Temperature control power supply cooperation method and electronic device

Abstract

This application discloses a temperature-controlled power supply coordination method and electronic device, relating to the field of temperature-controlled power supply coordination technology. The method includes real-time acquisition of node current data, node temperature data, and capacitor status information of multiple node servers; calculation of the output phase difference of the power supply unit based on the node current data, node temperature data, and capacitor status information, and generation of a first output current allocation parameter; determination of a first heat dissipation control parameter of the heat dissipation execution unit based on the node temperature data; joint analysis of the first output current allocation parameter and the first heat dissipation control parameter to obtain a temperature-power coordinated adjustment amount, and generation of a second output current allocation parameter and a second heat dissipation control parameter. This application enables dynamic coupling optimization of power supply and temperature control within the server chassis, ensuring a stable and balanced distribution of temperature and current in the node servers, thereby solving the problems of frequency reduction protection caused by instantaneous overheating of nodes and uneven temperature and current distribution between nodes in related technologies.

Description

Technical Field

[0001] This application relates to the field of temperature-controlled power supply coordination technology, and in particular to a temperature-controlled power supply coordination method and electronic device. Background Technology

[0002] In server systems, power supply and temperature control are critical aspects of ensuring stable server operation. Existing node servers typically distribute current to multiple computing nodes through a shared power module, supplemented by capacitor buffers to cope with short-term load fluctuations; temperature control systems mostly use fan arrays or liquid cooling plates to manage heat dissipation, regulating node temperatures through unified airflow or zoned temperature control.

[0003] These solutions are widely used in practical engineering, but they have certain limitations. For example, the response speed of the power supply system is usually in the microsecond range, while the response speed of the heat dissipation system is in the millisecond range, which is three orders of magnitude different. Under sudden load conditions, the nodes may overheat instantaneously, triggering frequency reduction protection and causing performance degradation. At the same time, in 4U or compact chassis, fixed phase interleaved or parallel power supply designs and overall heat dissipation systems are difficult to cope with asymmetrical loads and local thermal coupling effects, which can easily cause uneven temperature and current distribution between nodes. In addition, existing solutions lack analysis of the electro-thermal interaction within the server chassis, ignoring thermal interference, electrothermal effects and environmental heat absorption effects within the chassis, which can easily form a negative feedback loop, causing system oscillation, resulting in temperature control and power supply imbalance, and posing risks of local overheating and overcurrent. Summary of the Invention

[0004] This application provides a temperature control and power supply coordination method, device, and electronic equipment that can achieve dynamic coupling optimization of power supply and temperature control in server chassis, ensure stable and balanced distribution of temperature and current of node servers, and improve the overall reliability of the system. This aims to at least solve the problems of frequency reduction protection caused by instantaneous overheating of nodes and uneven temperature and current distribution between nodes in related technologies.

[0005] This application provides a temperature-controlled power supply coordination method for a control board in a temperature-controlled power supply coordination device. The temperature-controlled power supply coordination device also includes multiple node servers, a heat dissipation execution unit, and a power module. The power module includes a power supply unit and a capacitor unit. The method includes:

[0006] Real-time acquisition of node current data, node temperature data, and capacitor status information of capacitor units in power modules from multiple node servers;

[0007] Based on node current data, node temperature data, and capacitor status information, the output phase difference of the power supply unit in the power module is determined, and the first output current allocation parameters are generated.

[0008] Based on the node temperature data, determine the first heat dissipation control parameter of the heat dissipation execution unit;

[0009] Based on the chassis thermal coupling model, the first output current distribution parameters and the first heat dissipation control parameters are jointly analyzed to obtain the temperature-electricity coordinated adjustment amount.

[0010] The second output current distribution parameter and the second heat dissipation control parameter are generated based on the temperature and electricity coordination adjustment amount, so as to control the power supply unit and the heat dissipation execution unit respectively.

[0011] On the other hand, this application also provides an electronic device, including: a memory for storing a computer program; and a processor for implementing any of the above-mentioned temperature control and power supply coordination methods when executing the computer program.

[0012] This application achieves rapid control of node temperature and reduces the risk of instantaneous overheating by real-time acquisition of current, temperature, and capacitor status information of multiple node servers, calculating the output phase difference, and generating first output current distribution parameters and first heat dissipation control parameters. Simultaneously, by incorporating the mutual influence of current distribution and heat dissipation control into joint analysis, it effectively mitigates conflicts between power supply and heat dissipation targets and oscillations caused by negative feedback loops, achieving coordinated optimization of power supply and temperature control within the chassis. Furthermore, by dynamically generating and timely updating the second output current distribution parameters and the second heat dissipation control parameters, it adapts to uneven load distribution between nodes, local thermal coupling, and changes in the chassis environment, balancing the temperature and current distribution of each node, avoiding the risk of local overheating or overcurrent, and reducing the possibility of node power outages or fault propagation due to abnormal temperature or power consumption. Therefore, compared to traditional fixed-parameter control schemes, it enables dynamic adjustment of the control strategy, improving the reliability, stability, and overall performance maintenance capability of the server cluster under complex load and environmental conditions. Attached Figure Description

[0013] 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.

