Server power supply system, server, topology generation method and storage medium

Abstract

This application discloses a server power supply system, a server, a topology generation method, and a storage medium, relating to the field of power electronics technology. This application adopts a two-level architecture and physical isolation design of "high-voltage centralized conversion - medium- and low-voltage distributed distribution" using a first DC bus and a second DC bus. The second DC bus provides an adaptive mounting platform for load nodes such as supercapacitor nodes and energy storage nodes. The number of load nodes can be determined based on their own power requirements and the voltage and current rating of the second DC bus. This separates high-voltage power conversion from load node power supply, achieving physical isolation between the high-voltage DC of the first DC bus and the low-voltage DC of the second DC bus. This ensures that the high-voltage DC input is safely, efficiently, and stably distributed to each load node within the server after conversion. This solves the safety hazards and reliability risks caused by directly connecting high-voltage DC to load nodes in related technologies, as well as the poor load node mounting adaptability.

Description

Technical Field

[0001] This application relates to the field of server power supply technology, and in particular to a server power supply system, a server, a topology generation method, and a storage medium. Background Technology

[0002] To address the surge in AI computing power and the demand for improved energy efficiency in data centers, high-voltage direct current (HVDC) has become a direction for power supply upgrades. However, in related technologies, HVDC (High-Voltage Direct Current) input poses safety hazards and reliability risks due to the direct connection of high-voltage direct current to the server motherboard. Summary of the Invention

[0003] This application provides a server power supply system, server, method, medium, and product to at least solve the technical problems of safety hazards and reliability risks caused by direct connection of high-voltage DC power to the server motherboard in related technologies.

[0004] This application provides a server power supply system, comprising: a first DC bus for providing high-voltage DC power to the server at a voltage higher than a first preset voltage; at least one second DC bus for providing low-voltage DC power to the server's load nodes at a voltage lower than a second preset voltage, wherein the first preset voltage is higher than the second preset voltage; at least one voltage conversion module connected to the first DC bus, with its input terminal connected to the first DC bus and its output terminal connected to at least one second DC bus, converting the high-voltage DC power into low-voltage DC power; and load nodes connected to at least one second DC bus, the number of second DC buses connected to the load nodes being determined based on the first power requirement of the load nodes, the first voltage level of the second DC bus, and the current carrying capacity of the second DC bus.

[0005] This application also provides a server, including the server power supply system of the above embodiments; at least one load node, the load node being connected to at least one second DC bus, the number of second DC buses connected to the load node being determined based on the first power demand of the load node, the first voltage level and current carrying level of the second DC bus.

[0006] This application also provides a topology generation method for generating the topology of the server power supply system described in the above embodiments. The method includes the following steps: obtaining at least one of the second voltage level of the first DC bus, the first voltage level of the second DC bus, the current carrying level of the second DC bus, and the first power demand of the load nodes; calculating the number of power modules in the voltage conversion module based on the first voltage level and the second voltage level; calculating the number of the second DC bus based on the first voltage level, the current carrying level, and the first power demand; and generating the topology of the server power supply system based on the number of power modules, the number of the second DC bus, and the number of voltage conversion modules.

[0007] This application also provides a non-volatile computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the steps of the topology generation method as described in the above embodiments.

[0008] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the topology generation method described in the above embodiments.

[0009] This application utilizes a two-tier architecture and physical isolation design based on a first DC bus and a second DC bus, employing "high-voltage centralized conversion - medium- and low-voltage distributed distribution." The second DC bus provides a suitable mounting platform for load nodes such as supercapacitor nodes and energy storage nodes. The number of load nodes can be determined based on their power requirements and the voltage and current rating of the second DC bus. This separates high-voltage power conversion from load node power supply, achieving physical isolation between the high-voltage DC of the first DC bus and the low-voltage DC of the second DC bus. This ensures that the high-voltage DC input is safely, efficiently, and stably distributed to each load node within the server after conversion. This solves the safety hazards and reliability risks caused by directly connecting high-voltage DC to load nodes in related technologies, as well as the poor load node mounting adaptability. Attached Figure Description

[0010] To more clearly illustrate the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 An architecture diagram of a server power supply system provided in this application embodiment;

[0012] Figure 2 A diagram of a server high-voltage DC multi-bus power supply architecture provided in this application embodiment;

[0013] Figure 3 This is a schematic diagram of the high-voltage DC multi-bus power supply architecture provided in the embodiments of this application;

[0014] Figure 4 This is a schematic diagram of the high-voltage conversion board architecture provided in an embodiment of this application;

[0015] Figure 5 This application provides a schematic diagram of the structure of a server according to an embodiment of the present application.

[0016] Figure 6 This is a schematic diagram of the chip motherboard architecture provided in an embodiment of this application;

[0017] Figure 7 This is a schematic diagram of the supercapacitor board architecture provided in an embodiment of this application;

[0018] Figure 8 This is a schematic diagram of the high-voltage direct current multi-bus scheme provided in the embodiments of this application;

[0019] Figure 9 This is a schematic diagram of a topology generation method provided in an embodiment of this application. Detailed Implementation

[0020] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the protection scope of this application.

[0021] It should be noted that, in the description of this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. The terms "first," "second," etc., in this application are used to distinguish similar objects and are not used to describe a specific order or sequence.

[0022] With the surge in AI computing power, single GPU power consumption exceeds 1200W, and future rack computing density will exceed 500kW / rack, creating a need to reduce PUE. High-voltage direct current (HVDC) technology has become a key direction for data center power supply upgrades due to its high efficiency and ease of deployment, and is evolving towards higher voltage levels such as 800V. However, the adaptation of related technologies to the server side still faces significant bottlenecks, with the problems in adapting to high-voltage HVDC manifesting in the following ways:

[0023] 1. Key energy storage nodes, such as supercapacitors / BBUs (Battery Backup Units), cannot be directly connected;

[0024] Second, directly connecting high-voltage DC power to the server motherboard poses safety hazards and reliability risks.

[0025] Third, the lack of a board-level high-voltage DC power distribution scheme makes it difficult to meet the precise power supply requirements inside the server.

[0026] Specifically, supercapacitors and BBUs, as core components ensuring power supply resilience, cannot be effectively connected or operate in high-voltage HVDC environments due to their inherent low-voltage operating characteristics. Although related technologies have enabled HVDC input access, they have not solved this compatibility bottleneck, directly threatening core functions such as data center power outage protection. As HVDC upgrades to the 800V level, directly introducing kilovolt-level high-voltage DC power into the server motherboard area filled with precision components can easily cause serious safety issues such as arcing, breakdown, and EMI (Electromagnetic Interference), and fault isolation is difficult, exacerbating system reliability risks.

[0027] There is a lack of mature solutions in related technologies to safely and efficiently distribute high-voltage DC power to low-voltage chips at the board level, such as GPUs (Graphics Processing Units) and CPUs (Central Processing Units), making it impossible to achieve precise power supply. Even if the input access problem is solved, it is still difficult to meet the power supply requirements of internal server components.

[0028] In summary, the relevant technologies have significant shortcomings in areas such as high-voltage compatibility of key energy storage components, high-voltage safety protection and fault isolation of server motherboards, and fine-grained board-level power distribution for core chips such as GPUs, making it difficult to adapt to the power supply requirements of HVDC technology upgrades and high-performance servers.

[0029] To address the aforementioned issues, this application proposes a server power supply architecture adapted to high-voltage direct current (HVDC) input. This architecture achieves high-voltage adaptation and mounting of key energy storage nodes (supercapacitors / BBUs), safe isolation between HVDC power and the server motherboard, and refined allocation of board-level HVDC power. To enable those skilled in the art to better understand this application, further detailed description is provided below with reference to the accompanying drawings and specific embodiments.

[0030] The specific application environment architecture or specific hardware architecture on which the server power supply system depends is described here.

[0031] Figure 1This is a schematic diagram of a server power supply system provided in an embodiment of this application.

[0032] like Figure 1 As shown, the server power supply system 10 includes: a first DC bus 101, at least one second DC bus 102, and at least one voltage conversion module 103.

