A power balance control method, device, equipment and medium

Abstract

The application discloses a power balance control method, device and equipment and a medium, and belongs to the technical field of servers. The method comprises the following steps: determining the average current of all power modules in a server according to the actual currents of the power modules, and obtaining the average current; determining the impedance of a path where a target power module is located according to the output voltage of the target power module and the actual current of the target power module, and obtaining the target impedance; determining the target difference between the actual current of the target power module and the average current, and judging whether the target difference is within a preset range; if not, determining the target correction amount required by the target power module according to the target impedance and the target difference, correcting the output voltage of the target power module according to the target correction amount, and continuing to perform the step of determining the average current of all power modules in the server according to the actual currents of the power modules until the target difference is within the preset range. Through the method, the current balance precision of each power module in the server can be improved.

Description

Technical Field

[0001] This invention relates to the field of server technology, and in particular to a power balancing control method, apparatus, device, and medium. Background Technology

[0002] As data center computing density increases, a single server typically requires multiple power modules connected in parallel to ensure its safety and reliability during operation. To extend the lifespan of each power module and prevent localized overheating, it is usually necessary to balance the load current of each power module within the server.

[0003] In related technologies, the drooping current method or the master-slave method are commonly used to balance the load current of each power module in a server. However, neither of these methods can detect impedance changes along the path of the power module; they rely solely on pre-set control parameters or strategies to balance the output current of each power module, resulting in poor current sharing accuracy. Currently, there is no effective solution to this technical problem.

[0004] Therefore, improving the current sharing accuracy of each power module in a server is a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide a power balancing control method, apparatus, device, and medium to solve the technical problem of poor accuracy in current sharing control of multiple power modules in a server in the prior art. The specific solution is as follows: To address the aforementioned technical problems, this invention provides a power balancing control method applied to a server's baseboard management controller; the server uses multiple power modules connected in parallel for power supply; including: The average current of all power modules in the server is determined based on the actual current of all power modules in the server, thus obtaining the mean current. The impedance of the path where the target power supply is located is determined based on the output voltage of the target power supply and the actual current of the target power supply, thus obtaining the target impedance; the target power supply is any one of multiple power modules. Determine the target difference between the actual current of the target power supply and the average current, and determine whether the target difference is within a preset range; If not, then determine the target correction amount required for the target power supply based on the target impedance and the target difference, and correct the output voltage of the target power supply based on the target correction amount. Continue to execute the step of determining the average current of all power modules in the server based on the actual current of all power modules in the server, until the target difference is within the preset range.

[0006] In one specific embodiment of this application, the step of determining the average current of all power modules in the server based on the actual current of all power modules in the server, and obtaining the average current, includes: The actual current of all power modules in the server is collected using a shunt and an analog-to-digital converter, and the average current of all power modules in the server is determined based on the actual current of all power modules in the server, thus obtaining the average current.

[0007] In one specific embodiment of this application, determining the impedance of the path where the target power source is located based on the output voltage and actual current of the target power source, and obtaining the target impedance, includes: Determine whether the server is in a light-load running state or an idle state; If not, continue with the step of determining whether the server is in a light-load running state or an idle state; If so, the output voltage of the target power supply is read via the power management bus; The actual current of the target power supply is acquired using the shunt and the analog-to-digital converter, and the load bus voltage of the server is read. Based on the impedance calculation model, and by determining the impedance of the path where the target power supply is located according to the output voltage of the target power supply, the load bus voltage of the server, and the actual current of the target power supply, the target impedance is obtained. The expression for the impedance calculation model is as follows: ; In the formula, The target impedance is... The output voltage of the target power supply. This refers to the load bus voltage of the server. The actual current of the target power supply.

[0008] In one specific embodiment of this application, it further includes: The impedance of the path where the target power source is located is stored according to a preset period to obtain target stored data; If the target stored data indicates that the connector or cable connected to the target power supply is abnormal, a warning message will be displayed. If the target power supply is determined to be abnormal based on the target stored data, the target power supply will be switched to standby mode.

[0009] In one specific embodiment of this application, determining the target correction amount required for the target power supply based on the target impedance and the target difference includes: Based on the proportional-integral-derivative control algorithm, the target correction amount required for the target power supply is determined according to the target impedance and the target difference.

[0010] In one specific embodiment of this application, the step of determining the target correction amount required for the target power supply based on the proportional-integral-derivative control algorithm and according to the target impedance and the target difference includes: Based on the target model, and according to the target impedance and the target difference, the target correction amount required for the target power supply is determined; The expression for the target model is: ; In the formula, This is the target correction amount. This is the proportionality coefficient. The target difference, The target impedance is denoted as .

[0011] In one specific embodiment of this application, the proportional coefficient is dynamically set based on the target difference and the load power of the server.