[0014] Figure 1 A flowchart of a temperature-controlled power supply coordination method provided in this application embodiment;

[0015] Figure 2 This is a schematic diagram of an output phase difference calculation method provided in an embodiment of this application;

[0016] Figure 3 A diagram illustrating the steps of temperature-controlled power supply coordination provided in this application embodiment;

[0017] Figure 4A structural diagram of a temperature control and power supply coordination device provided in an embodiment of this application;

[0018] Figure 5 A flowchart illustrating the execution of a temperature-controlled power supply coordination method provided in this application embodiment;

[0019] Figure 6 An electronic device diagram is 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] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0023] In one embodiment, such as Figure 1 As shown, this application provides a temperature-controlled power supply coordination method, including:

[0024] Step 101: Collect node current data, node temperature data, and capacitor status information of capacitor units in power modules from multiple node servers in real time.

[0025] Step 102: Based on the node current data, node temperature data, and capacitor status information, determine the output phase difference of the power supply unit in the power module and generate the first output current allocation parameters.

[0026] Step 103: Determine the first heat dissipation control parameters of the heat dissipation execution unit based on the node temperature data;

[0027] Step 104: Based on the chassis thermal coupling model, perform joint analysis on the first output current distribution parameters and the first heat dissipation control parameters to obtain the temperature-electricity coordinated adjustment amount;

[0028] Step 105: Generate the second output current distribution parameter and the second heat dissipation control parameter based on the temperature-electricity coordinated adjustment amount, so as to control the power supply unit and the heat dissipation execution unit respectively.

[0029] The embodiments of this application collect real-time data on the current, temperature, and capacitance status of multiple node servers, calculate the output phase difference, and generate first output current distribution parameters and first heat dissipation control parameters. This enables rapid control of node temperature under sudden load conditions, reducing the risk of instantaneous overheating. Simultaneously, by incorporating the mutual influence of current distribution and heat dissipation control into joint analysis, it effectively alleviates conflicts between power supply and heat dissipation targets and oscillations caused by negative feedback loops, achieving coordinated optimization of power supply and temperature control within the chassis system. Furthermore, by dynamically generating and timely updating the second output current distribution parameters and the second heat dissipation control parameters, it can adapt to uneven load distribution between nodes, local thermal coupling, and changes in the chassis environment, balancing the temperature and current distribution of each node, avoiding the risk of local overheating or overcurrent, and reducing the possibility of node power outages or fault propagation due to abnormal temperature or power consumption. Compared to traditional fixed-parameter control schemes, this approach enables dynamic adjustment of the control strategy, improving the reliability, stability, and overall performance maintenance capabilities of the server cluster under complex load and environmental conditions.

[0030] In one embodiment, such as Figure 2 As shown, based on node current data, node temperature data, and capacitor status information, the output phase difference of the power supply unit in the power module is determined, including:

[0031] The node current difference is obtained by analyzing the node current data, and the fundamental phase difference is determined based on the node current difference;

[0032] The node temperature data is parsed to obtain the node temperature difference, and a temperature compensation term is generated based on the node temperature difference.

[0033] The capacitor status information is analyzed to obtain the real-time voltage and rated voltage of the capacitor unit, and a capacitor compensation term is generated based on the ratio of the real-time voltage to the rated voltage.

[0034] The output phase difference is obtained by superimposing the basic phase difference, temperature compensation term, and capacitance compensation term.

[0035] Specifically, in this embodiment, when calculating the output phase difference of the power supply unit, the node current difference, node temperature difference, and the ratio of capacitor unit voltage to rated voltage are simultaneously analyzed, and these are converted into basic phase difference, temperature compensation term, and capacitor compensation term, respectively, and then superimposed on the output. Therefore, it can maintain unified quantification and linkage correction of phase under three operating conditions: unbalanced load, local overheating, and capacitor voltage drop. Compared with the scheme of controlling only current or fixed phase, it can reduce voltage ripple and phase imbalance probability and suppress transient drops caused by capacitor over-discharge, thereby improving power supply stability and anti-interference capability.

[0036] In one embodiment, multiple node servers are configured, including a first node server and a second node server. The output phase difference of the power supply unit in the power module is determined based on node current data, node temperature data, and capacitor status information, including:

[0037] The basic phase difference between the two nodes is determined based on the node current difference between the first and second node servers, and is expressed as:

[0038] ;

[0039] in, I represents the fundamental phase difference between two nodes. A I represents the real-time current of the first node server. B K represents the real-time current of the second node server. imbal K is the load imbalance coefficient, ranging from 0.8 to 1.2, used to adjust the base phase difference so that the phase compensation increases or decreases with changes in the load difference; imbal K can be calculated based on the following formula: imbal =0.8+0.4·tanh(2·I rated |I A -I B |);

[0040] Where tanh(·) is the hyperbolic tangent function, I rated This represents the node's rated current, a normalized reference value used to calculate the load imbalance coefficient.

[0041] A dual-node temperature compensation term is generated based on the node temperature difference between the first and second node servers, and is represented as follows:

[0042] ;

[0043] in, T represents the two-node temperature compensation term. A The node temperature (T) represents the temperature of the first node server. B The node temperature (T) represents the temperature of the second node server. thresh The threshold value for the temperature difference between the two nodes is 20℃, which is used to normalize the temperature difference so that the temperature compensation increases linearly. sgn(·) is the sign function used to determine the direction of the temperature difference.