[0033] The system includes a first DC bus 101, which provides high-voltage DC power to the server at a voltage higher than a first preset voltage; at least one second DC bus 102, which provides low-voltage DC power to the server's load nodes at a voltage lower than a second preset voltage, wherein the first preset voltage is higher than the second preset voltage; at least one voltage conversion module 103, which is mounted on the first DC bus 101, with its input connected to the first DC bus 101 and its output connected to at least one second DC bus 102, converting the high-voltage DC power to low-voltage DC power; and load nodes are mounted on at least one second DC bus 102, the number of second DC buses 102 mounted on the load nodes being determined based on the first power requirement of the load nodes, the first voltage level of the second DC bus 102, and the current carrying capacity level of the second DC bus 102.

[0034] It is understood that in this embodiment, the high voltage DC input is carried by the first DC bus with a voltage higher than the first preset voltage, and the voltage conversion module is used to reduce the high voltage level to the low voltage level of the second DC bus, which is lower than the second preset voltage. The first preset voltage is higher than the second preset voltage, thus constructing an adaptation bridge between the high voltage input and the load node connected to the second DC bus. Because the second DC bus is at a low voltage level, energy storage nodes with inherently low voltage characteristics, such as supercapacitors and BBUs, can be directly connected to it. These energy storage nodes, along with other load nodes of the server, obtain adapted power through the second DC bus. The number of second DC buses connected to a load node can be flexibly determined based on its own primary power requirement, the primary voltage level of the second DC bus, and the current carrying capacity. This solves the dual bottlenecks of critical energy storage components being unable to adapt to high-voltage HVDC environments and poor load node compatibility. At the same time, high-voltage DC power is transmitted only through the first DC bus, and the converted low-voltage power is supplied to various load nodes of the server (such as internal components like the motherboard, CPU, and GPU) through the second DC bus. This avoids safety hazards and reliability risks such as arcing and breakdown caused by direct high-voltage access from a physical isolation perspective. Furthermore, the setting of the second DC bus can provide a refined power distribution path for load nodes with different power requirements, such as GPUs and CPUs within the server, making up for the lack of board-level high-voltage DC distribution solutions. Through hierarchical voltage conversion and bus design, the compatibility, safety, and distribution efficiency issues of high-voltage HVDC in server applications are solved, providing a reliable solution for adapting high-voltage DC power supply to high-performance servers and providing stable power to their various load nodes.

[0035] It should be noted that the first preset voltage is a system-preset high-voltage threshold used to distinguish between high voltage and medium / low voltage. In server power supply scenarios, it is typically set to the 100V-200V range, but the specific value can be configured as needed. The first DC bus is the high-voltage power input channel of the server power supply system, corresponding to the high-voltage DC bus copper bus in the system. It provides the server with high-voltage DC power higher than the first preset voltage. Common voltage levels are 400V, ±400V, 800V, or other higher DC voltage levels. It provides energy input for the entire power supply system and can simultaneously support multiple server power supply architectures. Its function is to carry and transmit high-voltage DC power to subsequent voltage conversion components. The voltage conversion module refers to the component that realizes the conversion from high voltage to medium / low voltage, corresponding to the high-voltage conversion board in the system. At least one module is provided, mounted on the first DC bus, with one end connected to the first DC bus and the other end connected to the second DC bus. The voltage conversion module integrates a first circuit board, at least one high-voltage power supply module, voltage bus, interface and other structures. Its function is to reduce the high voltage level of the first DC bus to the medium / low voltage level of the second DC bus. It can also utilize the space between the high voltage copper bus and the intermediate voltage copper bus to arrange multiple high-efficiency power supply modules to improve conversion efficiency.

[0036] The second preset voltage is the system's preset low-voltage threshold, used in conjunction with the first preset voltage to classify voltage levels. It is typically set within the 30V-60V range, but specific values ​​can be configured as needed. The current carrying capacity refers to the maximum current threshold that the second DC bus is allowed to continuously carry; it is an indicator of the bus's power supply capacity and determines the upper limit of the load power that a single bus can drive. The second DC bus refers to the transmission and distribution channel for converted medium / low voltage electricity in the system, corresponding to the intermediate voltage bus copper bus in the system. At least one bus is provided, connected to the output of the voltage conversion module. It is used to provide low-voltage DC power (below the second preset voltage) to the server's load nodes (where the first preset voltage is higher than the second preset voltage). Common voltage levels include 54V, 48V, and 24V. Its function is to deliver the converted power to load nodes such as supercapacitor boards, lithium battery nodes, and chip motherboards. The number of second DC buses connected to load nodes is determined based on the load node's first power requirement, the first voltage level of the second DC bus, and the current carrying capacity, providing dedicated power supply paths for different load nodes. The first voltage level refers to the low-voltage DC voltage level transmitted on the second DC bus, i.e., the output voltage level after being stepped down by the voltage conversion module. It serves as the adaptation voltage reference for the load node and determines the voltage adaptation logic connected to the load node. Common values ​​include 54V, 48V, 24V, or other lower DC voltage levels. This application does not specify the material or implementation method of the first and second DC buses.

[0037] Specifically, such as Figure 2 As shown, the first DC bus corresponds to the high-voltage DC bus copper busbar in the diagram, which is the high-voltage power input channel for the entire system. The server rack is connected to the main power supply through it. The voltage conversion module corresponds to the high-voltage conversion board in the diagram, which is used to convert the high-voltage electricity input from the high-voltage DC busbar into intermediate voltages such as 54V, 48V, and 24V. The second DC bus corresponds to... Figure 2 The intermediate voltage busbar is the transmission channel for the converted low-voltage electricity, used to deliver the intermediate voltage to the supercapacitor board and the chip motherboard to achieve power distribution.

[0038] To further clarify, the number of second DC buses connected to the load nodes Based on the first required power P of the load node and the first voltage level of the second DC bus Current rating of the second DC bus The calculation yields the following formula: The first voltage level of the common second DC bus. Taking 54V as an example, the current carrying capacity of the second DC bus The first power requirement of the chip motherboard is 100kW, calculated as follows: (The value is 1000A). The value is 1.85, rounded up to 2 second DC buses. If the first power requirement of the load node is 200kW, then 4 second DC buses can be arranged, and the number of second interfaces will be adjusted accordingly. It should be noted that there is only one first DC bus, but the entire server can be connected to the first DC bus by several of the above power supply architectures at the same time.

[0039] In this application embodiment, the voltage conversion module includes a first circuit board and a first interface, a power module array and at least one second interface disposed on the first circuit board. The first interface is used to connect to a first DC bus, the second interface is used to connect to a second DC bus, and the power module array is used to convert high voltage DC power into low voltage DC power.

[0040] It is understood that in this embodiment of the application, the power module array, as the execution component for voltage conversion, undertakes the task of converting high-voltage DC to low-voltage DC between the first DC bus and the second DC bus. The first circuit board provides an integration carrier and a basis for collaborative operation. The first circuit board also integrates a first interface and at least one second interface, wherein the first interface is used to connect to the first DC bus and the second interface is used to connect to the second DC bus, thus clarifying the dedicated connection path between the voltage conversion module and the two buses. On the one hand, the integrated power module array can stably and efficiently convert high voltage to low voltage, ensuring the stability of the low voltage level of the second DC bus and providing a reliable mounting foundation for low-voltage energy storage nodes such as supercapacitors and BBUs, thus guaranteeing the compatibility of key energy storage components at the execution level. On the other hand, the integrated layout of the power module array makes the voltage conversion process easier to control, avoiding direct interaction between high-voltage circuits and low-voltage loads, further isolating high-voltage risks and reducing safety hazards caused by high-voltage access to server load nodes. Simultaneously, the power module array on the first circuit board, in conjunction with at least one second interface, can support multiple corresponding second DC buses, providing independent low-voltage power supply paths for different load nodes (such as GPUs and CPUs) within the server, supporting the implementation of a fine-grained board-level power distribution scheme from a hardware structure perspective. This enhances the stability, safety, and flexibility of voltage conversion.

