the power supply section.

[0005] Fast-acting fuse solution: cutting speed can be less than 10 ms However, fuses are disposable consumables that need to be replaced when the power is off, and they cannot be remotely controlled or self-reset.

[0006] Solid-state circuit breaker solution: Using components such as IGBTs can shorten the response time to the microsecond level and effectively avoid arc flash, but it has high on-resistance, requires fan cooling, is large in size, lacks reliable mechanical circuit breaking isolation, and has high cost.

[0007] Solid-state circuit breakers, parallel switches, and series switches combined solutions can solve the problems of conduction resistance and electrical isolation, but they are large in size, high in cost, complex in structure, and have poor reliability.

[0008] (iii) Fault handling is difficult in a series topology. For example... Figure 1 In the series connection shown, multiple SST power units are connected in series to the medium-voltage AC side, such as 10kV. When one power unit fails, it needs to be bypassed and short-circuited via a disconnect switch. At this time, due to the reduced number of units in series, each of the remaining units must bear a higher voltage. For example, if 10 units are connected in series and each shares 1kV, after one unit is short-circuited, the remaining 9 units must each share approximately 1.111kV. The units operate under overvoltage conditions for extended periods, accelerating component aging. Replacing the faulty unit requires completely disconnecting the entire system, severely impacting power supply continuity.

[0009] (iv) Defects in the protection logic. The existing protection logic is downstream detection and downstream disconnection, meaning that the circuit breaker or fuse in the branch circuit where the fault occurs disconnects the connection to the power supply. However, when a short-circuit fault occurs, the fault current will conduct throughout the entire circuit at a physical speed close to the speed of light, rising to a peak value that could damage the power supply and equipment within milliseconds. The tripping time of conventional DC circuit breakers is usually between tens of milliseconds, which is insufficient to disconnect the current before it reaches a destructive level. The fault current often needs to be disconnected by the protection components of the upstream power supply section, resulting in an expanded power outage area.

[0010] Even if solid-state circuit breakers are used to shorten the breaking time to 5-50 microseconds, the cost of configuring solid-state circuit breakers for each branch circuit is extremely high, and their heat generation and heat dissipation problems further reduce the reliability of the system.

[0011] (v) The single busbar architecture results in a wide range of fault impacts. In the traditional single main busbar architecture, all power units are connected in parallel to a common busbar. When a fault occurs in a branch, if the protection device of that branch fails to act in time, the fault current will directly affect the voltage of the entire busbar, causing other load branches connected to the same busbar to also suffer voltage drops or even power outages. The fault isolation granularity is coarse, and precise grid-like isolation cannot be achieved. Summary of the Invention

[0012] Purpose of the invention: To address the shortcomings of the prior art, this invention provides an HVDC power supply system based on matrix topology and its selective protection method.

[0013] Technical solution: According to a first aspect of the present invention, an HVDC power supply system based on matrix topology is provided, the system comprising: M power units are used to convert the medium-voltage AC power provided by the external medium-voltage AC bus into DC power. The input terminals of the M power units are connected in parallel to the external medium-voltage AC bus. N DC busbars, one end of each DC busbar is electrically connected to the corresponding load; M single-pole multi-throw (SPMD) switches are used, each SPMD switch including one moving contact and N load stationary contacts. The output terminal of the power unit is connected to the moving contact, and the other end of the DC bus is connected to the N sets of load stationary contacts of each SPMD switch. This allows any one of the M power units to be selectively connected to any one of the N loads through its corresponding SPMD switch, thus forming an M×N matrix topology. ; The controller is connected to the control terminal of each of the single-pole multi-throw switches and is used to control the moving contact of each of the single-pole multi-throw switches to selectively close with a corresponding set of load stationary contacts to establish or disconnect the power supply path between the corresponding power unit and the load.

[0014] Furthermore, including: The single-pole multi-throw switch also includes a pre-charge stationary contact, a discharge stationary contact, and a grounding stationary contact disposed on the switch base; The positive terminals of each of the DC busbars are connected in series with pre-charge resistors and then connected together to the pre-charge stationary contact. When the moving contact is rotated to the pre-charge position, the DC output terminal of the power unit forms a parallel pre-charge circuit with each of the distributed DC busbars via the moving contact, the pre-charge stationary contact and each of the pre-charge resistors. Each of the DC busbars is connected in series with a discharge resistor at its positive terminal and then connected to the discharge stationary contact. When the moving contact is rotated to the discharge position, each of the DC busbars forms a parallel discharge circuit with the DC output terminal of the power unit via the corresponding discharge resistor, the discharge stationary contact, and the moving contact. The positive terminals of each of the DC busbars are connected to the grounding stationary contact, which is connected to the protective ground. When the moving contact is rotated to the grounding position, the DC output terminal of the power unit is grounded via the moving contact and the grounding stationary contact.

[0015] Furthermore, including: The positive terminal of each of the DC busbars is also connected to N independent load stationary contacts located in the ON position of the switch, which are used to form the main power supply circuit of the corresponding load when the moving contact is rotated to the ON position.

[0016] Furthermore, including: The number of power units M is greater than or equal to the number of DC busbars N.