[0044] and, The constraint condition is that when the node temperature difference is less than 5℃, the temperature compensation term is 0;

[0045] The capacitor compensation term is generated based on the ratio of the real-time voltage to the rated voltage of the capacitor unit, and is expressed as:

[0046] ;

[0047] in, Represented as capacitance compensation term, V cap This represents the real-time voltage of the capacitor cell, V. nom Indicates the rated voltage of the capacitor unit;

[0048] The dual-node output phase difference is obtained by superimposing the dual-node basic phase difference, the dual-node temperature compensation term, and the capacitance compensation term, and is expressed as:

[0049] ;

[0050] in, This indicates the phase difference between the two node outputs, and the phase difference between the two node outputs is between 150° and 210°.

[0051] Specifically, this embodiment provides a method for calculating the output phase difference of two nodes when two node servers are set in the chassis. It is applicable to scenarios with 4U or compact rack space and provides a theoretical basis for the rapid adjustment of the power supply unit in this scenario.

[0052] In one embodiment, generating the first output current allocation parameters includes:

[0053] Calculate the current adjustment coefficients for multiple node servers based on the output phase difference and the preset phase deviation sensitivity coefficient;

[0054] The current adjustment coefficient is normalized based on the current total load current of multiple node servers.

[0055] The normalized current adjustment coefficient is multiplied sequentially by the rated current of the corresponding node server to obtain the current reference values ​​of multiple node servers, which are then used as the first output current allocation parameter.

[0056] Specifically, in this embodiment, the current adjustment coefficient of each node is obtained by mapping the output phase difference through the phase deviation sensitivity. Based on this, the total load of multiple nodes is normalized and then multiplied with the rated current of each node to generate a current reference value. Therefore, the distribution ratio can be continuously adjusted under the premise of ensuring the conservation of total current and the constraint of the rated capacity of a single node. Furthermore, by establishing a monotonic and traceable deterministic mapping between "phase and current shunting", the secondary oscillation caused by over-distribution or under-distribution can be reduced, thereby improving the controllability and feasibility of current distribution.

[0057] In one embodiment, the current adjustment coefficients of multiple node servers are calculated based on the output phase difference and a preset phase deviation sensitivity coefficient, including:

[0058] Set the corresponding reference phase based on the node current phase of the node server in the current balance state;

[0059] Calculate the difference between the output phase difference and the reference phase to obtain the phase deviation.

[0060] Calculate the ratio of the phase deviation to the preset scaling factor, and multiply it by the preset phase deviation sensitivity coefficient to obtain the node phase correction value;

[0061] The current adjustment coefficient of the corresponding node server is obtained by adding the node phase correction value to the preset constant offset value.

[0062] In one specific embodiment, a first node server and a second node server are arranged in the chassis. The node current phase of the two in the current balance state is 180°, which represents the theoretical phase difference when the output voltages of the two nodes are completely interleaved and complementary in a dual-node interleaved power supply system. Therefore, the reference phase is set to 180°.

[0063] The preset scaling factor is set to 30°, which is used to map the phase deviation to the current adjustment coefficient. The greater the phase deviation, the more current is adjusted. The specific value can be adjusted according to the hardware characteristics, control accuracy and load regulation capability.

[0064] A preset phase deviation sensitivity coefficient is used to determine the current offset between two nodes. It can be set to [-1, 1]. Positive values ​​deviate from the first node server, and negative values ​​deviate from the second node server.

[0065] The current adjustment coefficient is specifically expressed as follows:

[0066] , ;

[0067] Where, k A k represents the current adjustment coefficient of the first node server. B α represents the current adjustment coefficient of the second node server, and α represents the preset phase deviation sensitivity coefficient.

[0068] Specifically, in this embodiment, the phase deviation is first calculated by taking the reference phase under current balance as a reference, and then the node phase correction value is obtained by scaling factor and phase deviation sensitivity coefficient. The final current adjustment coefficient is formed by constant offset. Therefore, fine-grained adjustment can be achieved in the small deviation range, and excessive amplification can be avoided in the large deviation range. At the same time, it ensures that the algorithm structure of each node is consistent and the scope is uniform, eliminating the implicit nonlinear uncertainty, which is convenient for calibration and online maintenance, thereby reducing adjustment jitter and improving convergence predictability.

[0069] In one embodiment, the heat dissipation execution unit includes a fan array consisting of multiple deflecting fans. Based on node temperature data, the first heat dissipation control parameters of the heat dissipation execution unit are determined, including:

[0070] Analyze node temperature data to determine the real-time temperature of multiple node servers, and compare them with preset temperature thresholds respectively;

[0071] If the real-time temperature of the node server is less than or equal to the temperature threshold, it is determined to be a normal temperature node, and multiple peripheral deflection fans are set to operate according to preset normal temperature operating parameters.

[0072] If the real-time temperature of a node server is greater than the temperature threshold, it is identified as an overheated node, and the difference between the real-time temperature of the overheated node and the temperature threshold is taken as the cooling requirement.

[0073] Multiple peripheral deflecting fans are selected based on the distribution of the fan array, and deflection angle adjustment parameters are generated to form an airflow convergence zone at the overheating node.

[0074] Based on the cooling demand and the rated power of multiple peripheral deflection fans, determine the speed adjustment parameters of multiple peripheral deflection fans to generate the target air volume corresponding to the cooling demand in the airflow convergence area.

[0075] The ambient temperature operating parameters, speed adjustment parameters, and deflection angle adjustment parameters of multiple deflecting fans in the fan array are integrated as the first heat dissipation control parameter.