[0041] It should be noted that the first interface is a connection component integrated on the first circuit board of the voltage conversion module. Its function is to realize the physical connection and power transmission between the voltage conversion module and the first DC bus. It is used to access the high-voltage DC power on the first DC bus and does not have low-voltage power output or signal interaction functions, ensuring the stability and safety of high-voltage power transmission. The second interface is an output connection component integrated on the first circuit board. It is the power output channel between the voltage conversion module and the second DC bus. Its function is to transmit the low-voltage DC power converted by the power module array to the second DC bus for low-voltage power output. It cannot be connected to high-voltage power. At least one interface should be provided, and it can be flexibly configured according to the number of second DC buses and the power supply requirements of the load nodes. The interface specifications must match the low-voltage level, current carrying capacity, and connection standard of the second DC bus. A power module array refers to a functional unit (as opposed to scattered independent power modules) integrated on a first circuit board and composed of multiple power modules arranged in a preset layout. It is the execution component for voltage conversion, converting the high-voltage DC power input to the first interface into low-voltage DC power that meets the requirements of the second DC bus. The model and quantity of the power modules in the array need to be determined based on the high-voltage level of the first DC bus, the low-voltage level of the second DC bus, and the total power requirements of the system. Furthermore, the modules work collaboratively to improve conversion efficiency and stability. This application does not specifically limit the specific implementation method of the voltage conversion module.

[0042] Specifically, the first circuit board is the hardware integration carrier of the voltage conversion module. Located inside the voltage conversion module, it provides physical installation space and electrical connection paths for the power module array, the first interface, the second interface, and other circuit components. It ensures the coordinated operation of all components during voltage conversion and serves as the hardware foundation platform for the voltage conversion module's functionality. The power module array is the core functional unit within the voltage conversion module, integrated on the first circuit board. Its function is to convert the high-voltage DC input from the first DC bus into the low-voltage DC required by the second DC bus. Simultaneously, it optimizes the arrangement of the power module array using the space between the high-voltage copper bus and the intermediate voltage copper bus, improving overall conversion efficiency. It is the core component for voltage conversion execution.

[0043] In this embodiment of the application, the first interface is disposed on the first side of the first circuit board, and the second interface is disposed on the second side of the first circuit board, with the first side and the second side being opposite sides.

[0044] It is understood that in this embodiment, the first interface is located on the first side of the first circuit board, serving as a high-voltage input terminal directly connected to the first DC bus, introducing high-voltage electricity into the voltage conversion module. The second interface is located on the second side of the first circuit board (the first and second sides are opposite sides), serving as a low-voltage output terminal connected to the second DC bus. After the power module array completes the conversion from high-voltage DC to low-voltage DC, it directly outputs low-voltage electricity to the second DC bus through the second interface. On the one hand, the first and second interfaces are located on opposite sides of the circuit board, forming a physical separation layout between high-voltage input and low-voltage output, further strengthening the physical isolation between high-voltage circuits and low-voltage loads. Compared with the direct high-voltage connection method in related technologies, this significantly reduces safety hazards such as arcing and breakdown, and also reduces the impact of electromagnetic interference on server load nodes (such as precision components like GPUs and CPUs). On the other hand, at least one second interface can be connected to different second DC buses, which can provide dedicated mounting paths for low-voltage energy storage nodes such as supercapacitors and BBUs through independent second DC buses, and can also allocate independent low-voltage power supply paths for different load nodes inside the server, making the voltage conversion module more efficient and reliable.

[0045] It should be noted that in the embodiments of this application, "first side" and "second side" refer to two opposing edges or surfaces of the first circuit board, such as the left and right sides, front and back of the circuit board. The purpose of limiting "opposite sides" is to completely separate the first interface (high voltage end) and the second interface (low voltage end) in physical space, avoid the intersection of the paths of high voltage access and low voltage output, and further improve electrical safety and anti-interference capability. The specific orientation of the first side and the second side, such as left and right / front and back, can be flexibly adjusted according to the overall layout of the server, and this application does not make specific limitations.

[0046] Specifically, such as Figure 3As shown, the power supply and operating logic, core parameters, and module configuration details of the high-voltage conversion board node are as follows: This node first receives power input through the first DC bus. The input high-voltage electricity is introduced through the first interface on the first side of the first circuit board and directly transmitted to a power module array composed of several high-voltage power modules. After these high-voltage power modules perform voltage reduction processing, they are then connected to the corresponding second DC bus through multiple second interfaces on the second side of the first circuit board (opposite to the first side), thereby realizing voltage level conversion and subsequent power transmission. The input voltage of the high-voltage power module corresponds to the first voltage level, and the output voltage corresponds to the second voltage level. Simultaneously, due to the large installation space between the first and second DC buses, multiple high-efficiency high-voltage power modules can be arranged in this area, effectively improving the overall efficiency of the entire voltage conversion process. From the characteristics of the high-voltage power module itself, its physical structure is a rectangular module, typically around 20×40cm in size, and the circuit topology adopts an LLC (Inductor-Inductor-Capacitor) topology. This application does not specifically limit the physical structure and topology type of the high-voltage power module.

[0047] like Figure 4 As shown, the first interface is located on the first side of the first circuit board of the voltage conversion module. It is the physical connection port between the voltage conversion module and the first DC bus, used to connect to the first DC bus. Its function is to introduce the high-voltage electricity from the first DC bus into the voltage conversion module, providing input for the voltage conversion of the subsequent power module array. The second interface is located on the second side of the first circuit board of the voltage conversion module, and there is at least one. It is the physical connection port between the voltage conversion module and the second DC bus, used to connect to the second DC bus. Its function is to output the low-voltage electricity converted by the power module array to the second DC bus, supplying power to load nodes such as supercapacitors and chip motherboards. Each second interface corresponds to one second DC bus.

[0048] In this embodiment, the power module array includes multiple power modules. The high-voltage port of the power module is connected to a first bus, and the low-voltage port of the power module is connected to a second bus. The first bus is connected to a first interface, and the second bus is connected to at least one second interface. The voltage level accessed by the high-voltage port is higher than the voltage level accessed by the low-voltage port.

[0049] It is understood that in this embodiment, the power module array consists of multiple power modules. Voltage conversion is achieved through a transmission link consisting of a first interface, a first bus, a high-voltage port of the power module, a low-voltage port of the power module, a second bus, and a second interface. This ensures that the voltage conversion process is stable and efficient, and can continuously output a low-voltage level that meets the requirements of the second DC bus. This structured connection design provides a reliable mounting voltage foundation for low-voltage energy storage nodes such as supercapacitors and BBUs, avoids voltage fluctuations caused by unclear conversion links, further reduces the safety hazards of high-voltage access, and provides stable low-voltage power supply support for fine-grained board-level power distribution for server load nodes.

[0050] It should be noted that the high-voltage port is the interface on the power module used to connect to high-voltage electricity. It is only connected to the first bus, and its voltage level is higher than that of the low-voltage port and must match the high-voltage level of the first DC bus. It does not have low-voltage input or signal transmission functions. The low-voltage port is the interface on the power module used to output low-voltage electricity. It is connected to the second bus, and its output voltage level is consistent with the low-voltage level of the second DC bus. It is used for low-voltage power output and cannot be connected to high-voltage electricity. The first bus, also known as the high-voltage bus, is the high-voltage power transmission channel inside the voltage conversion module, used to connect the first interface to each... The high-voltage port of the power module is used to evenly distribute the high-voltage electricity introduced from the first interface to each power module. Its current carrying capacity must match the high-voltage level of the first DC bus and the total power of the system. The second bus, also known as the low-voltage bus, is the low-voltage power transmission channel inside the voltage conversion module. It is used to connect the low-voltage port of each power module to at least one second interface. Its function is to collect the low-voltage electricity converted by all power modules and transmit it to the second interface. Its specifications must match the low-voltage level of the second DC bus and the power supply requirements of the load nodes. This application does not specify the specific model or material of each component.

[0051] In this embodiment of the application, determining the number of power modules based on the total system power of the server and the number of voltage conversion modules includes: obtaining the module power of the power modules; calculating the module power of the voltage conversion modules based on the total system power and the number of modules; calculating the number of power modules based on the module power and the module power; and rounding up the number of power modules if the number of modules is not an integer.