[0017] Secondly, the present invention also provides a selective protection method for an HVDC power supply system based on matrix topology, the method comprising the following steps: S1: Real-time detection of the output current of each power unit; S2: When the output current of the target power unit is detected to exceed the preset threshold, the power electronic switching device inside the target power unit is controlled to perform a shutdown action to cut off the fault current within microseconds. S3: After the fault current returns to zero, control the moving contact of the single-pole multi-throw switch corresponding to the target power unit to switch from the on position to the discharge position or the ground position, so that a physical isolation break is formed between the target power unit and the load.

[0018] Furthermore, the method also includes: in step S3, after the moving contact is switched to the discharge position, each DC busbar forms a discharge circuit through the corresponding discharge resistor, the discharge stationary contact and the moving contact, so as to discharge the residual charge in the target power unit and the corresponding DC busbar.

[0019] Furthermore, including: Following step S3, the following is also included: S4: When it is determined that another power unit needs to take over the power supply to the load from the target power unit that has been turned off, the moving contact of the single-pole multi-throw switch corresponding to the other power unit is controlled to perform the following operations in sequence: First, switch to the pre-charge position, and perform buffer charging on the corresponding DC busbar through the pre-charge resistor. After the voltage is matched, switch to the on position and switch to the load stationary contact position that is electrically connected to the target load, thereby completing the dynamic switching of the power supply path.

[0020] Furthermore, including: After step S4, the method further includes: S5: When the power demand of any load exceeds the rated output power of a single power unit, control the moving contacts of the single-pole multi-throw switches corresponding to two or more power units to switch to the ON position simultaneously and lock them to the same load stationary contact electrically connected to the load, thereby enabling multiple power units to supply power to the load in parallel.

[0021] Furthermore, including: In step S5, before controlling the moving contacts of the single-pole multi-throw switches corresponding to two or more power units to simultaneously switch to the ON position, the method further includes: Each of the moving contacts is controlled to switch to the pre-charge position first, and the same distributed DC busbar is buffered and charged through the pre-charge resistor; The voltage of each moving contact in the circuit is detected. Once the voltage matches the voltage of the distributed DC busbar, each moving contact is controlled to switch to the ON position simultaneously and locked to the same set of load stationary contacts that are electrically connected to the distributed DC busbar.

[0022] Furthermore, the method also includes: S6: When the power demand of the target load drops below the rated output power of a single power unit, the moving contact of the single-pole multi-throw switch corresponding to at least one power unit in the parallel power units is controlled to exit the same group of load stationary contacts and switch to the load stationary contact electrically connected to other loads or switch to the discharge position, thereby realizing the conversion from parallel power supply to independent power supply.

[0023] Thirdly, the present invention also provides an electronic device, comprising: a processor, a storage medium, and a bus, wherein the storage medium stores machine-readable instructions executable by the processor, and when the electronic device is in operation, the processor communicates with the storage medium via the bus, and the processor executes the machine-readable instructions to perform the steps of the method as described in the second aspect.

[0024] Fourthly, the present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the method described in the second aspect.

[0025] Beneficial effects: Compared with the prior art, the present invention has the following advantages: This invention splits the output of a power unit into multiple branches, each with an independent switching switch. Multiple independent distributed DC busbars are correspondingly set on the load side, forming an M×N matrix topology between the power unit and the load. A single-pole multi-throw (SPMD) switch enables flexible switching of the power unit to any load. Any power unit can be selectively connected to any load via its corresponding SPMD switch, allowing for dynamic reconfiguration of the relationship between power units and loads based on load requirements. This achieves optimal path and load balancing, and allows multiple power units to be connected in parallel to supply power to the same load to meet high power demands, or for other healthy units to automatically take over when one power unit fails.

[0026] Meanwhile, this invention changes the traditional downstream protection logic. When a fault occurs in a branch circuit, the power electronic components inside the corresponding power unit directly cut off the fault current in microseconds. After the current returns to zero, the switch is used to switch to the discharge position or the ground position to achieve physical isolation. This two-level collaborative protection mechanism of electronic cut-off followed by mechanical isolation not only achieves a fault response speed in microseconds, but also avoids the need to configure expensive solid-state circuit breakers in downstream branches.

[0027] Compared to existing technologies, this invention shifts the protection responsibility from downstream distributed branch protection devices to the electronic switches of the upstream power unit itself. At the system architecture level, it fundamentally solves the contradiction between protection speed and selectivity in DC systems, achieves precise isolation and minimization of fault areas, significantly improves the reliability of HVDC power supply systems, and reduces the cost of downstream protection devices and system complexity. It is especially suitable for scenarios with extremely high requirements for power supply continuity and power density, such as intelligent computing centers.

[0028] In summary, this invention, through the combination of distributed DC busbars, matrix topology, single-pole multi-throw switches, and power unit self-disconnection technologies, brings substantial technical improvements to the protection speed, isolation accuracy, cost control, operation and maintenance convenience, and scenario adaptability of HVDC power supply systems. Moreover, the effects of each level in this application are not simply superimposed. Matrix topology provides the physical basis for precise isolation, power unit self-disconnection provides the execution means for rapid protection, and single-pole multi-throw switches provide the implementation path for dynamic reconfiguration and online maintenance. The three work together to form a complete technical solution that combines speed, accuracy, economy, and flexibility. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of the structural principle of an existing DC system, HVDC. Figure 2 This is a topology diagram of an HVDC power supply system based on matrix topology provided in Embodiment 1 of the present invention; Figure 3 This is a schematic diagram of the single-pole multi-throw switch according to Embodiment 1 of the present invention; Figure 4 This is a schematic diagram of the first combination structure between the power unit and the load according to Embodiment 1 of the present invention; Figure 5 This is a schematic diagram of the second combination structure between the power unit and the load as described in Embodiment 1 of the present invention; Figure 6 This is a schematic diagram of the third combination structure between the power unit and the load as described in Embodiment 1 of the present invention; Figure 7 This is the fault current curve diagram described in Embodiment 1 of the present invention; Figure 8 This is a flowchart of the selective protection method described in Embodiment 2 of the present invention. Detailed Implementation