[0076] Specifically, in this embodiment, the heat dissipation execution unit is constructed as a fan array composed of multiple deflecting fans. Based on the node temperature threshold, it distinguishes between normal temperature nodes and over-temperature nodes. For over-temperature nodes, it selects surrounding deflecting fans according to their spatial location to generate deflection angle adjustment parameters to form an airflow convergence area. Then, it calculates the speed adjustment parameters based on the cooling demand and the rated power of the fans. Finally, it integrates the normal temperature operation, angle, and speed adjustment parameters into the first heat dissipation control parameters. Therefore, without sacrificing the basic heat dissipation of normal temperature nodes, it can directionally improve the convective heat transfer coefficient at hot spots and shorten the hot spot cooling time. Compared with the unified air duct fixed speed scheme, it can reduce ineffective air volume and energy consumption and suppress local temperature superposition caused by heat backflow.

[0077] In a preferred embodiment, determining the first heat dissipation control parameter of the heat dissipation execution unit based on node temperature data further includes:

[0078] Determine if the node server is overheating, and use the difference between the real-time temperature of the overheating node and the temperature threshold as the cooling requirement, expressed as:

[0079] ;

[0080] Where ΔTi represents the cooling demand of the i-th node server, and Ti is the real-time temperature of the i-th node server. If the cooling demand is 0, the corresponding node server is a normal temperature node. The current operating parameters of the deflection fans around the normal temperature node are recorded as normal temperature operating parameters, and the current air volume is the target air volume.

[0081] Furthermore, in order to control the airflow, the cooling demand is mapped to the target airflow, expressed as:

[0082] ;

[0083] in K represents the target airflow of the i-th node server. i Let be the airflow-heat dissipation empirical coefficient for the i-th node server, representing the relationship between the temperature change of the corresponding node server and the required heat dissipation airflow;

[0084] Furthermore, in the fan array, the position of the j-th deflection fan is denoted as p. j Rated air volume is recorded as The rated speed is denoted as The location of the i-th node server within the chassis is denoted as n. i ;

[0085] Furthermore, taking the i-th node server as the center position, select M peripheral deflection fans from the fan array that are less than or equal to a preset distance threshold from the i-th node server. The preset distance threshold is based on the specific layout inside the chassis and the layout settings of the fan array.

[0086] When the deflection fan speed is set to S j The deflection angle is set to θ. j At that time, the effective airflow contribution of the j-th deflection fan to the i-th node server is expressed as:

[0087] ;

[0088] Among them, v ij (S j ,θ j ) represents the effective airflow contribution of the j-th deflection fan to the i-th node server, angle(n) i -p j ) represents the directional angle of the line connecting the j-th deflection fan and the i-th node server, D ij Let γ represent the distance between the j-th deflection fan and the i-th node server, where γ is the angle influence coefficient and is a positive empirical number. η is the distance attenuation coefficient for airflow within the chassis. j The fan efficiency coefficient is determined and set based on the fan's own performance and the influence of local airflow; therefore, the total airflow received by the i-th node server is:

[0089] ;

[0090] Where M is the number of peripheral deflection fans selected by the i-th node server, and V i This represents the total air volume received by the i-th node server.

[0091] Furthermore, when the number of overheating nodes is greater than one, to meet the heat dissipation requirements of each overheating node, it is necessary to adjust the deflection angle and speed of multiple peripheral deflection fans. The target angle for the peripheral deflection of the i-th node server is set as follows: The actual set angle is obtained by cutting within the mechanically adjustable range of the surrounding deflection fan, and is denoted as . When the same fan needs to serve multiple nodes at the same time, weights can be assigned according to the temperature difference between the nodes, and the target angle can be solved in a weighted manner, thereby achieving a compromise between different nodes;

[0092] After determining the actual set angles of multiple peripheral deflection fans, an airflow contribution matrix is ​​established, and the target airflow of multiple node servers is integrated. A matrix equation is constructed, and the fan speed vector is solved by constrained least squares to obtain the set speed of all deflection fans in the fan array. The first heat dissipation control parameters are then integrated to generate the first heat dissipation control parameters.

[0093] Specifically, this embodiment establishes a formulaic mapping relationship between node temperature difference, cooling demand, and target airflow, transforming heat dissipation control from empirical threshold control to quantifiable computational control. By incorporating parameters such as real-time node temperature, preset temperature threshold, fan rated power, and deflection angle into the formula, fan speed and deflection angle adjustments that match the actual needs of each node can be directly generated. This not only accurately locates the heat dissipation needs of overheated nodes but also ensures the matching between cooling demand and airflow supply at the mathematical model level, thereby achieving closed-loop prediction and adaptive adjustment of fan operating parameters. Compared to schemes that generate control parameters solely based on logical judgment, this embodiment avoids dependence on fixed thresholds and empirical parameters, improves the dynamic response speed and adjustment accuracy of heat dissipation control, and further enhances the system's temperature management capabilities under complex environments such as asymmetrical loads and local thermal coupling.

[0094] like Figure 3 As shown, in one embodiment, the first output current allocation parameter and the first heat dissipation control parameter are jointly analyzed to obtain the temperature-electricity coordinated adjustment amount, including:

[0095] Step 401: Obtain the node current allocation value of the node server according to the first output current allocation parameter, and generate the predicted power consumption change of the node server based on the node current allocation value.