[0012] To address the aforementioned technical problems, the present invention also provides a power balancing control device applied to a server's baseboard management controller; the server uses multiple power modules connected in parallel for power supply; comprising: The mean value determination module is used to determine the average current of all power modules in the server based on the actual current of all power modules in the server, and obtain the mean current. An impedance determination module is used to determine the impedance of the path where the target power supply is located based on the output voltage of the target power supply and the actual current of the target power supply, thereby obtaining the target impedance; the target power supply is any one of a plurality of power supply modules. The difference judgment module is used to determine the target difference between the actual current of the target power supply and the average current, and to determine whether the target difference is within a preset range. The current adjustment module is used to determine the target correction amount required by the target power supply based on the target impedance and the target difference when the judgment result of the difference judgment module is negative. At the same time, it corrects the output voltage of the target power supply based on the target correction amount and continues to execute the step of determining the average current of all power modules in the server based on the actual current of all power modules in the server until the target difference is within the preset range.

[0013] To address the aforementioned technical problems, the present invention also provides a power balancing control device, comprising: Memory, used to store computer programs; A processor is configured to implement the steps of a power balancing control method as disclosed above when executing the computer program.

[0014] To address the aforementioned technical problems, the present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of a power balancing control method as disclosed above.

[0015] Beneficial Effects: In the power balancing control method provided by this invention, the server uses multiple power modules connected in parallel for power supply. The baseboard management controller in the server determines the average current of all power modules in the server based on the actual current of all power modules in the server, obtaining the average current. Then, it determines the impedance of the path where the target power supply is located based on the output voltage of the target power supply and the actual current of the target power supply, obtaining the target impedance. Here, the target power supply is any one of the multiple power modules. After that, it determines the target difference between the actual current of the target power supply and the average current, and judges whether the target difference is within a preset range. If the target difference is not within the preset range, the baseboard management controller determines the target correction amount required for the target power supply based on the target impedance and the target difference, and corrects the output voltage of the target power supply based on the target correction amount, and continues to execute the step of determining the average current of all power modules in the server based on the actual current of all power modules in the server, until the target difference is within the preset range.

[0016] Compared to existing technologies, the method described in this invention can measure the impedance changes along the paths of all power modules in the server, and adjust and control the output current of each power module based on the impedance changes along each power module's path and the actual current of each power module. This ensures that the controlled output of each power module matches its actual operating state, thereby significantly improving the current sharing accuracy of each power module in the server. Furthermore, this method is directly controlled by the baseboard management controller in the server, further enhancing the baseboard management controller's global optimization capabilities when controlling the entire server.

[0017] Furthermore, by using a shunt and analog-to-digital converter to determine the actual current of each power module in this invention, the current sampling accuracy of the power modules can be further improved. The impedance calculation model allows for a more accurate calculation of the impedance along the path of the target power supply. Moreover, the proportional-integral-derivative (PID) control algorithm not only accurately calculates the target correction amount required by the target power supply, but also, by dynamically adjusting and setting the proportional coefficient in the PID algorithm, it can balance the response speed of the baseboard management controller when correcting the voltage of each power module with the stability of server operation.

[0018] Correspondingly, the power balancing control device, equipment, and medium provided by the present invention also have the above-mentioned beneficial effects. Attached Figure Description

[0019] To more clearly illustrate the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 A flowchart of a power balancing control method provided in an embodiment of the present invention; Figure 2 This is a hardware structure diagram of a server performing flow sharing control according to an embodiment of the present invention. Figure 3 A comparison chart showing the results of using the drooping method and the control method described in this application to perform current sharing control on each power module in the server; Figure 4 This is a structural diagram of a power balancing control device provided in an embodiment of the present invention; Figure 5 This is a structural diagram of a power balancing control device provided in an embodiment of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the protection scope of the present invention.

[0022] Please see Figure 1 , Figure 1 A flowchart of a power balancing control method provided in an embodiment of the present invention, the method comprising: Step S11: Determine the average current of all power modules in the server based on the actual current of all power modules in the server, and obtain the mean current; Step S12: Determine the impedance of the path where the target power supply is located based on the output voltage and actual current of the target power supply, and obtain the target impedance; the target power supply is any one of the multiple power supply modules. Step S13: Determine the target difference between the actual current and the average current of the target power supply, and determine whether the target difference is within the preset range; Step S14: If not, determine the target correction amount required for the target power supply based on the target impedance and the target difference, and correct the output voltage of the target power supply based on the target correction amount. Then continue to execute step S11: determine the average current of all power modules in the server based on the actual current of all power modules in the server, until the target difference is within the preset range.

[0023] This embodiment provides a power balancing control method, which is specifically described using the Baseboard Management Controller (BMC) in a server as the executing entity. The server is powered by multiple Power Supply Units (PSUs) connected in parallel. The Baseboard Management Controller communicates with each power supply unit via a Power Management Bus (PMBus).

[0024] In this control method, the baseboard management controller first periodically collects the actual current of all power modules in the server, and then determines the average current of all power modules in the server based on the actual current of all power modules in the server, thus obtaining the average current.