[0052] It is understood that, on the one hand, the embodiments of this application determine the module power based on the total system power and the number of voltage conversion modules, so that the power module configuration of each voltage conversion module matches the power demand it needs to bear, avoiding insufficient conversion capacity due to insufficient number of modules, or waste of resources due to excessive number of modules; on the other hand, the design of rounding up when the value is not an integer reserves power redundancy for the voltage conversion process, which can cope with the power demand changes when the server load fluctuates, and further ensure that the voltage conversion module can continuously and stably output low voltage level that meets the requirements of the second DC bus, thereby providing reliable and stable low voltage power supply support for each load node of the server, while improving the power adaptability and operational reliability of the entire power supply system.

[0053] Specifically, the calculation of the number of high-voltage power modules is as follows: Let the total number of power conversion boards be N1, the number of power modules required per power conversion board be N2, the total system power be P, the power of a single power module be P0, and the redundancy factor be K, typically ranging from 1.2 to 1.5 to ensure the reliability of the system power supply. The formula for calculating the number of power modules N2 required per power conversion board is N2 = K × P / (P0 × N1), and the result should be rounded up. For example, if the total system power is 8000W, the power of a single power module is 400W, and a total of 8 power conversion boards are deployed, then each power conversion board needs to handle 1000W of power. Taking the redundancy factor K = 1.2, substituting into the formula, we get N2 = 1.2 × 8000 / (400 × 8) = 3, meaning each power conversion board needs to be configured with 3 high-voltage power modules.

[0054] In this embodiment, the number of power modules is determined based on the server's total system power and the number of voltage conversion modules.

[0055] It is understood that the embodiments of this application determine the number of power modules based on the total system power of the server and the number of voltage conversion modules, ensuring that the conversion capacity of each voltage conversion module is precisely matched with the system power requirements. This avoids conversion overload caused by insufficient power modules and prevents resource waste caused by excessive numbers. At the same time, sufficient and suitable power modules can stably output low voltage levels that meet the requirements of the second DC bus, providing a reliable mounting base for low-voltage energy storage nodes such as supercapacitors and BBUs, and providing stable power support for the dedicated power supply paths of board-level components such as GPUs and CPUs (all of which are load nodes), further reducing the risk of failure caused by power mismatch.

[0056] It should be noted that the total system power refers to the total power required by all load nodes when the server is operating at full load. It is an indicator of the overall server power supply demand and determines the total conversion capacity that the power modules must provide. The number of voltage conversion modules refers to the total number of voltage conversion modules configured in the server power supply system. Each voltage conversion module needs to share a portion of the total system power, and its number directly affects the power handling capacity of a single module. Both serve as the basis for determining the number of power modules. The core is to determine the power share required by a single voltage conversion module by matching the total system power with the number of voltage conversion modules, thereby matching the corresponding number of power modules to ensure that the total conversion capacity of the power module array covers the power requirements. This application does not limit the specific value of the total system power or the upper limit of the number of voltage conversion modules; these can be flexibly adjusted according to the actual server configuration.

[0057] In this embodiment, the number of second interfaces is determined based on the number of second DC buses.

[0058] It is understood that the embodiments of this application determine the number of second interfaces based on the number of second DC buses, ensuring that each second DC bus can establish a stable connection with the voltage conversion module through a dedicated second interface. This avoids power outages or transmission conflicts caused by interface and bus mismatch, and can also adapt to the second voltage level to meet the mounting requirements of low-voltage energy storage nodes such as supercapacitors and BBUs, providing dedicated power supply paths for board-level components such as GPUs and CPUs, while reducing the risk of failure caused by improper connection adaptation.

[0059] It should be noted that the matching logic between the number of second interfaces and the number of second DC buses is "one second interface per second DC bus," ensuring a dedicated path for power transmission and avoiding uneven power distribution or signal interference caused by multiple buses sharing interfaces. Load nodes rely on the low-voltage electricity transmitted from the second DC buses. The number of second interfaces, acting as a bridge between the voltage conversion module and the second DC bus, is adjusted synchronously with the number of second DC buses, requiring no additional independent calculation. Meeting the "one-to-one" adaptation requirement ensures accurate power transmission from the voltage conversion module to the second DC bus, further improving the reliability and adaptability of the power supply system.

[0060] The server power supply system of this application embodiment carries high-voltage DC input through a first DC bus and uses a voltage conversion module to build an adaptation bridge between the high-voltage input and the low-voltage load. The voltage conversion module connects to the high-voltage electricity of the first DC bus through a first interface. After the internal power module array completes the high-voltage to low-voltage conversion, the low-voltage electricity is then transmitted to the second DC bus through a second interface. Since the second DC bus transmits low-voltage electricity, energy storage nodes with inherent low-voltage characteristics, such as supercapacitors and BBUs, can be directly mounted on it, solving the bottleneck of key energy storage components being unable to adapt to high-voltage HVDC environments. At the same time, the high-voltage DC electricity is transmitted on the high-voltage side of the first DC bus and the voltage conversion module. The converted low-voltage electricity is supplied to the server chip motherboard and board-level components such as GPU and CPU through the second DC bus, physically isolating the high voltage and low voltage, avoiding safety hazards and reliability risks such as arcing and breakdown caused by direct high voltage connection to the load end. The setup of multiple second DC buses, coupled with a dedicated second interface connection design, provides a refined power distribution path for board-level components with different power consumption requirements within the server. Each second DC bus can specifically handle the power supply needs of a particular load node, compensating for the lack of board-level high-voltage DC distribution solutions. Through a hierarchical design of centralized high-voltage transmission, precise module conversion, and low-voltage branch supply, the compatibility, security, and distribution efficiency issues of high-voltage HVDC in server applications are effectively resolved, providing a reliable solution for adapting high-voltage DC power supply to high-performance servers.

[0061] Through the above description of the embodiments, those skilled in the art can clearly understand that the server power supply system according to the above embodiments can be implemented by means of a general hardware architecture plus dedicated control software. Of course, it can also be implemented by customized hardware (such as dedicated chips, fixed logic circuits). However, in most scenarios, the former (general hardware architecture with software control logic) is a better implementation method.

[0062] Figure 5 This is a schematic diagram of the structure of a server provided in an embodiment of this application. The server 50 includes the structure of the server power supply system 10 described in the above embodiment.

[0063] In this embodiment, the server 50 further includes at least one load node, which is connected to at least one second DC bus. The number of second DC buses connected to the load node is determined based on the first power requirement of the load node, the first voltage level and the current carrying capacity of the second DC bus.

[0064] It is understood that the embodiments of this application connect power supply with load demand. The low-voltage electricity transmitted by the second DC bus can directly meet the connection requirements of low-voltage energy storage nodes such as supercapacitors and BBUs. At the same time, computing nodes such as GPUs and CPUs act as load nodes, drawing power by connecting to the corresponding second DC bus. The number of second DC buses connected to the load nodes is determined based on their own first power requirement, the first voltage level and current carrying capacity of the second DC bus. This avoids the safety risks caused by direct high-voltage access to load nodes and provides dedicated power supply paths for different load nodes through adaptive bus configuration. The embodiments of this application not only support the stable access and operation of energy storage nodes such as supercapacitors and BBUs in the system, ensuring power supply redundancy and power outage protection functions; but also provide precisely adapted power supply support for computing nodes such as GPUs and CPUs, meeting the power consumption requirements of core computing components of servers; at the same time, single or two types of load nodes can be flexibly configured according to actual scenarios, enabling the power supply system to adapt to different server power supply and computing needs, improving the applicability and functionality of the overall architecture.

[0065] In the embodiments of this application, the load node is at least one including an energy storage node and a computing node.

[0066] It is understood that the load node in this embodiment includes at least one energy storage node and a computing node. The energy storage node can be flexibly configured as at least one of a capacitor unit and a battery energy storage unit: the capacitor unit can specifically adjust the voltage peak of the server, smooth voltage fluctuations during the power supply process, and ensure power supply stability; the battery energy storage unit can provide backup power for the server in the event of a power outage or other emergencies, ensuring the safety of critical operations and data. This configuration allows the load nodes to be appropriately matched according to actual needs, achieving dynamic adjustment of power supply voltage and emergency backup redundancy through the energy storage node, while relying on the computing node to support the core operations of the server, further improving the reliability and adaptability of the server power supply system.