[0096] Step 402: Based on the first heat dissipation control parameters, predict the predicted heat dissipation of the node server when the heat dissipation execution unit executes the first heat dissipation control parameters;

[0097] Step 403: Input the predicted power consumption change and predicted heat dissipation into the pre-built chassis thermal coupling model, and combine the chassis environment heat absorption factor to generate the node predicted temperature of the node server.

[0098] Step 404: Generate the node temperature deviation of the node server based on the difference between the predicted node temperature and the preset target node temperature.

[0099] Step 405: Calculate the temperature-electricity coordinated adjustment amount corresponding to the node server based on the node temperature deviation.

[0100] Specifically, in this embodiment, in the joint analysis, the current allocation value of each node is first obtained from the first output current allocation parameter, and the predicted power consumption change is obtained accordingly. Then, the predicted heat dissipation of each node when executing the first heat dissipation control parameter is obtained from the first heat dissipation control parameter. Both are input into the pre-built chassis thermal coupling model, and the heat absorption factor of the chassis environment is considered to output the node predicted temperature. Subsequently, the difference between the node predicted temperature and the node target temperature is used to form the node temperature deviation and calculate the temperature-electricity coordinated adjustment. Therefore, it can explicitly characterize the electrothermal interaction and the temperature modulation of the chassis environment. Before the parameters are issued, the conflict between power supply shunting and heat dissipation scheduling is eliminated in advance. Compared with the approach of optimizing the two independently, the probability of overshoot and undershoot can be reduced and the number of round-trip corrections can be reduced, thereby improving the one-time effectiveness of coordinated control.

[0101] In a preferred embodiment, generating the predicted power consumption change of the node server based on the node current allocation value includes: obtaining the equivalent resistance of the node server, and calculating the predicted power consumption change of the node server by combining the corresponding node current allocation value and the difference between the current of the node.

[0102] The predicted heat dissipation of the node server when the heat dissipation execution unit executes the first heat dissipation control parameter includes: simulating and recording the total air volume received by the node server when the heat dissipation execution unit executes the first heat dissipation control parameter, and combining the current air temperature difference in the chassis, air specific heat capacity, air-receiving area of ​​the node server and air-heat conversion efficiency factor to comprehensively analyze and obtain the predicted heat dissipation of the node server.

[0103] By inputting the predicted power consumption change and predicted heat dissipation into the pre-built chassis thermal coupling model, and combining the chassis environment heat absorption factor, the predicted node temperature of the node server is generated, which can be expressed as:

[0104] ;

[0105] Among them, T i (t) represents the current temperature of the i-th node server, t represents the current time, and t+1 represents the time of the next measurement cycle. Let ΔP represent the node-predicted temperature of the i-th node server. i C represents the predicted power consumption change of the i-th node server. i Let γ be the node thermal capacity of the i-th node server to reflect the buffering effect of power consumption on temperature rise. ik H is the thermal coupling coefficient between the i-th node server and the k-th node server to describe the interaction between airflow distribution and heat within the chassis. env (t) represents the heat absorption factor of the chassis environment, including the effects of external ambient temperature and heat absorption by the phase change energy storage layer, which can be modeled according to the actual chassis settings.

[0106] It is worth noting that, The saturation function is used to characterize the marginal effect of heat dissipation, and is preferably:

[0107] ;

[0108] in, Q represents the predicted heat dissipation of the k-th node server. sat The heat dissipation saturation value indicates that the additional cooling effect begins to significantly decrease at the corresponding heat dissipation level;

[0109] It is worth noting that the predicted temperature of a node server is not only determined by the heat generated by its own power consumption increment and the heat dissipation generated by its own heat dissipation execution unit, but is also affected by the coupling effect of the heat dissipation behavior of other nodes inside the chassis.

[0110] Therefore, in the thermal coupling function for predicting temperature, the product term of heat dissipation of the node set is used for modeling, that is... Where, when k=i, this term corresponds to the direct suppression effect of the node's own heat dissipation on its temperature; when k≠i, this term represents the indirect regulation effect of the heat dissipation distribution of adjacent or surrounding nodes on the node temperature through airflow interference or heat diffusion. Through the above modeling method, the electro-thermal interaction and overall thermal coupling effect between multiple nodes inside the chassis can be reflected more accurately, so that the predicted temperature results are more consistent with the actual operating state.

[0111] Furthermore, based on the obtained node predicted temperature values ​​of the node servers, corresponding node temperature deviation values ​​are generated, and the corresponding temperature-electricity coordinated adjustment amount is calculated based on the node temperature deviation values, including:

[0112] Based on the obtained node temperature deviation value, determine the required temperature correction direction and correction intensity for the corresponding node server. Specifically, when the predicted node temperature is higher than the preset temperature threshold, it is determined to be an overheated node and should be corrected by reducing the power supply current or increasing the heat dissipation intensity. Conversely, when the predicted node temperature is lower than the preset temperature threshold, the power supply current can be appropriately increased or the heat dissipation intensity can be reduced to maintain it in a reasonable operating temperature range.

[0113] The current adjustable margin is obtained by the difference between the node current allocation value and the upper and lower limits of the node rated current. The current adjustable margin is used to represent the space that the node server can still adjust upward or downward based on the existing current allocation.

[0114] Based on the difference between the predicted heat dissipation of the node and the preset upper and lower limits of the heat dissipation intensity of the node, a heat dissipation adjustable margin is generated. The heat dissipation adjustable margin is used to represent the margin that the node server can still increase or decrease under the current heat dissipation conditions.