[0025] Assuming the server has One power module, and this The actual current of each power module is respectively , , ... Then the baseboard management controller can obtain to The average current is obtained by averaging the average values.

[0026] Subsequently, the baseboard management controller reads the output voltage reported internally by the target power supply via the power management bus. , ( (and will be based on the output voltage of the target power supply) and the actual current of the target power supply Determine the impedance of the path containing the target power source to obtain the target impedance. The target impedance can be approximated as the sum of the physical contact impedances, including the cable resistance, connector resistance, and PCB (Printed Circuit Board) trace resistance, from the target power output terminal to the load junction point.

[0027] Then, the baseboard management controller calculates the target difference between the actual current and the average current of the target power supply, and determines whether the target difference is within a preset range. The preset range is a numerical range set in advance based on the server's average accuracy requirements.

[0028] If the target difference is within the preset range, it indicates that the output current of the target power supply is close enough to the average current and has reached the current sharing target. In this case, monitoring of the target power supply can continue without performing subsequent voltage correction processes. The board management controller can then correct the output voltage of the next power module as required or wait for the next round of voltage correction.

[0029] If the target difference is not within the preset range, it means that the output current of the target power supply deviates significantly from the average current. At this time, the board management controller will correct the output voltage of the target power supply so that the output current of the target power supply can be within the preset range, thereby achieving the current sharing target.

[0030] Specifically, the board management controller determines the target correction amount required by the target power supply based on the target impedance and target difference, and simultaneously corrects the output voltage of the target power supply according to the target correction amount. If the output current of the target power supply is too high, the board management controller will issue a command to the target power supply to reduce the target correction amount, thereby reducing its own output value. If the output current of the target power supply is too low, the board management controller will issue a command to the target power supply to increase the target correction amount, thereby increasing its own output value.

[0031] After the baseboard management controller has corrected the output voltage of the target power supply, and the output voltage of all power modules in the server has reached a relatively stable state, the baseboard management controller will continue to execute step S11, re-collect the actual current of all power modules in the server, and recalculate the average current of all power modules based on the actual current of all power modules to obtain a new average current. After that, the baseboard management controller will continue to correct the power modules whose output current exceeds the preset range, thus forming a closed-loop iterative control logic, until the difference between the actual current and the average current of all power modules is within the preset range. This achieves the purpose of high-precision current sharing of multiple power modules in the server.

[0032] It should be noted that in this application, the current sharing control of each power module in the server is entirely the responsibility of the baseboard management controller, which enables the baseboard management controller to perform more centralized active closed-loop control of each power module.

[0033] Compared to existing technologies, the method described in this embodiment can measure the impedance changes along the paths of all power modules in the server, and adjust and control the output current of each power module based on the impedance changes along the paths of each power module and the actual current of each power module. This ensures that the controlled output of each power module matches its actual operating state, thereby significantly improving the current sharing accuracy of each power module in the server. Furthermore, this method is directly controlled by the baseboard management controller in the server, further enhancing the global optimization capability of the baseboard management controller when controlling the server as a whole.

[0034] Based on the above embodiments, this embodiment further explains and optimizes the technical solution. As a preferred implementation, the above steps: determining the average current of all power modules in the server based on the actual current of all power modules in the server, and obtaining the average current, include: The actual current of all power modules in the server is collected using a shunt and an analog-to-digital converter, and the average current of all power modules in the server is determined based on the actual current of all power modules in the server, thus obtaining the average current.

[0035] To more accurately acquire the actual current of all power modules in the server, a shunt can be pre-connected between the output terminal of each power module and the motherboard load, and an analog-to-digital converter (ADC) can be embedded inside the baseboard management controller. The shunt and ADC are then used to acquire the actual current of all power modules in the server.

[0036] It should be noted that the resistance range of the shunt is [missing information]. Measurement accuracy is Within this range, the baseboard management controller uses an analog-to-digital converter to acquire the differential voltage across each shunt in real time at a sampling frequency higher than 1kHz. By combining this with the known resistance values ​​of the shunts, the actual current of all power modules in the server can be calculated. Using this measurement method, the current output accuracy of the power modules can reach [percentage missing]. .

[0037] In practical applications, Hall effect sensors can also be used to replace shunts; however, the measurement accuracy of Hall effect sensors at room temperature is limited. Compared to a shunt, the measurement accuracy of a Hall sensor is poor. Furthermore, the measurement accuracy of a Hall sensor is also affected by factors such as magnetoresistance and temperature. Therefore, in this embodiment, a shunt is chosen to collect the actual current of the power module.

[0038] In addition, the cost of a shunt is lower than that of a Hall sensor. Therefore, when using a shunt to collect the actual current of all power modules in a server, the cost required to sample the actual current of the power modules can be further reduced.

[0039] Obviously, the technical solution provided in this embodiment can not only collect the output current value of each power module more accurately, but also reduce the cost of sampling each power module.