[0067] Specifically, such as Figure 2 As shown, the power output from the voltage conversion module is connected to multiple second DC buses. Energy storage nodes (such as lithium battery nodes and other battery energy storage units) are a type of load node, connected to their corresponding second DC buses, and have no direct electrical connection to the first DC bus. The chip motherboard corresponds to computing nodes, and its number does not need to match the number of voltage conversion modules and energy storage nodes. Only one first DC bus is configured, but the entire server can be simultaneously connected to the first DC bus via several of the aforementioned power supply architectures, achieving flexible expansion of power supply capacity. For example... Figure 6 As shown, the computing node corresponds to the chip motherboard in the diagram. Its core includes components such as the CPU and GPU, serving as the core computing carrier for the load node. Power is drawn from the second DC bus and further converted within the board to precisely supply power to the CPU, GPU, and other core computing components, which is crucial for the server's computing power output. For example... Figure 7 As shown, the capacitor unit corresponds to the supercapacitor plate in the diagram. As an important component of the energy storage node, it is connected to the second DC bus. Its main function is to regulate voltage peaks and smooth power supply fluctuations, while also providing short-term emergency power during power outages. Battery energy storage units (such as lithium battery nodes) are also energy storage nodes. Their function is to provide servers with backup power for longer periods, adapting to long-term backup power and peak shaving needs, and ensuring the safety of critical operations and data during sudden power outages. In summary, energy storage nodes enhance power supply stability through energy storage and discharge regulation, while computing nodes support core computing power output. The synergistic effect of these two types of load nodes significantly improves the stability, redundancy, and resilience of the power supply system.

[0068] In this embodiment of the application, the energy storage node includes at least one of a capacitor unit and a battery energy storage unit. The capacitor unit is used to regulate the peak voltage of the server, and the battery energy storage unit is used to provide backup power to the server.

[0069] It is understood that the energy storage node in this embodiment includes at least one of a capacitor unit and a battery energy storage unit. The capacitor unit is specifically used to regulate the peak voltage of the server, while the battery energy storage unit is responsible for providing backup power to the server. This configuration effectively avoids the impact of voltage fluctuations on the equipment and ensures critical operation during sudden power outages, making the server power supply more stable and reliable.

[0070] In this embodiment, the capacitor unit includes a second circuit board, a capacitor module, a first voltage conversion device, and at least one third interface. The second circuit board is provided with the capacitor module, the first voltage conversion device, and at least one third interface. The capacitor module is connected to one end of the first voltage conversion device, the third interface is connected to the other end of the first voltage conversion device, and the third interface is connected to a second DC bus. The first voltage conversion device realizes voltage conversion between the capacitor module and the second DC bus.

[0071] It is understood that the first voltage conversion device in this application embodiment can specifically complete the voltage matching between the capacitor module and the second DC bus, ensuring that the capacitor module can be stably connected to the system and effectively play the function of regulating voltage peak; while the connection design of the third interface and the second DC bus ensures reliable connection between the capacitor unit and the power supply system, and the integrated circuit board layout improves the compactness and working stability of the capacitor unit structure.

[0072] Specifically, such as Figure 7As shown, the second circuit board is the substrate of the capacitor unit, used to support the supercapacitor module, the first voltage conversion device (corresponding to the voltage conversion board in the diagram), and the third interface (corresponding to the copper bus interface of the intermediate voltage bus in the diagram). It provides physical installation space and electrical connection paths for each component, ensuring the coordinated operation of the internal circuits of the capacitor unit. The capacitor module, corresponding to the supercapacitor module in the diagram, consists of multiple supercapacitors connected in series. It is used to divide and adapt the voltage to the intermediate bus voltage and is the energy storage component of the capacitor unit, responsible for storing electrical energy and releasing it when needed, such as smoothing voltage fluctuations and providing emergency power. The number of supercapacitors connected in series needs to be calculated and determined based on the second voltage level and the withstand voltage level of each supercapacitor. The first voltage conversion device, corresponding to the voltage conversion board in the diagram, connects the supercapacitor module and the third interface. Its function is to convert the output voltage of the supercapacitor module to a voltage level matching the second DC bus, ensuring that the capacitor unit can stably connect to the second DC bus and effectively perform voltage regulation.

[0073] In this embodiment of the application, at least one supercapacitor is provided in the capacitor module, and the number of supercapacitors in the capacitor module is determined according to the first voltage level and the voltage rating of the capacitor.

[0074] It is understood that, by matching the voltage rating of the capacitors with the number of supercapacitors at the first voltage level, the embodiments of this application ensure that the capacitor module as a whole can stably withstand the first voltage level, avoid damage caused by insufficient voltage rating, and at the same time ensure that the capacitor module can reliably perform the function of regulating voltage peak, thereby improving the safety and effectiveness of the capacitor unit's operation.

[0075] It should be noted that the withstand voltage rating refers to the maximum DC voltage that a single supercapacitor can withstand stably and continuously under normal operating conditions. It is one of the electrical parameters of a supercapacitor and determines its safe operating voltage range in a circuit. In this application, the withstand voltage rating of a single supercapacitor is typically low. If directly connected to a circuit corresponding to the first voltage level, insufficient withstand voltage could lead to supercapacitor breakdown, damage, or even safety risks. Therefore, it is necessary to improve the overall withstand voltage capability of the module by connecting multiple supercapacitors in series. The total withstand voltage after series connection must match the first voltage level, and the withstand voltage rating of each individual supercapacitor is the basis for calculating the number of supercapacitors to be connected in series.

[0076] In this embodiment, the battery energy storage unit includes a third circuit board, a battery module, a second voltage conversion device, and at least one fourth interface. The third circuit board is provided with at least one of the battery module, the second voltage conversion device, and the fourth interface. The battery module is connected to one end of the second voltage conversion device, the fourth interface is connected to the other end of the second voltage conversion device, and the fourth interface is connected to a second DC bus. The second voltage conversion device realizes voltage conversion between the battery module and the second DC bus.

[0077] It is understood that, by integrating the battery module, the second voltage conversion device, and the fourth interface on the third circuit board, this embodiment of the application achieves efficient adaptation between the battery energy storage unit and the second DC bus. The second voltage conversion device can accurately complete the voltage matching between the battery module and the second DC bus, ensuring that the battery module can be stably connected to the power supply system. The connection design of the fourth interface and the second DC bus ensures reliable docking between the battery energy storage unit and the system, enabling it to provide backup power in a timely manner when needed. At the same time, the integrated circuit board layout makes the collaborative work of each component more stable, improving the structural compactness and functional reliability of the battery energy storage unit.

[0078] It should be noted that the third circuit board is the overall hardware carrier of the battery energy storage unit. It can house at least one of the following: a battery module, a second voltage conversion device, and a fourth interface. This provides physical installation space and electrical connection paths for each component, ensuring the coordinated operation of the internal circuitry of the battery energy storage unit. The third circuit board directly supports the battery module and related auxiliary circuits (such as protection circuits and monitoring circuits), further optimizing the installation layout and electrical connection stability of the battery module, and improving the structural compactness and reliability of the battery energy storage unit. The second voltage conversion device is a voltage matching component between the battery module and the second DC bus. One end connects to the battery module, and the other end connects to the fourth interface. Its function is to convert the output voltage of the battery module to a voltage level matching the second DC bus, ensuring that the battery energy storage unit can stably connect to the second DC bus and achieve efficient charging and discharging adaptation. The fourth interface is the physical connection port between the battery energy storage unit and the second DC bus. At least one interface is located on the third circuit board, with one end connected to the second voltage conversion device and the other end connected to the second DC bus. Its function is to transmit the voltage-converted power from the battery module to the second DC bus, or to draw power from the second DC bus to charge the battery module.

[0079] In this embodiment of the application, at least one backup battery is provided in the battery module, and the number of backup batteries in the battery module is determined according to the first voltage level of the second DC bus and the capacity of the backup batteries.