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] Example 1: As Figure 2 As shown, this embodiment provides an HVDC power supply system based on matrix topology. The system includes power units, DC busbars, single-pole multi-throw switches, and a controller. Specifically: M power units are used to convert the medium-voltage AC power supplied by the external medium-voltage AC bus into DC power. The input terminals of the M power units are connected in parallel to the external medium-voltage AC bus. In this embodiment, the power units are AC / DC power units, and their core function is to convert AC power into DC power. Their internal structure includes the following modules: (1) Input rectification and EMI filtering stage: It converts AC power into pulsating DC power through a rectifier bridge and smooths it through a filter capacitor, while suppressing electromagnetic interference through an EMI filter; (2) Power factor correction (PFC) stage: It improves the power factor and reduces harmonic pollution through high-frequency switching transistors IGBT or SiC MOSFET and PFC control chip; (3) DC / DC isolation conversion stage: It achieves electrical isolation and converts DC power into a stable voltage required by the load through primary-side switching transistors (IGBT or SiC MOSFET), high-frequency transformer, secondary-side rectifier and output filter circuit; (4) Output disconnect switch: It sets a solid-state circuit breaker or electronic contactor at the DC output of the power unit to completely isolate the power unit from the distributed DC bus when a serious fault is detected. Among them, the switching transistors of the PFC stage, the primary-side switching transistors of the DC / DC conversion stage and the output disconnect switch are all controllable power electronic switching devices, and their turn-off speed can reach the microsecond level. When the controller detects a fault current, it can cut off the output of the power unit within microseconds by turning off the drive signals of these power electronic switching devices, thus achieving rapid fault isolation.

[0032] Therefore, the power electronic switching devices inside the power unit refer to the IGBT / SiCMOSFET switches in the PFC stage and DC / DC conversion stage, as well as the solid-state circuit breakers / electronic contactors at the output of the power unit. These are components that are inherent in the power unit itself, not external protection devices, and require no additional cost. This two-stage collaborative mechanism of electronic disconnection followed by mechanical isolation not only achieves the ultra-fast response speed of the solid-state circuit breaker level to protect equipment safety, but also provides reliable physical isolation through the mechanical switch contacts to ensure operation and maintenance safety. Furthermore, it eliminates the need to configure expensive solid-state circuit breakers in each downstream branch, overcoming the inherent defects of solid-state circuit breakers, such as high conduction losses, the need for heat dissipation, and the lack of reliable mechanical isolation.

[0033] There are N DC busbars, one end of which is electrically connected to the corresponding load. In this embodiment, the DC busbars are multiple DC buses that are independently led out from the output of the AC / DC power unit via a single-pole multi-throw switch and directly connected to the corresponding load. Each busbar is an independent power distribution channel, supplying power to a single load only, or to a single load after a simple busbar connection. The busbars are physically isolated from each other.

[0034] Specifically, it is installed between the output side of the single-pole multi-throw switch and the input side of the load, as shown in the attached figure. Figure 2 As shown, M AC / DC power units are independently connected to N loads in a one-to-one manner. Each power unit selectively connects to an independent distributed DC busbar via its corresponding single-pole multi-throw switch, supplying power to one load. For example, if the rated output power of each AC / DC power unit is 100kW, then the maximum available power for each load is 100kW. Therefore, when the load power demand is ≤100kW, the attached... Figure 4 That's acceptable.

[0035] Therefore, in this embodiment, N distributed DC busbars are independently led out, each busbar connected only to its corresponding single load, and the busbars are physically isolated from each other. When a branch fails, only the power unit corresponding to that branch performs electronic disconnection and switch isolation, completely breaking the connection between the faulty power unit and its corresponding busbar, without affecting the normal power supply of other power units to the loads on their respective busbars. In contrast, in a traditional single-busbar architecture, a short circuit in any branch can cause a voltage drop on the busbar, affecting all loads connected to that busbar. This invention fundamentally eliminates the systemic risk of a single fault causing a large-scale impact, precisely limiting the power outage area to within the fault grid.

[0036] M single-pole multi-throw (SPMD) switches are used, each SPMD switch including one moving contact and N load stationary contacts. The output terminal of the power unit is connected to the moving contact. Therefore, the number of SPMD switches is the same as the number of power units. The other end of the DC bus is connected to the N sets of load stationary contacts of each SPMD switch, so that any power unit among the M power units can be selectively connected to any one of the N loads through its corresponding SPMD switch, thus forming an M×N matrix topology. .