[0040] In a preferred embodiment, the above steps, including determining the impedance of the path where the target power supply is located based on the output voltage and actual current of the target power supply to obtain the target impedance, include: Determine whether the server is in a lightly loaded or idle state; If not, continue with the steps to determine whether the server is in a light-load or idle state. If so, the output voltage of the target power supply is read through the power management bus; The actual current of the target power supply is collected using a shunt and an analog-to-digital converter, and the load bus voltage of the server is read. Based on the impedance calculation model, and by determining the impedance of the path where the target power supply is located according to the output voltage of the target power supply, the load bus voltage of the server, and the actual current of the target power supply, the target impedance is obtained. The expression for the impedance calculation model is as follows: ; In the formula, For the target impedance, The output voltage of the target power supply. This refers to the load bus voltage of the server. This represents the actual current of the target power supply.

[0041] In this embodiment, to improve server resource utilization, the impedance of each power module path can be calibrated when the server is in a light-load or idle state. When the server is in a light-load or idle state, the baseboard management controller reads the output voltage of the target power supply through the power management bus. .

[0042] At the same time, the board management controller also uses shunts and analog-to-digital converters to acquire the actual current of the target power supply. It will also read the server's load bus voltage via the power management bus. Then, the baseboard management controller will output the target power supply voltage. Server load bus voltage and the actual current of the target power supply The input is fed into the impedance calculation model, and the model's output is used to determine the impedance of the path containing the target power source, thus obtaining the target impedance. .

[0043] It should be noted that in practical applications, in order to further improve the target impedance The calculation accuracy can also be improved in calculating target impedance. At this time, only the target power supply is turned on, and all other power supplies in the server except the target power supply are turned off. This can eliminate the influence of other power supply modules in the server on the calculated impedance value of the path where the target power supply is located.

[0044] Furthermore, this application utilizes an impedance calculation model to automatically measure the impedance of each power module's path within the server, and can also transform hardware aging and manufacturing tolerances into quantifiable digital models. Moreover, the baseboard management controller calculates the required target correction amount for the target power supply based on the target difference between the actual current and the average current of the target power supply, as well as the impedance along the target power supply's path. It then issues millivolt-level differentiated voltage regulation commands to the target power supply via the power management bus, enabling more precise current sharing control of each power module within the server.

[0045] Clearly, the technical solution provided in this embodiment can improve the resource utilization of the server and accurately calculate the impedance of the path where the target power supply is located.

[0046] Based on the above embodiments, this embodiment further explains and optimizes the technical solution. As a preferred implementation, the above control method further includes: The impedance of the path where the target power source is located is stored according to a preset period to obtain the target stored data; If the target stored data indicates that the connector or cable connected to the target power supply is abnormal, a warning message will be displayed. If the target power supply is determined to be abnormal based on the target stored data, the target power supply will be switched to standby mode.

[0047] In this embodiment, to enable the substrate management controller to more clearly understand the real-time operating status of the target power supply, the impedance of the path where the target power supply is located can be stored according to a preset period to obtain target stored data. For example, the preset period can be set to 1 day. In practical applications, the target stored data can be stored in the non-volatile memory of the substrate management controller to ensure the security and reliability of the target stored data during data storage.

[0048] Specifically, if the baseboard management controller determines, based on the target stored data, that the impedance of the path of the target power supply exceeds the first preset threshold for 30 consecutive days, it indicates that the connector or cable connected to the target power supply is abnormal. At this time, a warning message can be issued so that the staff can repair or replace the connector or cable connected to the target power supply.

[0049] Alternatively, if the board management controller determines, based on the target stored data, that the impedance change of the path of the target power supply exceeds the first preset slope over several consecutive days, it can also indicate that there is an abnormality in the connector or cable connected to the target power supply. In this case, the board management controller can also issue a warning message.

[0050] If the baseboard management controller determines, based on the target storage data, that the impedance of the path of the target power supply exceeds the second preset threshold on a certain day, it indicates that the target power supply has malfunctioned. In order to avoid safety accidents, the target power supply can be switched to standby mode, and the server load can be gradually transferred to other power modules to achieve soft isolation of the target power supply and thus avoid sudden power outages caused by contact deterioration.

[0051] Alternatively, if the board management controller determines, based on the target stored data, that the impedance change of the target power supply on its path exceeds the second preset slope on a certain number of days, it can also be determined that the target power supply has malfunctioned. In this case, the board management controller can also switch the target power supply to standby mode.

[0052] Clearly, the technical solution provided in this embodiment can further ensure the security and reliability of the server during operation.

[0053] In addition, in practical applications, the impedance of the paths of various power modules in the server will change over time, which will generate a large amount of impedance calculation data. In this case, after the target storage data is used up, it can be deleted to avoid the target storage data occupying storage space in the server.

[0054] Based on the above embodiments, this embodiment further explains and optimizes the technical solution. As a preferred implementation, the above step of determining the target correction amount required by the target power supply based on the target impedance and the target difference includes: Based on the proportional-integral-derivative control algorithm, the target correction amount required by the target power supply is determined according to the target impedance and the target difference.