[0080] It is understood that in this embodiment, the number of backup batteries in the battery module is determined based on the first voltage level and the backup battery capacity, which realizes the precise configuration of the battery module in terms of voltage adaptation and backup power supply capability. This ensures that the battery module can match the voltage level of the second DC bus and stably connect to the power supply system; and that the number of batteries can be configured according to capacity requirements to ensure sufficient backup power duration for the server in scenarios such as power outages, further improving the reliability and power supply guarantee capability of the battery energy storage unit.

[0081] It should be noted that the capacity of a backup battery refers to the amount of electrical energy that a single backup battery can store, usually measured in ampere-hours (AH). It is a parameter that measures the battery's energy storage capacity. Its value, along with the first voltage level, determines the number of backup batteries in the battery module. This must be considered in conjunction with the required backup power duration to ensure that the battery module can both meet the voltage requirements of the first voltage level and provide sufficient backup power capacity.

[0082] In this embodiment of the application, the computing node includes a fourth circuit board, at least one graphics processor, at least one central processing unit, a power supply module, and at least one fifth interface. The fourth circuit board is provided with at least one of the graphics processor, the central processing unit, the power supply module, and the fifth interface. One end of the power supply module is connected to at least one of the graphics processor and the central processing unit, and the other end of the power supply module is connected to the fifth interface. The fifth interface is connected to the second DC bus.

[0083] It is understood that, in the embodiments of this application, by integrating a graphics processing unit (GPU), a central processing unit (CPU), a power supply module, and a fifth interface on the fourth circuit board, the computing node achieves efficient power supply connection with the second DC bus. The fifth interface connects to the second DC bus to obtain low-voltage power, and then the power supply module distributes appropriate power to the core computing components such as the GPU and CPU. The integrated circuit board layout makes the various components work together more closely, which not only ensures the stability of the computing node's power supply from the power supply system, but also specifically meets the high computing power supply requirements of the GPU and CPU, thereby improving the reliability and power supply adaptability of the computing node.

[0084] It should be noted that the fourth circuit board is the hardware integration carrier of the computing node, i.e., the substrate of the chip motherboard. It houses at least one of the following: a graphics processor (GPU), a central processing unit (CPU), a power supply module, and a fifth interface. This provides physical mounting space and electrical connection paths for each component, ensuring the coordinated operation of the internal circuitry of the computing node. The GPU is the core computing element within the computing node responsible for graphics rendering and massively parallel computing. It is commonly used in deep learning and high-performance computing scenarios and requires an appropriate voltage from the power supply module for stable operation; it is one of the components responsible for computing power output. The CPU is the core element within the computing node responsible for instruction execution, system control, and core logic operations. It dominates basic operations such as data processing and task scheduling and also relies on the appropriate voltage provided by the power supply module. The power supply module is a voltage adaptation component within the computing node. One end connects to at least one of the GPU and CPU, and the other end connects to the fifth interface. Its function is to further convert the second DC bus voltage introduced by the fifth interface into a lower-level operating voltage required by the GPU and CPU, ensuring that the computing elements receive a stable and appropriate power input. The fifth interface is the physical connection port between the computing node and the second DC bus. It is located on the fourth circuit board, with at least one interface. One end is connected to the power supply module, and the other end is connected to the second DC bus. Its function is to obtain medium / low voltage power from the second DC bus and transmit it to the power supply module. It is the "power input port" of the computing node.

[0085] In this embodiment, the number of fifth interfaces is determined based on the first voltage level and the second power requirement of the computing node.

[0086] It is understood that the number of fifth interfaces in this embodiment is determined based on the first voltage level and the second power requirement of the computing node, thereby achieving a precise configuration of the power supply interface between the computing node and the second DC bus. This not only matches the first voltage level to ensure the compatibility of the interface with low-voltage power supply, but also meets the power requirements of the computing node to avoid interface overload or insufficient power supply, ensuring that core computing components such as GPU and CPU can stably obtain the required power, and improving the reliability and adaptability of the computing node power supply.

[0087] It should be noted that the second power requirement refers to the total power required by a single computing node under normal operating conditions, that is, the sum of the power requirements of all core computing components within the computing node, such as the central processing unit (CPU) and graphics processing unit (GPU). As a power parameter of the computing node, it, together with the first voltage level, determines the number of fifth interfaces. Since the fifth interface needs to draw power from the second DC bus and transmit it to the power supply module, its current carrying capacity is fixed. According to the formula "power = voltage × current", the total current required needs to be calculated based on the second power requirement and the first voltage level. Then, combined with the current carrying capacity of a single fifth interface, the number of interfaces is determined to ensure that the fifth interface can stably transmit the power required by the computing node and avoid power outages or performance limitations of computing components due to insufficient interface current carrying capacity.

[0088] Specifically, for chipset motherboards, such as Figure 6 As shown, there are intermediate voltage bus copper bus interfaces, and the number of interfaces is calculated from the above formula for the number of second DC buses. Calculations show that the chip motherboard draws power from the central copper busbar, which is then converted by the power supply module to the voltage level required by the CPU / GPU chip, ultimately powering the chip.

[0089] For a description of the features in the embodiment corresponding to this server, please refer to the relevant description of the embodiment corresponding to the server power supply system, which will not be repeated here.

[0090] The server in this embodiment achieves performance improvement by integrating a power supply system adapted to high-voltage direct current (HVDC) input. This solves the adaptation problem of high-voltage HVDC on the server side, supports stable mounting of energy storage nodes such as supercapacitors and BBUs, and meets the high computing power supply requirements of computing nodes such as GPUs and CPUs. Furthermore, the structured design and bus isolation of the voltage conversion module eliminate the safety hazards of direct high-voltage connection and ensure power supply stability. At the same time, each component, such as the power module, interface, and energy storage / computing unit, can be flexibly configured according to voltage level and power requirements, which improves the server's overall compatibility with high-voltage power supply, power supply reliability, and adaptability to different load scenarios.

[0091] The following will illustrate the process of a server high-voltage DC multi-bus scheme through a specific embodiment, such as... Figure 8 As shown, the specific steps are as follows:

[0092] In step one, the high-voltage DC voltage and intermediate bus voltage level are selected. The high-voltage DC input voltage of the system (e.g., 400V, ±400V, 800V, etc.) and the voltage level of the intermediate bus after conversion are determined. (e.g., 54V, 48V, 24V, etc.) to provide a core voltage dimension design basis for subsequent voltage conversion and component configuration.

[0093] In step two, the number of intermediate busbar voltage copper busbars is determined based on the system power rating and current requirements. This is done in conjunction with the formula... Quantitatively calculate the number of intermediate busbar copper bars. For example, when =54V =1000A, P=100kW ≈1.85, rounded up to 2. This calculation ensures that the intermediate bus can meet the power demand, avoids overload, and provides a refined power distribution path for energy storage and computing nodes.

[0094] In step three, interfaces for the high-voltage conversion board, supercapacitor board, and chip motherboard are reserved based on the number of intermediate voltage busbars. Hardware connection interfaces are planned based on the number of intermediate busbars: the voltage conversion module, i.e., the high-voltage conversion board, needs to reserve interfaces for connecting to the first DC bus (high-voltage input) and multiple second DC buses (intermediate voltage output); energy storage nodes, i.e., the supercapacitor board and computing nodes, i.e., the chip motherboard, need to reserve interfaces for mounting to the intermediate busbars. The number of chip motherboards can differ from the number of high-voltage conversion boards and supercapacitor boards, improving architectural flexibility.

[0095] In step four, the number and total number of supercapacitors connected in series are selected based on the intermediate bus voltage, required capacitance, and withstand voltage rating to form a supercapacitor plate. Taking a supercapacitor as an example, combined with the formula... = Calculate the number of series connections, then combine this with the total capacity requirement to calculate the total number of connections. The formula is as follows: .For example, =54V =3V、 When =1.2, =22 units; if the total capacitance requirement is 300F and the capacitance of each capacitor is 300F, then the total number is 484 units. This ensures that the voltage withstand capability of the supercapacitor module matches the intermediate bus voltage, while meeting the energy storage capacitance requirement, and realizing voltage peak regulation and emergency power supply.