[0037] This embodiment breaks through the limitations of traditional single main bus or one-to-one fixed wiring architectures, constructing a complete M×N cross switch matrix by connecting M AC / DC power units and N loads through M single-pole multi-throw switches and N distributed DC sub-buses. In this matrix: M power units form rows and N distributed DC busbars and their corresponding loads form columns. At each intersection of the rows and columns, there is an optional electrical connection node, which is the connection between a set of load stationary contacts of a single-pole multi-throw switch and the corresponding busbar. Therefore, by instructing the moving contact of each single-pole multi-throw switch to selectively close with the corresponding stationary contact, the controller can connect any power unit to any load at any time. Multiple intersections on the same column can be closed simultaneously, meaning multiple power units can be connected in parallel to supply power to the same load. However, only one intersection on the same row can be closed at any given time, meaning a power unit can only be connected to one load at any given time. This is determined by the physical structure of a single-pole multi-throw switch, which has one moving contact and N stationary contacts.

[0038] In this embodiment, the matrix topology provides a unified physical platform for advanced power supply strategies such as dynamic reconfiguration, precise isolation, parallel expansion, and redundancy replacement. This gives the system architecture of the present invention inherent flexibility and scalability. That is, M and N can be flexibly configured according to actual scenarios, there is no fixed physical binding relationship between power units and loads, and all connection relationships are defined by software and dynamically scheduled.

[0039] In this embodiment, the number of power units M is greater than or equal to the number of DC busbars N. This is because each single-pole multi-throw switch can only close with one set of stationary contacts at the same time, that is, each power unit can only supply power to one load at the same time. Therefore, M power units can only cover a maximum of M loads at the same time. In other words, if all N loads are to be covered, M ≥ N, which is a necessary condition for achieving N+X redundancy.

[0040] In this embodiment, as Figure 3As shown, a single-pole multi-throw switch physically consists of a rotatable common moving contact and multiple sets of load stationary contacts distributed in specific angular regions on the switch base. The common moving contact is located at the center of the switch and can rotate around its axis, always maintaining an electrical connection with the DC output terminal of the AC / DC power unit. The multiple sets of load stationary contacts are distributed in different angular regions of the switch base, with varying numbers of stationary contacts in each region, each connected to a different external circuit. When the contact rotates to a certain angle, it can only close with the load stationary contact in that angular region, forming a unique conductive electrical path.

[0041] In this embodiment, the single-pole multi-throw switch, in addition to the multiple sets of load stationary contacts mentioned above, also includes a pre-charge stationary contact, a discharge stationary contact, and a grounding stationary contact disposed on the switch base. Therefore, the pre-charge position, the discharge position, and the grounding position are three fixed angular regions reached by the moving contact as it rotates on the switch base. When the moving contact rotates to a certain position, it makes mechanical contact and electrically connects with the fixed load stationary contact within that position.

[0042] More specifically, the pre-charge position is a fixed angle area on the switch base. The pre-charge stationary contact is the only set of fixed conductive contacts in this area. When the moving contact rotates into this fixed angle area, it closes with the pre-charge stationary contact. When the module starts, the moving contact briefly stops at this position, limiting the surge current through the pre-charge resistor, and slowly charging the capacitors of each load bus to near the rated voltage, avoiding the large current surge when directly connected. Similarly, the discharge position is another fixed angle area on the switch base. The discharge stationary contact is the only set of fixed conductive contacts in this area. When the moving contact rotates into this angle area, it closes with the pre-charge stationary contact. When the stationary contact closes and the module is out of service, the internal electronic switch turns off first, and the moving contact rotates to this position and stays there briefly. This discharges the residual charge on the module output terminal and the corresponding busbar to a safe level through the discharge resistor, preventing injury from residual voltage. The grounding position is the third fixed angle area on the switch base. The grounding stationary contact is the only set of fixed conductive contacts in this area. When the moving contact rotates into this angle area, it closes with the grounding stationary contact. When the module is under maintenance or out of service for a long time, the moving contact is locked in this position, forcibly grounding the output terminal of the AC / DC module, establishing a reliable mechanical isolation break, and ensuring the safety of maintenance personnel.

[0043] Based on the position of the moving contact, the switch can achieve the following: the positive terminals of each of the DC busbars are connected in series with pre-charge resistors and then connected together to the pre-charge stationary contact. When the moving contact is rotated to the pre-charge position, the DC output terminal of the power unit forms a parallel pre-charge circuit with each of the distributed DC busbars via the moving contact, the pre-charge stationary contact, and each of the pre-charge resistors. Each of the DC busbars is connected in series with a discharge resistor at its positive terminal and then connected to the discharge stationary contact. When the moving contact is rotated to the discharge position, each of the DC busbars forms a parallel discharge circuit with the DC output terminal of the power unit via the corresponding discharge resistor, the discharge stationary contact, and the moving contact. The positive terminals of each of the DC busbars are connected to the grounding stationary contact, which is connected to the protective ground. When the moving contact is rotated to the grounding position, the DC output terminal of the power unit is grounded via the moving contact and the grounding stationary contact.

[0044] To enable the selective connection of any of the M power units to any of the N loads—that is, the dynamic reconfiguration capability of the matrix topology—in this embodiment, the positive terminals of each of the DC busbars are also connected to N independent stationary load contacts located in the ON position of the switch. This forms the main power supply circuit for the corresponding load when the moving contact rotates to the ON position. For example, if a power unit 2 fails, the system needs another healthy power unit to take over and continue supplying power to load 2. If power unit 3 has only one stationary contact in the ON position (i.e., fixedly connected to load 3), then load 3 cannot supply power to load 2, and the fault cannot be recovered.