[0115] Based on the adjustable current margin and adjustable heat dissipation margin, current tendency weight and heat dissipation tendency weight are generated, and combined with the node temperature deviation value, the initial current adjustment amount and the initial heat dissipation adjustment amount are generated.

[0116] Global constraints are set based on the total power supply current and total heat dissipation capacity in the chassis. Based on the global constraints, the initial current adjustment and the initial heat dissipation adjustment are scaled to generate the current calibration adjustment and the heat dissipation calibration adjustment, which are then integrated as the temperature and power synergistic adjustment.

[0117] Specifically, this embodiment enables coordinated control of power supply and heat dissipation, solving the problem of instantaneous overheating of nodes caused by fast power supply response but slow heat dissipation response in traditional systems; by adjusting the current and heat dissipation driven by node temperature deviation, it avoids uneven temperature distribution caused by asymmetrical loads and local thermal coupling under fixed phase power supply and unified heat dissipation strategies; global constraint processing ensures that the overall power and heat dissipation capacity of the system do not exceed the limit, avoiding system oscillation or overcurrent risks caused by local optimization; the final generated temperature and power coordinated adjustment amount can make the temperature of each node converge stably to the target value, improving the overall operational reliability and performance stability of the server.

[0118] In one embodiment, generating a second output current distribution parameter and a second heat dissipation control parameter based on the temperature-electricity coordinated adjustment amount includes:

[0119] By analyzing the temperature-electricity coordinated adjustment, we can obtain the node current fine-tuning and heat dissipation control fine-tuning.

[0120] The first output current distribution parameter of the power supply unit is corrected according to the node current fine-tuning amount to obtain the intermediate output current distribution parameter;

[0121] The first heat dissipation control parameter of the heat dissipation execution unit is corrected according to the heat dissipation control fine-tuning amount to obtain the intermediate heat dissipation control parameter;

[0122] The intermediate output current allocation parameters and intermediate heat dissipation control parameters are iteratively analyzed together. Each iteration uses the node predicted temperature and node current phase of the previous iteration as input until the node temperature deviation of multiple node servers is less than the stable threshold and the corresponding node current phase tends to stabilize.

[0123] The intermediate output current allocation parameters and intermediate heat dissipation control parameters obtained through iterative convergence are used as the second output current allocation parameters and the second heat dissipation control parameters.

[0124] Specifically, in this embodiment, the temperature-electricity coordinated adjustment amount is analyzed into current fine-tuning amount and heat dissipation fine-tuning amount to modify the first output current allocation parameter and the first heat dissipation control parameter respectively to obtain intermediate parameter pairs. Based on this, joint analysis is iteratively performed until the temperature deviation of each node is less than the stability threshold and the node current phase tends to be stable. Then, the intermediate parameter pairs obtained by convergence are output as the second output current allocation parameter and the second heat dissipation control parameter, thus forming a convergence process with a clear stopping criterion. Furthermore, the phase stability target and temperature stability target are simultaneously satisfied by iterative small-step linkage fine-tuning, avoiding oscillations caused by large-step alternating corrections, thereby improving the repeatability of the stable landing point and adjusting the convergence speed.

[0125] In one embodiment, the method further includes:

[0126] When one or more monitoring dimension values ​​exceed the preset threshold, or when the node server's runtime reaches the preset monitoring cycle, the second output current allocation parameter and the second heat dissipation control parameter are regenerated.

[0127] Specifically, this embodiment can adjust the operation of the power supply unit and heat dissipation unit in a timely manner by monitoring and updating the node server in real time, so as to respond promptly to the operating status inside the chassis.

[0128] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method.

[0129] like Figure 4 As shown, embodiments of this application also provide a temperature-controlled power supply coordination device, including a control board, multiple node servers, a heat dissipation execution unit, and a power module; wherein,

[0130] The power module includes a power supply unit and a capacitor unit;

[0131] The heat dissipation unit includes a fan array consisting of multiple deflecting fans;

[0132] The control board is configured to implement any of the above-mentioned temperature control and power supply coordination methods.

[0133] This embodiment provides a temperature control and power supply coordination device. By functionally partitioning and clearly interconnecting the control board, multiple node servers, heat dissipation execution unit, and power module, the control loop of acquisition-computation-execution can be closed in the hardware structure, achieving a system integration effect with short measurement and control paths and low feedback delay. Compared with loosely coupled external control, this structure facilitates online calibration and module replacement, and reduces the risk of control mismatch caused by interface inconsistencies.

[0134] In a preferred embodiment, the number of node servers is set to two, and the power supply unit is preferably a dual-channel interleaved pulse width modulation power supply, i.e., a PWM power supply, with an adjustable phase difference of 180°±5°.

[0135] Furthermore, the capacitor unit is preferably a supercapacitor group formed by connecting six Maxwell 2.7V / 3000F capacitors in parallel, using an active balancing circuit with a balancing current of 10A. The supercapacitor group has an electron spin resonance of less than 2mΩ at 25℃, a cycle life of more than 150,000 cycles, and an instantaneous power support of 50A / 100ms.

[0136] Furthermore, the fan array is preferably a magnetic levitation deflection fan array, which consists of 8 independent fan units, arranged with Halbach permanent magnet array (NdFeB N52 grade), driven by a carbon fiber rotor weighing 180g, with a maximum air volume of 45CFM±3%, a deflection angle range of 0-90°, an angle resolution of 0.1°, and a step response time of less than 50ms.