[0096] In step five, a high-voltage power supply module with a high-voltage DC input and an intermediate bus voltage output is selected to form a high-voltage conversion board. High-voltage power supply modules adapted to both the first and second voltage levels are selected (e.g., 400V input, 54V output). Utilizing the space advantage of the high-voltage and intermediate voltage copper busbars, multiple high-efficiency modules are arranged to improve conversion efficiency and support power output from multiple intermediate busbars, ensuring voltage conversion stability and system power supply capacity.

[0097] In summary, the embodiments of this application have at least the following beneficial effects:

[0098] (1) Improved adaptability of energy storage nodes. A middle bus scheme for server high voltage DC (HVDC) architecture is proposed. The middle bus copper bus serves as a supercapacitor node for instantaneous high-power backup, peak shaving and energy storage. It provides physical mounting interfaces and electrical connection points for long-term power backup and peak shaving and valley filling, solving the problem that energy storage nodes cannot be effectively connected under high voltage HVDC input and realizing flexible adaptation of energy storage functions.

[0099] (2) Power supply safety is guaranteed. A two-level architecture and physical isolation design of "centralized high-voltage conversion and distributed medium and low voltage distribution" is adopted. The lines and conversion modules that carry high-voltage HVDC are completely separated from the intermediate bus copper bus and subsequent load areas that only carry medium and low voltage (such as 48V and 24V) in terms of physical space, wiring path and connectors. This achieves physical isolation between high-voltage power conversion and motherboard-level equipment, eliminates the safety risks such as arcing and breakdown caused by high voltage being directly introduced into the motherboard area, and ensures that the high-voltage DC is safely, efficiently and stably distributed to the server components after conversion.

[0100] (3) Effective optimization of power usage efficiency. By utilizing the large space between the high-voltage DC and the intermediate voltage copper busbar, a high-voltage power conversion board is arranged to perform a centralized and efficient "high-voltage to medium / low-voltage" primary conversion. Then, the board-level power module performs the final fine voltage regulation, reducing energy loss in the conversion process and thus improving the overall power usage efficiency of the data center.

[0101] The following will illustrate the method for generating the topology of a server power supply system through a specific embodiment. This method is used to generate the topology of the server power supply system described in the above embodiment, such as... Figure 9 As shown, the specific steps are as follows:

[0102] In step S901, at least one of the following is obtained: the second voltage level of the first DC bus, the first voltage level of the second DC bus, the current carrying capacity of the second DC bus, and the first required power of the load node.

[0103] It is understood that the embodiments of this application obtain these parameters, covering high-voltage input, low-voltage output, bus load capacity, and load power requirements, to lay the foundation for subsequent topology design. This ensures that the generation of the subsequent topology is based on actual voltage adaptation requirements, bus load capacity, and electrical load requirements, guaranteeing that the final generated topology can accurately match the core functional requirements of the power supply system, such as high-low voltage conversion, bus configuration, and load node power supply, thus laying a data foundation for the rationality and reliability of the topology.

[0104] In step S902, the number of power modules in the voltage conversion module is calculated based on the first voltage level and the second voltage level, and the number of second DC buses is calculated based on the first voltage level, the current carrying capacity level, and the first required power.

[0105] It is understood that the embodiments of this application determine the number of power modules by voltage level to match the high-low voltage conversion requirements, and calculate the number of buses by combining voltage, current rating, and load power to adapt to power supply carrying and distribution requirements. This ensures that the power modules can stably complete voltage conversion, guarantees that the second DC bus can safely carry power, and accurately distributes it to each load node. This provides a specific quantitative reference for the specific construction of the topology, making the entire system design more reasonable and adaptable.

[0106] In this embodiment of the application, the calculation of the number of second DC buses based on the first voltage level, current level, and first required power includes: calculating the power of the second DC bus based on the first voltage level and current level; calculating a preliminary number of second DC buses based on the first required power and the power of the second DC bus, and identifying whether the preliminary number is an integer. If the number of second DC buses is not an integer, the preliminary number of second DC buses is rounded up to obtain the final number of second DC buses.

[0107] It is understandable that the calculation of the number of second DC buses in this embodiment involves first calculating the bus power, then calculating a preliminary number, and rounding up non-integer values. This ensures that the bus quantity configuration both closely matches the actual carrying capacity and meets load requirements. First, the carrying power of a single bus is calculated based on the first voltage level and current carrying capacity. Then, a preliminary number is obtained by comparing this with the first power requirement of the load nodes. Rounding up avoids bus overload due to insufficient quantity. This ensures that the second DC bus can stably carry power while fully meeting the power supply needs of each load node, providing a rigorous calculation basis for the rational configuration of buses in the topology and guaranteeing the safety and reliability of the power supply system.

[0108] In this embodiment, the load node includes an energy storage node, the energy storage node includes a capacitor unit, and the capacitor module of the capacitor unit is provided with at least one supercapacitor. The method further includes: obtaining the withstand voltage rating of the supercapacitor provided in the capacitor module; identifying whether the withstand voltage rating is an integer; if the withstand voltage rating is not an integer, rounding the withstand voltage rating up; calculating the number of supercapacitors in the energy storage module based on the first voltage rating and the rounded withstand voltage rating; and calculating the number of energy storage modules based on the energy storage demand value and the energy storage value of the energy storage module.

[0109] It is understood that the embodiments of this application, by clarifying the calculation logic of the number of supercapacitors, including the rounding of the withstand voltage level, the calculation of the number of supercapacitors based on the first voltage level and the withstand voltage level, and the calculation of the number of modules combined with the energy storage requirements, achieve precise configuration of the capacitor modules in terms of voltage adaptation and energy storage requirements. This ensures that the capacitor modules can stably withstand the second voltage level, avoiding failures caused by withstand voltage mismatch, and can also accurately meet the energy storage regulation requirements of the server, thereby improving the rationality and reliability of the capacitor unit design.

[0110] It should be noted that the energy storage demand value refers to the total energy storage capacity requirement of the server power supply system for the energy storage units. For example, to meet short-term emergency power supply or voltage peak regulation, the system requires a total capacitor capacity of 300F. The energy storage value of the energy storage module refers to the energy storage capacity of a single energy storage module. Its value is related to the capacitance value of a single capacitor and the number of capacitors connected in series. Connecting capacitors in series will reduce the capacitance value of the module and is the basis for calculating the total number of energy storage modules required.

[0111] Specifically, taking supercapacitors as an example, such as Figure 7 As shown, on a supercapacitor board, several supercapacitors are connected in series to form a supercapacitor module. The series voltage division allows the supercapacitor module's withstand voltage to approach the voltage of the intermediate busbar copper bus, enabling it to be connected to the intermediate voltage busbar copper bus via a simple voltage conversion board. The number of supercapacitors within a single supercapacitor module... With intermediate bus voltage level Voltage rating of supercapacitors The relevant formula is... = Assuming the typical voltage rating of supercapacitors is 2.7-3.3V, take... 3V, intermediate bus voltage level The voltage rating is 54V. Because prolonged discharge will cause the supercapacitor's voltage rating to decrease, a voltage withstand factor is set for the supercapacitor. The value is 1.2, resulting in a calculated number of supercapacitors within the module of 21.6. Rounding up to 22, this is sufficient to meet the requirements for connecting to the intermediate voltage bus copper busbar. The total number of supercapacitors required is... It can be calculated using the following formula: For the required total number of capacitors, The capacitance value of a single supercapacitor is given by the formula: Taking a required capacitance of 300F as an example, the capacitance of a single supercapacitor is usually 300F. In the example above, there are 22 supercapacitors in a single module. Therefore, the total number of capacitors required is 484, which meets the required capacitance.

[0112] In step S903, the topology of the server power supply system is generated based on the number of power modules, the number of second DC buses, and the number of voltage conversion modules.

[0113] It is understood that, based on the quantities of power modules, second DC buses, and voltage conversion modules calculated in the first two steps, this application's embodiments transform the quantitative configuration of these key components into a specific server power supply system topology. This provides a clear quantitative basis for topology generation, ensuring that the generated topology accurately corresponds to the system's actual power conversion and distribution needs. Furthermore, by matching the number of voltage conversion modules, the connection logic between components is ensured to be reasonable. This guarantees that the generated topology accurately matches the actual application requirements.