[0045] If there are N stationary contacts in the energized position of the 3 power units, and the corresponding loads are denoted as 1 to N, the controller can switch the moving contact from the stationary contact corresponding to load 3 to the stationary contact corresponding to load 2 according to the instruction, and seamlessly complete the replacement.

[0046] Therefore, in this embodiment, the system also includes a controller connected to the control terminal of each of the single-pole multi-throw switches, for controlling the moving contact of each of the single-pole multi-throw switches to selectively close with a corresponding set of load stationary contacts, so as to establish or disconnect the power supply path between the corresponding power unit and the load.

[0047] In this embodiment, the controller collects data from each branch in real time using current and voltage sensors, performs fault diagnosis, and reports the system operating status to the upper-level management system. Simultaneously, it receives power scheduling commands from the upper level, i.e., it monitors the output current of each AC / DC power unit in real time. When the current in a branch exceeds a preset threshold, the controller executes two levels of protection within microseconds: First, it sends a shutdown signal to the AC / DC power unit corresponding to the faulty branch, causing its internal IGBT / SiC MOSFET switches to turn off within microseconds, cutting off the fault current; second, after the fault current returns to zero, it controls the single-pole multi-throw switch corresponding to that branch to switch from the ON position to the PAR or ground position, establishing a reliable mechanical isolation barrier. Therefore, this mechanism allows the present invention to achieve a microsecond-level fault response speed without configuring expensive solid-state circuit breakers in each downstream branch.

[0048] Based on the above topology, the controller can issue the following dynamic routing commands according to the real-time power requirements of the load, including: (1) Path selection: The moving contact of each single-pole multi-throw switch selects which load stationary contact to close, thereby realizing the basic function that any power unit can be connected to any load.

[0049] (2) Path switching: When it is necessary to change the correspondence between a power unit and the load, the controller controls the moving contact of the switch to execute the following sequence in sequence: exit the current on position → enter the discharge position to discharge residual charge → enter the pre-charge position to buffer → enter the target on position.

[0050] (3) Parallel scheduling: When the power demand of a certain load exceeds the capacity of a single power unit, the controller instructs the switches of two or more power units to simultaneously select the same load stationary contact to realize the parallel power supply of multiple modules.

[0051] (4) Redundancy: When a power unit fails, the controller instructs another power unit to switch to the load originally connected to the failed unit, thereby achieving seamless succession.

[0052] The controller in this embodiment extends the control scope from single-branch protection to global scheduling of the entire M×N matrix topology. This is achieved by real-time acquisition of system status, execution of two-level protection logic, and dynamic scheduling of power paths.

[0053] Based on the parallel scheduling in (3) above, this embodiment provides another topology, in which multiple AC / DC power units are simultaneously locked to the same set of load static contacts on the same distributed DC busbar through their respective single-pole multi-throw switches, thereby realizing that multiple power units supply power to the same load in parallel, such as Figure 4As shown, the common moving contact of power units U1 and U2 is simultaneously located at position ③, indicating that both power units are in a discharging state, meaning they are shutting down. The internal electronic switches are off, and the moving contacts are in the discharging position, releasing residual charge through the discharge resistor. The common moving contact of power units U3 and U4 is simultaneously located at position ①, indicating that both power units are in a pre-charging state. This means that the two power units are slowly charging the bus capacitor on the load side through a series pre-charging resistor to suppress the large current surge generated when directly connected.

[0054] Similarly, parallel scheduling can be determined in the following two ways, such as Figure 4 As shown, when the moving contacts of the two power units are simultaneously locked in position 2, i.e., the ON position, their DC output terminals are physically connected in parallel through the same distributed DC busbar, jointly supplying power to load 1. However, this method cannot achieve dynamic reconfiguration because each switch's position 2 is fixedly connected to a certain DC busbar, and the load busbar it is connected to cannot be changed by the switch itself.

[0055] like Figure 5 As shown, power unit U1 is positioned at position ④ via a common moving contact and supplies power to load 4. Power units U2, U3, and U4 are simultaneously positioned at position ① via a common moving contact, indicating that all three power units are in a pre-charge state. That is, two power units are slowly charging the bus capacitor on the load side through a series pre-charge resistor, as shown. Figure 6 As shown, power units U1, U2, U3 and U4 are simultaneously located at position ① via a common moving contact, indicating that all four power units are in a pre-charge state. That is, two power units slowly charge the bus capacitor on the load side through a series pre-charge resistor.

[0056] Therefore, based on this, this application provides another solution, which can use a single-pole multi-throw switch to achieve dynamic reconfiguration, that is, there are N independent load static contacts in the switch's on position, which are respectively connected to N different distributed DC sub-buses, that is, each switch itself can select among N buses.

[0057] Specifically, to achieve N+X redundancy, the number of AC / DC power units M configured in this embodiment is X more than the actual number of loads N that need to work simultaneously, and X≥1. The extra units serve as hot backups, which can seamlessly take over when any working unit fails, ensuring uninterrupted power supply to the load. For example, 4 power units are equipped with 3 loads, which is 3+1 redundancy, and X=1; 5 power units are equipped with 3 loads, which is 3+2 redundancy, and X=2.