[0137] In one embodiment, the temperature-controlled power supply coordination device further includes a chassis; the chassis includes:

[0138] Box;

[0139] The air duct is located inside the enclosure and is connected to the heat dissipation unit.

[0140] Phase change thermal storage layer, which is installed on the wall of the air duct.

[0141] In a preferred embodiment, the phase change thermal storage layer uses paraffin wax with a phase change point of 45°C as the matrix, multilayer graphene (0.5wt%) as the reinforcing phase, and aluminum nitride nanoparticles as the thermal conductive material; its thermal conductivity can reach 68W / mK, its thermal storage density is 210J / g, and its cycle stability is greater than 10,000 times.

[0142] Specifically, in this embodiment, by further configuring the chassis, the built-in air duct and the phase change heat storage layer set on the wall of the air duct, the air duct and the fan array are connected to constrain and guide the airflow path. The phase change layer absorbs heat during the peak heat stage and releases heat during the slow release stage. Therefore, it can suppress the impact of short-term heat pulses on the node temperature and reduce the demand for sudden increases in air volume. It can achieve complementary upper-level collaborative control, smooth the temperature trajectory, expand the available adjustment margin for heat dissipation, and gain a response time window for the control algorithm under abnormal operating conditions.

[0143] like Figure 5 As shown, in one embodiment, the temperature control and power supply coordination process for the node servers within the chassis includes:

[0144] Sensors are installed inside the chassis to collect current data, temperature data, and capacitor unit status information from multiple node servers and transmit them to the control board.

[0145] The control board calculates the output phase difference of the power supply unit and generates the first output current allocation parameter based on the current data, temperature data and capacitor status information. It also generates the first heat dissipation control parameter based on the temperature data. After joint analysis, it obtains the temperature-electricity coordinated adjustment amount and generates the second output current allocation parameter and the second heat dissipation control parameter through iterative correction. The second output current allocation parameter is sent to the power supply unit and the second heat dissipation control parameter is sent to the heat dissipation execution unit.

[0146] The power supply unit adjusts the current according to the second output current distribution parameters;

[0147] The heat dissipation unit adjusts the fan speed and deflection angle according to the second heat dissipation control parameters;

[0148] During this process, the sensor continuously detects faults. If a fault is detected, it is fed back to the control board so that the control board can perform the corresponding fault handling operation. For example, when an abnormal current is detected, the control board sends a control command to the power module to enable the capacitor buffer.

[0149] like Figure 6 As shown, embodiments of this application also provide an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to perform the steps in any of the above-described temperature-controlled power supply coordination method embodiments.

[0150] 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.

[0151] The above provides a detailed description of a temperature-controlled power supply coordination method and electronic device provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only intended to help understand 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 this application.

Claims

1. A temperature-controlled power supply coordination method, characterized in that, A control board for a temperature-controlled power supply coordination device, the temperature-controlled power supply coordination device further comprising multiple node servers, a heat dissipation execution unit, and a power module, the power module comprising a power supply unit and a capacitor unit, the method comprising: The node current data, node temperature data, and capacitor status information of the capacitor unit are collected in real time from multiple node servers. Based on the node current data, the node temperature data, and the capacitor status information, the output phase difference of the power supply unit is determined, and the first output current allocation parameter is generated. Based on the node temperature data, the first heat dissipation control parameter of the heat dissipation execution unit is determined; Based on the chassis thermal coupling model, the first output current distribution parameters and the first heat dissipation control parameters are jointly analyzed to obtain the temperature-electricity coordinated adjustment amount; The second output current distribution parameter and the second heat dissipation control parameter are generated based on the temperature-electricity coordinated adjustment amount, so as to control the power supply unit and the heat dissipation execution unit respectively.

2. The temperature-controlled power supply coordination method according to claim 1, characterized in that, The step of determining the output phase difference of the power supply unit based on the node current data, the node temperature data, and the capacitor state information includes: The node current difference is obtained by parsing the node current data, and the fundamental phase difference is determined based on the node current difference; The node temperature data is parsed to obtain the node temperature difference, and a temperature compensation term is generated based on the node temperature difference; The capacitor state information is analyzed to obtain the real-time voltage and rated voltage of the capacitor unit, and a capacitor compensation term is generated based on the ratio of the real-time voltage to the rated voltage. The output phase difference is obtained by superimposing the basic phase difference, the temperature compensation term, and the capacitance compensation term.

3. The temperature-controlled power supply coordination method according to claim 1, characterized in that, Setting up multiple node servers, including a first node server and a second node server, and determining the output phase difference of the power supply unit based on the node current data, the node temperature data, and the capacitor status information includes: The basic phase difference between the two nodes is determined based on the node current difference between the first node server and the second node server, and is expressed as: ； in, I represents the basic phase difference between the two nodes. A I represents the real-time current of the first node server. B K represents the real-time current of the second node server. imbal The load imbalance coefficient is calculated based on the following formula: K imbal =0.8+0.4·tanh(2·I rated |I A -I B |), where tanh(·) is the hyperbolic tangent function, I rated Indicates the node's rated current; A dual-node temperature compensation term is generated based on the node temperature difference between the first node server and the second node server, and is expressed as follows: ； in, T represents the dual-node temperature compensation term. A The node temperature (T) represents the temperature of the first node server. B The node temperature (T) represents the temperature of the second node server. thresh This represents the two-node temperature difference threshold, and sgn(·) is the sign function used to determine the direction of the temperature difference; A capacitance compensation term is generated based on the ratio of the real-time voltage to the rated voltage of the capacitor unit, expressed as: ； in, Represented as the capacitance compensation term, V cap V represents the real-time voltage of the capacitor unit. nom This indicates the rated voltage of the capacitor unit; The dual-node output phase difference is obtained by superimposing the dual-node basic phase difference, the dual-node temperature compensation term, and the capacitance compensation term, and is expressed as: ； in, This indicates the output phase difference of the two nodes.