[0114] For a description of the features in the embodiment corresponding to the topology generation method of the server power supply system, please refer to the relevant description of the embodiment corresponding to the server power supply system, which will not be repeated here.

[0115] The server power supply system topology generation method in this application obtains key parameters such as voltage level, current carrying capacity, and power requirements, and quantitatively calculates the number of core components such as power modules and the second DC bus, thereby generating a suitable topology. This provides clear data for topology design, ensuring that the number of each component accurately matches the requirements for high-low voltage conversion, power carrying capacity, and distribution. Furthermore, rigorous calculation logic, such as rounding off non-integer values, avoids insufficient configuration or redundancy. Ultimately, the generated topology highly matches the actual application scenario, improving the scientific nature and reliability of the system design and providing solid support for the physical construction and functional implementation of the power supply system.

[0116] Embodiments of this application also provide a non-volatile computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the steps of the server power supply system topology generation method as described in the above embodiments.

[0117] In one exemplary embodiment, the aforementioned non-volatile computer-readable storage medium may include, but is not limited to, various media that can store computer programs for a long time and whose data is not lost after power failure, such as USB flash drives, read-only memory (ROM), portable hard drives, magnetic disks, optical disks, and solid-state drives (SSDs).

[0118] The embodiments of this application also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in the server power supply system topology generation method embodiment described above.

[0119] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0120] The foregoing has provided a detailed description of a shared resource access method, apparatus, device, medium, and product provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are only intended to aid in understanding the method and core ideas of this application. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of the claims of this application.

Claims

1. A server power supply system, characterized in that, include: The first DC bus is used to provide the server with high-voltage DC power that is higher than the first preset voltage; At least one second DC bus, the second DC bus providing low-voltage DC power to the load nodes of the server at a voltage lower than a second preset voltage, wherein the first preset voltage is higher than the second preset voltage; At least one voltage conversion module is mounted on the first DC bus. The input terminal of the voltage conversion module is connected to the first DC bus, and the output terminal of the voltage conversion module is connected to at least one second DC bus. The voltage conversion module converts the high-voltage DC power into the low-voltage DC power. When the voltage conversion module is connected to multiple second DC buses, the voltage levels of the multiple second DC buses are the same. The load node is connected to at least one second DC bus, and the number of second DC buses connected to the load node is determined based on the first power requirement of the load node, the first voltage level and current carrying level of the second DC bus; The load node includes a supercapacitor, which is disposed on the second circuit board. The number and total number of supercapacitors connected in series on the second circuit board are determined according to the second bus voltage and the required capacitance and withstand voltage level to form a supercapacitor board. The supercapacitor board is disposed between the two voltage conversion modules.

2. The server power supply system according to claim 1, characterized in that, The voltage conversion module includes a first circuit board and a first interface, a power module array, and at least one second interface disposed on the first circuit board. The first interface is used to connect to the first DC bus, the second interface is used to connect to the second DC bus, and the power module array is used to convert the high voltage DC power into the low voltage DC power.

3. The server power supply system according to claim 2, characterized in that, The first interface is located on the first side of the first circuit board, and the second interface is located on the second side of the first circuit board. The first side and the second side are opposite sides.

4. The server power supply system according to claim 2 or 3, characterized in that, The power module array includes multiple power modules. The high-voltage port of the power module is connected to a first bus, and the low-voltage port of the power module is connected to a second bus. The first bus is connected to a first interface, and the second bus is connected to at least one second interface. The voltage level connected to the high-voltage port is higher than the voltage level connected to the low-voltage port.

5. The server power supply system according to claim 4, characterized in that, The number of power modules is determined based on the total system power of the server and the number of voltage conversion modules.

6. The server power supply system according to claim 2, characterized in that, The number of the second interfaces is determined based on the number of the second DC buses.

7. A server, characterized in that, include: The server power supply system according to any one of claims 1-6; At least one load node is connected to at least one second DC bus, the number of second DC buses connected to the load node being determined based on the first power requirement of the load node, the first voltage level and current carrying capacity of the second DC bus.

8. The server according to claim 7, characterized in that, The load node is at least one of an energy storage node and a computing node.

9. The server according to claim 8, characterized in that, The energy storage node includes at least one of a capacitor unit and a battery energy storage unit. The capacitor unit is used to regulate the peak voltage of the server, and the battery energy storage unit is used to provide backup power to the server.

10. The server according to claim 9, characterized in that, The capacitor unit includes a second circuit board, a capacitor module, a first voltage conversion device, and at least one third interface. The second circuit board is provided with the capacitor module, the first voltage conversion device, and at least one third interface. The capacitor module is connected to one end of the first voltage conversion device, and the third interface is connected to the other end of the first voltage conversion device. The third interface is connected to the second DC bus. The first voltage conversion device realizes voltage conversion between the capacitor module and the second DC bus.

11. The server according to claim 10, characterized in that, The capacitor module contains at least one supercapacitor, and the number of supercapacitors in the capacitor module is determined based on the first voltage level and the voltage rating of the capacitor.

12. The server according to claim 9, characterized in that, The battery energy storage unit includes a third circuit board, a battery module, a second voltage conversion device, and at least one fourth interface. The third circuit board is provided with at least one of the battery module, the second voltage conversion device, and the fourth interface. The battery module is connected to one end of the second voltage conversion device, and the fourth interface is connected to the other end of the second voltage conversion device. The fourth interface is connected to the second DC bus. The second voltage conversion device realizes voltage conversion between the battery module and the second DC bus.

13. The server according to claim 12, characterized in that, The battery module is equipped with at least one backup battery, and the number of backup batteries in the battery module is determined according to the first voltage level of the second DC bus and the capacity of the backup batteries.

14. The server according to claim 9, characterized in that, The computing node includes a fourth circuit board, at least one graphics processor, at least one central processing unit, a power supply module, and at least one fifth interface. The fourth circuit board is provided with at least one of the graphics processor, the central processing unit, the power supply module, and the fifth interface. One end of the power supply module is connected to at least one of the graphics processor and the central processing unit, and the other end of the power supply module is connected to the fifth interface. The fifth interface is connected to the second DC bus.

15. The server according to claim 14, characterized in that, The number of fifth interfaces is determined based on the first voltage level and the second power requirement of the computing node.

16. A method for generating topology structures, characterized in that, The method is used to generate the topology of the server power supply system according to any one of claims 1-6, wherein the method includes the following steps: Obtain at least one of the following: the second voltage level of the first DC bus, the first voltage level of the second DC bus, the current carrying capacity of the second DC bus, and the first power demand of the load node; Based on the first voltage level and the second voltage level, calculate the number of power modules in the voltage conversion module, and based on the first voltage level, the current carrying capacity and the first required power, calculate the number of the second DC bus; The topology of the server power supply system is generated based on the number of power modules, the number of the second DC bus, and the number of voltage conversion modules.

17. The topology generation method according to claim 16, characterized in that, The step of calculating the number of second DC buses based on the first voltage level, the first current rating, and the first power demand includes: Calculate the power of the second DC bus based on the first voltage level and the current rating; The initial number of the second DC bus is calculated based on the first required power and the power of the second DC bus, and it is identified whether the initial number is an integer. If the number of the second DC bus is not an integer, the initial number of the second DC bus is rounded up to obtain the final number of the second DC bus.

18. The topology generation method according to claim 17, characterized in that, The load node includes an energy storage node, the energy storage node includes a capacitor unit, the capacitor unit has at least one supercapacitor within its capacitor module, and further includes: Obtain the withstand voltage rating of the supercapacitor set within the capacitor module; It identifies whether the withstand voltage rating is an integer. If the withstand voltage rating is not an integer, it rounds up to the nearest integer. The number of supercapacitors in the energy storage module is calculated based on the first voltage level and the rounded withstand voltage level; the number of energy storage modules is calculated based on the energy storage demand and the energy storage value of the energy storage module.

19. A non-volatile computer-readable storage medium, characterized in that, The non-volatile computer-readable storage medium stores a computer program, wherein the computer program, when executed by a processor, implements the steps of the topology generation method as described in any one of claims 16 to 18.

20. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the topology generation method as described in any one of claims 16 to 18.

Images