[0055] Because the proportional-integral-derivative (PID) control algorithm is not only simple and easy to implement, but also has strong robustness, it only requires adjusting the relevant operating parameters of the target power supply according to the actual output value of the target power supply, so that the target power supply can eventually converge to the desired operating state.

[0056] Therefore, in this embodiment, a proportional-integral-derivative (PID) control algorithm is used, and the target correction amount required for the target power supply is determined based on the target impedance and the target difference. Specifically, a PID controller can be added to the server, and the PID controller can calculate the target correction amount required for the target power supply based on the target impedance and the target difference.

[0057] Obviously, the technical solution provided in this embodiment allows for a relatively simple method to accurately calculate the target correction amount required by the target power supply.

[0058] As a preferred embodiment, the above steps, including: determining the target correction amount required by the target power supply based on the proportional-integral-derivative control algorithm and the target impedance and target difference, include: Based on the target model, the target correction amount required for the target power supply is determined according to the target impedance and the target difference; The expression for the target model is: ; In the formula, For the target correction amount, This is the proportionality coefficient. For the target difference, The target impedance.

[0059] In this embodiment, to accurately calculate the target correction amount required by the target power supply, a mathematical model corresponding to the target correction amount required by the target power supply can be created based on the proportional-integral-differential algorithm, thus obtaining the target model. Then, by substituting the target impedance and target difference into the target model, the target correction amount required by the target power supply can be calculated using the target model.

[0060] Obviously, the technical solution provided in this embodiment can be used to calculate the target correction amount required by the target power supply more accurately using a mathematical model.

[0061] As a preferred implementation, the scaling factor is dynamically set based on the target difference and the server's load power.

[0062] In practical applications, to balance the response speed of the baseboard management controller when correcting the voltage of each power module and the stability of server operation, a target difference can also be used. The scaling factor is dynamically set based on the server's load power. .

[0063] if Then Setting it to 1 allows the board management controller to more quickly pull the actual current output value of the target power supply to the average current level. if Then Setting it to 0.6 allows the board management controller to more smoothly transition the actual current output value of the target power supply to the average current level; if Then Setting it to 0.3 allows the board management controller to more precisely adjust the actual current output value of the target power supply to the average current level, preventing overshoot.

[0064] If the server is under light load (the server's load power is less than 20% of the server's rated power), then... Increase it appropriately, for example, by Multiplying by 1.2 allows the baseboard management controller to address the issue of insufficient correction due to the low average current across all power modules in the server.

[0065] Obviously, the technical solution provided in this embodiment allows for a more reasonable setting and adjustment of the scaling factor.

[0066] Based on the above embodiments, this embodiment further explains and optimizes the technical solution. As a preferred implementation, the above control method further includes: The complex programmable logic device (CPL) reports its own heartbeat signal to the server in real time, so that when the CPL detects an abnormality in the baseboard management controller, it locks the enable pin state of all power modules in the server; when the CPL detects a faulty power module in the server, it disconnects the faulty power module from the server.

[0067] In practical applications, in order to further improve the overall reliability of the server during operation, the Complex Programmable Logic Device (CPLD) in the server will also monitor the heartbeat signal of the baseboard management controller and the Power Good signal of each power module in the server in real time.

[0068] If the complex programmable logic device (CPL) detects that the heartbeat signal of the baseboard management controller has been lost for more than a preset time (e.g., 2 seconds), the CPL will lock the enable pin states of all power modules in the server to remain unchanged, thus allowing the server to continue operating and preventing safety incidents. At the same time, the CPL will also provide corresponding warning messages, enabling maintenance personnel to be promptly aware of the baseboard management controller malfunction and perform timely repairs.

[0069] If a complex programmable logic device (CPL) detects the disappearance of the Power Good signal of a power module, it may be due to an overcurrent in that power module. In this case, the CPL will pull down the power enable pin of the faulty power module and disconnect the faulty power module from the server, thereby protecting the motherboard hardware in the server.

[0070] Clearly, the technical solution provided in this embodiment can further improve the security of the server during operation.

[0071] To enable those skilled in the art to more clearly understand the implementation principle of the present invention, the foregoing disclosure is specifically described here through an application scenario embodiment.

[0072] Please see Figure 2 , Figure 2This is a hardware structure diagram of a server for current sharing control according to an embodiment of the present invention. The server includes a baseboard management controller, a complex programmable logic device (CPL), a first power module, a second power module, and a motherboard load. The baseboard management controller also embeds an analog-to-digital converter (ADC). Shunts Rs1 and Rs2 are connected in series to the output terminals of the first and second power modules, respectively. The shunts have a resistance of 0.25Ω and an accuracy of 0.1%. The differential signals at both ends of the shunts are connected to the ADC inside the baseboard management controller.