[0058] In this embodiment, the system is configured as follows: M=4 power units, N=3 loads, and the corresponding distributed DC busbars 1, 2, and 3 form a 4×3 matrix topology. Based on the above system configuration, the work is allocated. Power units 1, 2 and 3 are connected to the corresponding load stationary contacts 1, 2 and 3 respectively. Power unit 4 is connected to any load stationary contact, but at this time the output power is zero or extremely low, and it is in hot standby mode. The controller monitors the output current of each power unit in real time through the current sensor. When the output current of power unit 1 drops to zero or exceeds the normal operating range, it is determined that load 1 has failed. Then, the following operations are performed: the power electronic switching device inside the power unit 1 turns off automatically within microseconds, and the controller instructs the single-pole multi-throw switch corresponding to the power unit 1 to switch its moving contact from the load stationary contact 1 to the discharge position, i.e., position 3, to discharge the residual charge; then it switches to the ground position, i.e., position 4, to achieve complete physical isolation. Next, the controller instructs power unit 4 to perform the following switching: If power unit 4 is currently locked to other stationary contacts, first switch its moving contact from the current position to the pre-charge position, i.e., position 1. Through the pre-charge resistor buffer, make the output voltage of the power unit match the voltage of the sub-bus 1 where the load stationary contact 1 is located. After the voltage is matched, the moving contact switches from the pre-charge position to the on position and precisely switches to the load stationary contact 1. Finally, the output of the power unit 4 supplies power to the load 1 through the load stationary contact 1 and the busbar 1.

[0059] Therefore, based on the cross-connection capability of the matrix topology, each single-pole multi-throw switch corresponding to each power unit in this invention is equipped with N sets of fixed load stationary contacts, which are respectively connected to N distributed DC busbars. The controller can instruct the common moving contact of each switch to switch between different stationary contacts according to the real-time power demand of the load, so as to realize the flexible switching of any power unit to any load.

[0060] like Figure 7 The horizontal axis X-axis represents time. ms The y-axis represents the time from fault occurrence to complete disconnection. The vertical axis (Y-axis) represents the fault current amplitude (A), indicating the magnitude of the current in the fault circuit. Three curves are shown: the curve for this scheme, the conventional circuit breaker scheme, and the ideal disconnection curve. When a fault occurs, the current rapidly rises to its peak value, followed by a reversing peak (negative overshoot), and then quickly returns to zero.

[0061] The methods in this solution, such as sending a shutdown signal to the AC / DC power unit corresponding to the faulty branch, causing its internal switches to turn off within microseconds and interrupting the fault current, all directly interrupt the fault current through the power electronic switching devices inside the power unit. This method is fast and efficient. di / dtThe fault current rise is limited to a very low level, and the total disconnection time is shortened by about 3 to 5 times compared with conventional DC circuit breakers, reducing the energy surge by several orders of magnitude and fundamentally avoiding secondary hazards such as bus voltage drop, equipment overload overheating, and contact arcing. The reverse spike appearing on the waveform is not abnormal, but a physical phenomenon of inductor energy release during rapid electronic switch shutdown. The faster the electronic switch shuts off, the more obvious the reverse spike, but the shorter the energy release time, and the overall hazard is far lower than the long-term continuous high-current surge of conventional solutions. Conventional DC circuit breakers have a slow shutdown speed, so the inductor energy release is gradual, and the induced electromotive force amplitude is far insufficient to form a significant reverse spike; instead, the arc resistance gradually increases, consuming most of the inductor energy in the arc. Therefore, the fault current exhibits a longer period of oscillating decay without a reverse spike. The energy release of the inductor stored in this application is U = -L· di / dt This is the root cause of reverse spikes, which occur when electronic switches interrupt fault currents of hundreds of amperes at extremely high speeds. di / dt A reverse induced electromotive force (EMF) much higher than the bus voltage is generated on the parasitic inductance. This EMF drives current to flow along the freewheeling loop, forming a transient current in the opposite direction to the original fault current, which manifests as a distinct negative spike in the waveform. Although this reverse spike has a high instantaneous energy density, its duration is extremely short, and the energy is confined within the power unit and will not be conducted to other busbars, posing no substantial harm to the overall system. The existence of the reverse spike precisely demonstrates the extremely fast turn-off speed; that is, the faster the turn-off, the higher the efficiency. di / dt The larger the value, the more pronounced the reverse spike, but the shorter the total energy release time, and the overall thermal effect is far lower than that of the long-term continuous high-current impact of conventional solutions.

[0062] Therefore, the microsecond-level electronic shutdown of the power unit undertakes all arc-extinguishing tasks, ensuring that the downstream load switch is in a current-free interruptible state when a fault occurs. Thus, the downstream load switch does not need any DC breaking capacity or arc-extinguishing capability; its sole function is to establish a mechanical break under current-free conditions, which is precisely the operating condition that load switches excel at. Compared to existing technologies where each branch must be equipped with a DC circuit breaker or solid-state circuit breaker, the cost of downstream equipment in this invention can be significantly reduced, while also reducing heat sources and vulnerable components, further improving system reliability.

[0063] Example 2: This example also provides a selective protection method for HVDC power supply systems based on matrix topology, such as... Figure 8 As shown, the method includes the following steps: S1: Real-time detection of the output current of each power unit; S2: When the output current of the target power unit is detected to exceed the preset threshold, the power electronic switching device inside the target power unit is controlled to perform a shutdown action to cut off the fault current within microseconds. S3: After the fault current returns to zero, control the moving contact of the single-pole multi-throw switch corresponding to the target power unit to switch from the on position to the discharge position or the ground position, so that a physical isolation break is formed between the target power unit and the load.