4. The temperature-controlled power supply coordination method according to claim 1, characterized in that, The parameters for generating the first output current allocation include: Based on the output phase difference and the preset phase deviation sensitivity coefficient, the current adjustment coefficients of the multiple node servers are calculated; The current adjustment coefficient is normalized based on the current total load current of the multiple node servers. The normalized current adjustment coefficient is multiplied sequentially by the rated current of the corresponding node server to obtain multiple current reference values ​​for the node servers, which are then used as the first output current allocation parameter.

5. The temperature-controlled power supply coordination method according to claim 4, characterized in that, The step of calculating the current adjustment coefficients of multiple node servers based on the output phase difference and a preset phase deviation sensitivity coefficient includes: Based on the node current phase of the node server in the current balance state, set the corresponding reference phase; The phase deviation is obtained by calculating the difference between the output phase difference and the reference phase. Calculate the ratio of the phase deviation to the preset scaling factor, and multiply it by the preset phase deviation sensitivity coefficient to obtain the node phase correction value; The node phase correction value is added to the preset constant offset value to obtain the current adjustment coefficient corresponding to the node server.

6. The temperature-controlled power supply coordination method according to claim 1, characterized in that, The heat dissipation execution unit includes a fan array composed of multiple deflecting fans. Determining the first heat dissipation control parameter of the heat dissipation execution unit based on the node temperature data includes: The node temperature data is analyzed to determine the real-time temperature of multiple node servers, and each temperature is compared with a preset temperature threshold. If the real-time temperature of the node server is less than or equal to the temperature threshold, it is determined to be a normal temperature node, and multiple peripheral deflection fans are set to operate according to preset normal temperature operating parameters. If the real-time temperature of the node server is greater than the temperature threshold, it is determined to be an overheated node, and the difference between the real-time temperature of the overheated node and the temperature threshold is taken as the cooling demand. Multiple peripheral deflection fans are selected based on the distribution of the fan array, and deflection angle adjustment parameters are generated to form an airflow convergence zone at the overheating node. Based on the cooling demand and the rated power of the multiple peripheral deflection fans, determine the speed adjustment parameters of the multiple peripheral deflection fans to generate a target airflow corresponding to the cooling demand in the airflow convergence area. The normal temperature operating parameters, speed adjustment parameters, and deflection angle adjustment parameters of the multiple deflecting fans in the fan array are integrated as the first heat dissipation control parameter.

7. The temperature-controlled power supply coordination method according to claim 1, characterized in that, The joint analysis of the first output current distribution parameters and the first heat dissipation control parameters to obtain the temperature-electricity coordinated adjustment amount includes: The node current allocation value of the node server is obtained based on the first output current allocation parameter, and the predicted power consumption change of the node server is generated based on the node current allocation value. Based on the first heat dissipation control parameters, predict the predicted heat dissipation of the node server when the heat dissipation execution unit executes the first heat dissipation control parameters; The predicted power consumption change and the predicted heat dissipation are input into the pre-built chassis thermal coupling model, and combined with the chassis environment heat absorption factor, the node predicted temperature of the node server is generated. The node temperature deviation of the node server is generated based on the difference between the predicted node temperature and the preset target node temperature. Based on the node temperature deviation, calculate the temperature-electricity coordinated adjustment amount corresponding to the node server.

8. The temperature-controlled power supply coordination method according to claim 6, characterized in that, The step of generating the second output current distribution parameter and the second heat dissipation control parameter based on the temperature-electricity coordinated adjustment amount includes: The temperature-electricity coordinated adjustment amount is analyzed to obtain the node current fine-tuning amount and the heat dissipation control fine-tuning amount; The first output current allocation parameter of the power supply unit is corrected according to the node current fine-tuning amount to obtain the intermediate output current allocation parameter; The first heat dissipation control parameter of the heat dissipation execution unit is corrected according to the heat dissipation control fine-tuning amount to obtain the intermediate heat dissipation control parameter; The intermediate output current allocation parameters and the intermediate heat dissipation control parameters are iteratively executed in the joint analysis. Each iteration uses the node predicted temperature and node current phase of the previous iteration as input until the node temperature deviation of multiple node servers is less than the stable threshold and the corresponding node current phase tends to stabilize. The intermediate output current allocation parameters and the intermediate heat dissipation control parameters obtained through iterative convergence are used as the second output current allocation parameters and the second heat dissipation control parameters.

9. The temperature-controlled power supply coordination method according to claim 1, characterized in that, The method further includes: When one or more monitoring dimension values ​​exceed a preset threshold, or when the node server's runtime reaches a preset monitoring cycle, the second output current allocation parameter and the second heat dissipation control parameter are regenerated.

10. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the temperature-controlled power supply coordination method as described in any one of claims 1 to 9.

Images