[0073] The I2C / SMBus (Inter-Integrated Circuit / System Management Bus) interface on the baseboard management controller is connected to the first and second power modules via the PMBus protocol. Furthermore, the baseboard management controller can read and write the internal registers of both the first and second power modules. The power supply plane of the motherboard load is connected to the outputs of two shunts. The motherboard load includes, but is not limited to, CPU (central processing unit), memory, and PCIe (peripheral component interconnect express) devices.

[0074] Complex programmable logic devices (CPLs) are connected to the baseboard management controller via GPIO (General Purpose Input Output). Furthermore, CPLs can also control the enable pins of various power modules on the server.

[0075] When the server powers on for the first time or when the server recalibrates the impedance of each power module path due to current sharing control logic, the baseboard management controller will perform the following steps: Step 1: When the server is in a light-load or idle state (CPU usage less than 5%), the baseboard management controller records the output voltage of the first power module and the output voltage of the second power module, respectively obtaining... and .

[0076] Simultaneously, the baseboard management controller also measures the actual current of the first power module and the actual current of the second power module through the shunt, and obtains the results respectively. and Furthermore, the baseboard management controller also reads the server's load bus voltage to obtain... .

[0077] Step 2: The substrate management controller will calculate the impedance of the paths containing the first power module and the second power module based on the impedance calculation model, and obtain... and ; ; .

[0078] Step 3: The baseboard management controller will... and Store the data in Flash memory and mark the corresponding measurement time.

[0079] Step 4: Assume the server load has increased, and the total current has reached 100A. Ideally, the actual current drawn by the first and second power modules should be 50A. Because... If the current sharing between the first and second power modules is achieved solely by the drooping method, the actual current of the first power module may be 48A, and the actual current of the second power module may be 52A, with a deviation of 4%.

[0080] In this scenario, the baseboard management controller obtains the actual currents of the first and second power modules respectively through the shunt. and Then the baseboard management controller will calculate the average current of the first power module and the second power module as follows: ,and and The difference between them is , and The difference between them is .

[0081] Step 5: The baseboard management controller determines the error range ( Choose a smaller scaling factor And from this, the voltage correction amounts required for the first power module and the second power module are calculated as follows: and ; ; .

[0082] Step Six: The baseboard management controller sends a "voltage margin reduced" message to the first power module via the power management bus. The command "increase voltage margin" was sent to the second power module. The instruction is "".

[0083] Step 7: After waiting 200ms, resample the actual current of the first and second power modules. The reason for waiting 200ms is to allow the first and second power modules time to respond to the commands issued by the baseboard management controller.

[0084] Subsequently, the actual currents of the first and second power modules became 49.4A and 50.6A, respectively, and the deviation between the first and second power modules and the ideal average current was reduced to 1.2%. Under these circumstances, the output currents of the first and second power modules were iteratively adjusted according to the above steps until the deviation between the first and second power modules and the ideal average current was less than 1%.

[0085] Step 8: After the server has been running for a period of time, the baseboard management controller uses recorded impedance data to indicate that the impedance of the path where the first power module is located has decreased. Gradually rise to Meanwhile, the impedance of the path where the second power module is located remains stable. At this time, the baseboard management controller can send a warning message to the maintenance personnel through the server's web page: "The impedance of the power supply path of the first power supply module has increased abnormally. It is recommended to check the power connector."

[0086] When the maintenance personnel replaced the connection cable for the first power module during the planned maintenance window, the impedance of the path containing the first power module returned to normal. This avoids power supply failures caused by poor contact.

[0087] It should be noted that, assuming a different scenario, the server's total current is 150A, and the baseboard management controller obtains the actual currents of the first and second power modules through the shunt, respectively. and ,so and The difference between them is , and The difference between them is At this point, the baseboard management controller will select a larger scaling factor based on the error magnitude. And from this, the voltage correction amounts required for the first power module and the second power module are calculated as follows: and .

[0088] in, ; .

[0089] Then, the baseboard management controller sends a "voltage margin increase" message to the first power module via the power management bus. The command "voltage margin reduced" was sent to the second power module. The instruction is "".

[0090] After the baseboard management controller corrects the output voltages of the first and second power modules once, the output current of the first power module becomes 73A and the output current of the second power module becomes 77A. The difference between the actual current and the ideal average current of the first and second power modules is reduced to 2.7%, entering the small error range. Under these circumstances, the proportional coefficient can be adjusted. Set it to 0.3 to further finely adjust the output current of the first and second power modules.

[0091] Please see Figure 3 , Figure 3 This is a comparison chart showing the results of current sharing control of various power modules in a server using the drooping method and the control method described in this application. Figure 3 In the diagram, the horizontal axis represents time, and the vertical axis represents current deviation. Curve 1 represents current sharing control of the power modules in the server using the method described in this application, and curve 2 represents current sharing control of the power modules in the server using the drooping method. Figure 3 The results show that the current sharing control accuracy of the control method described in this application is significantly higher than that of the drooping method. The control method described in this application ensures that each power module in the server always operates within its optimal efficiency range, thus significantly reducing the overall power consumption of the server.