[0064] Furthermore, the method also includes: in step S3, after the moving contact is switched to the discharge position, each DC busbar forms a discharge circuit through the corresponding discharge resistor, the discharge stationary contact and the moving contact, so as to discharge the residual charge in the target power unit and the corresponding DC busbar.

[0065] In this embodiment, after step S3, the method further includes: S4: When it is determined that another power unit needs to take over the power supply to the load from the target power unit that has been turned off, the moving contact of the single-pole multi-throw switch corresponding to the other power unit is controlled to perform the following operations in sequence: First, switch to the pre-charge position, and perform buffer charging on the corresponding DC busbar through the pre-charge resistor. After the voltage is matched, switch to the on position and switch to the load stationary contact position that is electrically connected to the target load, thereby completing the dynamic switching of the power supply path.

[0066] This embodiment also includes: S5: When the power demand of any load exceeds the rated output power of a single power unit, control the moving contacts of the single-pole multi-throw switches corresponding to two or more power units to switch to the ON position simultaneously and lock them to the same load stationary contact electrically connected to the load, thereby enabling multiple power units to supply power to the load in parallel.

[0067] Furthermore, including: In step S5, before controlling the moving contacts of the single-pole multi-throw switches corresponding to two or more power units to simultaneously switch to the ON position, the method further includes: Each of the moving contacts is controlled to switch to the pre-charge position first, and the same distributed DC busbar is buffered and charged through the pre-charge resistor; The voltage of each moving contact in the circuit is detected. Once the voltage matches the voltage of the distributed DC busbar, each moving contact is controlled to switch to the ON position simultaneously and locked to the same set of load stationary contacts that are electrically connected to the distributed DC busbar.

[0068] The specific implementation method of step S5 is as described in the above-mentioned parallel scheduling technology, and will not be repeated here.

[0069] Furthermore, the method also includes: S6: When the power demand of the target load drops below the rated output power of a single power unit, the moving contact of the single-pole multi-throw switch corresponding to at least one power unit in the parallel power units is controlled to exit the same group of load stationary contacts and switch to the load stationary contact electrically connected to other loads or switch to the discharge position, thereby realizing the conversion from parallel power supply to independent power supply.

[0070] Specifically, the controller initiates the parallel disconnection process when it detects that the following three conditions are met: Condition 1: The real-time power demand of the target load has been reduced to below the safe threshold of the rated output power of a single power unit; Condition 2: Duration confirmation: Power drop lasts for a certain period of time to avoid frequent switching caused by instantaneous load fluctuations; Condition 3: There are other loads that require capacity expansion, or the released power units need to be redistributed to achieve global load balancing.

[0071] In one example, if the initial power units 1 and 2 are connected in parallel to supply power to load 1, the target state is that power unit 2 is removed from load 1 and redistributed to load 2. First, the controller instructs the internal electronic switch of power unit 2 to gradually reduce the output current. The controller instructs the moving contact of switch 2 to switch from the pre-charge position to the on position, thereby connecting the corresponding target load stationary contact to form a new power supply combination.

[0072] Therefore, this embodiment achieves row isolation capability based on the matrix topology, allowing each power unit to be independently disconnected. Each power unit is independently connected to a corresponding single-pole multi-throw switch. When a power unit needs to be replaced, only the switch corresponding to that unit needs to be switched to the ground position, and that unit is completely isolated from all distributed DC sub-buses, without affecting other power units and loads. In existing series topology schemes, replacing any power unit requires a complete power outage; while single-bus parallel schemes can achieve partial online maintenance, planned power outages are still required when bus operations are involved. Therefore, this embodiment can safely complete the replacement operation under full system load, significantly shortening the mean time to repair and improving system availability.

[0073] Example 3: This example also provides an electronic device, including: a processor, a storage medium and a bus. The storage medium stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the storage medium via the bus. The processor executes the machine-readable instructions to perform the steps of the selective protection method implemented by the HVDC power supply system based on matrix topology as described in Example 2.

[0074] Finally, this embodiment also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the selective protection method implemented in the HVDC power supply system based on matrix topology as described in Embodiment 2.

[0075] The device embodiments described above are merely illustrative. The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0076] Through the detailed description of the above embodiments, those skilled in the art can clearly understand that each implementation can be achieved by means of software plus necessary general-purpose hardware platforms, and of course, it can also be achieved by hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the prior art, can be embodied in the form of software products. These computer software products can be stored in computer-readable storage media, including read-only memory, random access memory, programmable read-only memory, erasable programmable read-only memory, programmable read-only memory, electronically erasable rewritable read-only memory, read-only optical disc or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium capable of having or storing data.

[0077] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the invention.

[0078] Obviously, those skilled in the art can make various modifications and variations to the embodiments of the present invention without departing from the spirit and scope of the embodiments of the present invention. Thus, if these modifications and variations to the embodiments of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention also intends to include these modifications and variations.

Claims

1. An HVDC power supply system based on matrix topology, characterized in that, include: M power units are used to convert the medium-voltage AC power provided by the external medium-voltage AC bus into DC power. The input terminals of the M power units are connected in parallel to the external medium-voltage AC bus. N DC busbars, one end of each DC busbar is electrically connected to the corresponding load; M single-pole multi-throw (SPMD) switches are used, each SPMD switch including one moving contact and N load stationary contacts. The output terminal of the power unit is connected to the moving contact, and the other end of the DC bus is connected to the N sets of load stationary contacts of each SPMD switch. This allows any one of the M power units to be selectively connected to any one of the N loads through its corresponding SPMD switch, thus forming an M×N matrix topology. ; The controller is connected to the control terminal of each of the single-pole multi-throw switches and is used to control the moving contact of each of the single-pole multi-throw switches to selectively close with a corresponding set of load stationary contacts to establish or disconnect the power supply path between the corresponding power unit and the load.

2. The HVDC power supply system based on matrix topology according to claim 1, characterized in that, The single-pole multi-throw switch also includes a pre-charge stationary contact, a discharge stationary contact, and a grounding stationary contact disposed on the switch base; The positive terminals of each of the DC busbars are connected in series with pre-charge resistors and then connected together to the pre-charge stationary contact. When the moving contact is rotated to the pre-charge position, the DC output terminal of the power unit forms a parallel pre-charge circuit with each of the distributed DC busbars via the moving contact, the pre-charge stationary contact and each of the pre-charge resistors. Each of the DC busbars is connected in series with a discharge resistor at its positive terminal and then connected to the discharge stationary contact. When the moving contact is rotated to the discharge position, each of the DC busbars forms a parallel discharge circuit with the DC output terminal of the power unit via the corresponding discharge resistor, the discharge stationary contact, and the moving contact. The positive terminals of each of the DC busbars are connected to the grounding stationary contact, which is connected to the protective ground. When the moving contact is rotated to the grounding position, the DC output terminal of the power unit is grounded via the moving contact and the grounding stationary contact.

3. The HVDC power supply system based on matrix topology according to claim 2, characterized in that, The positive terminal of each of the DC busbars is also connected to N independent load stationary contacts located in the ON position of the switch, which are used to form the main power supply circuit of the corresponding load when the moving contact is rotated to the ON position.

4. The HVDC power supply system based on matrix topology according to claim 1, characterized in that, The number of power units M is greater than or equal to the number of DC busbars N.

5. A selective protection method for an HVDC power supply system based on the matrix topology described in any one of claims 1-4, characterized in that, The method includes the following steps: S1: Real-time detection of the output current of each power unit; S2: When the output current of the target power unit is detected to exceed the preset threshold, the power electronic switching device inside the target power unit is controlled to perform a shutdown action to cut off the fault current within microseconds. S3: After the fault current returns to zero, control the moving contact of the single-pole multi-throw switch corresponding to the target power unit to switch from the on position to the discharge position or the ground position, so that a physical isolation break is formed between the target power unit and the load.

6. The selective protection method according to claim 5, characterized in that, The method further includes: in step S3, after the moving contact is switched to the discharge position, each DC busbar forms a discharge circuit through the corresponding discharge resistor, the discharge stationary contact and the moving contact, so as to discharge the residual charge in the target power unit and the corresponding DC busbar.

7. The selective protection method according to claim 6, characterized in that, Following step S3, the following is also included: S4: When it is determined that another power unit needs to take over the power supply to the load from the target power unit that has been turned off, the moving contact of the single-pole multi-throw switch corresponding to the other power unit is controlled to perform the following operations in sequence: First, switch to the pre-charge position, and perform buffer charging on the corresponding DC busbar through the pre-charge resistor. After the voltage is matched, switch to the on position and switch to the load stationary contact position that is electrically connected to the target load, thereby completing the dynamic switching of the power supply path.

8. The selective protection method according to claim 7, characterized in that, After step S4, the method further includes: S5: When the power demand of any load exceeds the rated output power of a single power unit, control the moving contacts of the single-pole multi-throw switches corresponding to two or more power units to switch to the ON position and lock them to the same load stationary contact electrically connected to the load, thereby realizing that multiple power units are connected in parallel to supply power to the load.

9. The selective protection method according to claim 8, characterized in that, In step S5, before controlling the moving contacts of the single-pole multi-throw switches corresponding to two or more power units to simultaneously switch to the ON position, the method further includes: Each of the moving contacts is controlled to switch to the pre-charge position first, and the same distributed DC busbar is buffered and charged through the pre-charge resistor; The voltage of each moving contact in the circuit is detected. Once the voltage matches the voltage of the distributed DC busbar, each moving contact is controlled to switch to the ON position simultaneously and locked to the same set of load stationary contacts that are electrically connected to the distributed DC busbar.

10. The selective protection method according to claim 9, characterized in that, The method also includes: S6: When the power demand of the target load drops below the rated output power of a single power unit, the moving contact of the single-pole multi-throw switch corresponding to at least one power unit in the parallel power units is controlled to exit the same group of load stationary contacts and switch to the load stationary contact electrically connected to other loads or switch to the discharge position, thereby realizing the conversion from parallel power supply to independent power supply.

11. An electronic device, characterized in that, include: The device includes a processor, a storage medium, and a bus, wherein the storage medium stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the storage medium via the bus, and the processor executes the machine-readable instructions to perform the steps of the selective protection method for a matrix topology-based HVDC power supply system as described in any one of claims 5 to 10.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the steps of the selective protection method for an HVDC power supply system based on matrix topology as described in any one of claims 5 to 10.

Images