[0092] Please see Figure 4 , Figure 4 This is a structural diagram of a power balancing control device provided in an embodiment of the present invention. The device includes: The mean value determination module 21 is used to determine the average current of all power modules in the server based on the actual current of all power modules in the server, and obtain the mean current. Impedance determination module 22 is used to determine the impedance of the path where the target power supply is located based on the output voltage of the target power supply and the actual current of the target power supply, so as to obtain the target impedance; the target power supply is any one of a plurality of power supply modules; The difference judgment module 23 is used to determine the target difference between the actual current of the target power supply and the average current, and to determine whether the target difference is within a preset range. The current adjustment module 24 is used to determine the target correction amount required by the target power supply based on the target impedance and the target difference when the judgment result of the difference judgment module is negative. At the same time, it corrects the output voltage of the target power supply based on the target correction amount and continues to execute the step of determining the average current of all power modules in the server based on the actual current of all power modules in the server until the target difference is within the preset range.

[0093] In one specific embodiment of this application, the mean determination module 21 includes: The average current determination unit is used to collect the actual current of all power modules in the server using a shunt and an analog-to-digital converter, and determine the average current of all power modules in the server based on the actual current of all power modules in the server, thereby obtaining the average current.

[0094] In one specific embodiment of this application, the impedance determination module 22 includes: A status determination unit is used to determine whether the server is in a light-load running state or an idle state. The status monitoring unit is used to continue executing the step of determining whether the server is in a light-load running state or an idle state when the determination result of the status judgment unit is negative. A voltage reading unit is used to read the output voltage of the target power supply through the power management bus when the determination result of the state judgment unit is yes. The bus reading unit is used to collect the actual current of the target power supply using the shunt and the analog-to-digital converter, and to read the load bus voltage of the server. An impedance calculation unit is used to determine the impedance of the path where the target power supply is located based on an impedance calculation model, the output voltage of the target power supply, the load bus voltage of the server, and the actual current of the target power supply, so as to obtain the target impedance. The expression for the impedance calculation model is as follows: ; In the formula, The target impedance is... The output voltage of the target power supply. This refers to the load bus voltage of the server. The actual current of the target power supply.

[0095] In one specific embodiment of this application, it further includes: The data storage module is used to store the impedance of the path where the target power source is located according to a preset period to obtain target stored data; The information prompting module is used to provide a warning message if it is determined from the target stored data that the connector or cable connected to the target power supply is abnormal. The power switching module is used to switch the target power supply to standby mode if it is determined from the target stored data that the target power supply is abnormal.

[0096] In one specific embodiment of this application, the current adjustment module 24 includes: The correction calculation unit is used to determine the target correction amount required by the target power supply based on the proportional-integral-derivative control algorithm and the target impedance and the target difference.

[0097] In one specific embodiment of this application, the current adjustment unit includes: The correction calculation subunit is used to determine the target correction amount required for the target power supply based on the target model and according to the target impedance and the target difference; The expression for the target model is: ; In the formula, This is the target correction amount. This is the proportionality coefficient. The target difference, The target impedance is denoted as .

[0098] The power balancing control device provided in this embodiment of the invention has the beneficial effects of the power balancing control method disclosed above.

[0099] Please see Figure 5 , Figure 5 This is a structural diagram of a power balancing control device provided in an embodiment of the present invention. The device includes: Memory 31 is used to store computer programs; The processor 32 is configured to implement the steps of a power balancing control method as disclosed above when executing the computer program.

[0100] The processor 32 may include one or more processing cores, such as a quad-core processor or an octa-core processor. The processor 32 may be implemented using at least one hardware form of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). The processor 32 may also include a main processor and a coprocessor. The main processor, also known as the Central Processing Unit (CPU), is used to process data in the wake-up state; the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, the processor 32 may integrate a Graphics Processing Unit (GPU), which is responsible for rendering and drawing the content to be displayed on the screen. In some embodiments, the processor 32 may also include an Artificial Intelligence (AI) processor, which handles computational operations related to machine learning.

[0101] The memory 31 may include one or more computer-readable storage media, which may be non-transitory. The memory 31 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices or flash memory devices. In this embodiment, the memory 31 is used to store at least the following computer program 301, which, after being loaded and executed by the processor 32, can implement the relevant steps of a power balancing control method disclosed in any of the foregoing embodiments. In addition, the resources stored in the memory 31 may also include an operating system 302 and data 303, and the storage method may be temporary or permanent storage. The operating system 302 may include Windows, Unix, Linux, etc. The data 303 may include, but is not limited to, the data involved in the power balancing control method.

[0102] In some embodiments, the power balancing control device may further include a display screen 33, an input / output interface 34, a communication interface 35, a power supply 36, and a communication bus 37.

[0103] It is understood that if the methods in the above embodiments are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the current technology, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and executes all or part of the steps of the methods in the various embodiments of the present invention. The aforementioned storage medium includes: USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), electrically erasable programmable ROM, registers, hard disks, removable disks, CD-ROMs, magnetic disks, or optical disks, and other media capable of storing program code.

[0104] The power balancing control device provided in this embodiment of the invention has the beneficial effects of the power balancing control method disclosed above.

[0105] Accordingly, embodiments of the present invention also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of a power balancing control method as disclosed above.

[0106] The computer-readable storage medium provided in this embodiment of the invention has the beneficial effects of the power balancing control method disclosed above.

[0107] The power balancing control method, apparatus, device, and medium provided by the present invention have been described in detail above. The various embodiments in the specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section. It should be noted that those skilled in the art can make several improvements and modifications to the present invention without departing from the principles of the invention, and these improvements and modifications also fall within the protection scope of the present invention.

[0108] It should also be noted that, in this specification, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, 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. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

Claims

1. A power balancing control method, characterized in that, A baseboard management controller for a server; the server is powered by multiple power modules connected in parallel; including: The average current of all power modules in the server is determined based on the actual current of all power modules in the server, thus obtaining the mean current. The impedance of the path where the target power supply is located is determined based on the output voltage of the target power supply and the actual current of the target power supply, thus obtaining the target impedance; the target power supply is any one of multiple power modules. Determine the target difference between the actual current of the target power supply and the average current, and determine whether the target difference is within a preset range; If not, then determine the target correction amount required for the target power supply based on the target impedance and the target difference, and correct the output voltage of the target power supply based on the target correction amount. Continue to execute the step of determining the average current of all power modules in the server based on the actual current of all power modules in the server, until the target difference is within the preset range.

2. The power balancing control method according to claim 1, characterized in that, The step of determining the average current of all power modules in the server based on their actual currents, and obtaining the average current, includes: The actual current of all power modules in the server is collected using a shunt and an analog-to-digital converter, and the average current of all power modules in the server is determined based on the actual current of all power modules in the server, thus obtaining the average current.

3. The power balancing control method according to claim 2, characterized in that, The step of determining the impedance of the path containing the target power source based on the output voltage and actual current of the target power source, to obtain the target impedance, includes: Determine whether the server is in a light-load running state or an idle state; If not, continue with the step of determining whether the server is in a light-load running state or an idle state; If so, the output voltage of the target power supply is read via the power management bus; The actual current of the target power supply is acquired using the shunt and the analog-to-digital converter, and the load bus voltage of the server is read. Based on the impedance calculation model, and by determining the impedance of the path where the target power supply is located according to the output voltage of the target power supply, the load bus voltage of the server, and the actual current of the target power supply, the target impedance is obtained. The expression for the impedance calculation model is as follows: ； In the formula, The target impedance is... The output voltage of the target power supply. This refers to the load bus voltage of the server. The actual current of the target power source.

4. The power balancing control method according to claim 1, characterized in that, Also includes: The impedance of the path where the target power source is located is stored according to a preset period to obtain target stored data; If the target stored data indicates that the connector or cable connected to the target power supply is abnormal, a warning message will be displayed. If the target power supply is determined to be abnormal based on the target stored data, the target power supply will be switched to standby mode.

5. The power balancing control method according to claim 1, characterized in that, The step of determining the target correction amount required for the target power supply based on the target impedance and the target difference includes: Based on the proportional-integral-derivative control algorithm, the target correction amount required for the target power supply is determined according to the target impedance and the target difference.

6. The power balancing control method according to claim 5, characterized in that, The method based on the proportional-integral-derivative (PID) control algorithm, and determining the target correction amount required for the target power supply according to the target impedance and the target difference, includes: Based on the target model, and according to the target impedance and the target difference, the target correction amount required for the target power supply is determined; The expression for the target model is: ； In the formula, This is the target correction amount. This is the proportionality coefficient. The target difference, The target impedance is denoted as .

7. The power balancing control method according to claim 6, characterized in that, The scaling factor is dynamically set based on the target difference and the server's load power.

8. A power balancing control device, characterized in that, A baseboard management controller for a server; the server is powered by multiple power modules connected in parallel; including: The mean value determination module is used to determine the average current of all power modules in the server based on the actual current of all power modules in the server, and obtain the mean current. An impedance determination module is used to determine the impedance of the path where the target power supply is located based on the output voltage of the target power supply and the actual current of the target power supply, thereby obtaining the target impedance; the target power supply is any one of a plurality of power supply modules. The difference judgment module is used to determine the target difference between the actual current of the target power supply and the average current, and to determine whether the target difference is within a preset range. The current adjustment module is used to determine the target correction amount required by the target power supply based on the target impedance and the target difference when the judgment result of the difference judgment module is negative. At the same time, it corrects the output voltage of the target power supply based on the target correction amount and continues to execute the step of determining the average current of all power modules in the server based on the actual current of all power modules in the server until the target difference is within the preset range.

9. A power balancing control device, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the steps of a power balancing control method as described in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of a power balancing control method as described in any one of claims 1 to 7.

Images