A manifold shunt system and liquid separation method with adaptive heat distribution of electronic components

Abstract

The present invention relates to the field of manifold shunt technology, and discloses a manifold shunt system and liquid separation method with adaptive thermal distribution of electronic components. The system includes a temperature acquisition module, a cluster analysis module, a target heat exchange function construction module, a target space layout determination module, a cold air demand determination module, and a heat dissipation execution module, wherein: the temperature acquisition module is used to obtain the real-time temperature of the electronic components in the integrated circuit chip; the cluster analysis module is used to calculate the difference between the real-time temperature and the real-time temperature of the adjacent components of the electronic component to obtain the temperature gradient of the electronic component, and cluster analysis is performed on the electronic components based on the temperature gradient and the real-time temperature to obtain multiple heat load characteristic clusters of the integrated circuit chip; the present invention improves the heat dissipation efficiency of the electronic components in the integrated circuit chip by optimizing the shunt layout of the manifold.

Description

Technical Field

[0001] The present invention relates to the technical field of manifold shunting, and in particular to a manifold shunting system and a liquid separation method with adaptive heat distribution of electronic components. Background Art

[0002] In the field of integrated circuit chip thermal management, as the integration density and power density of electronic components continue to increase, the thermal field within the chip exhibits significant temporal and spatial non-uniformity. Traditional manifold branches use a pre-set static layout and equal flow distribution strategy to dissipate heat from electronic components.

[0003] First, the static layout lacks the ability to adjust the layout of the heat load group distribution within the integrated circuit chip. The manifold branches cannot be dynamically adjusted according to the real-time heat distribution of the electronic components. When the heat load of the integrated circuit chip is unevenly distributed, the manifold heat dissipation efficiency will be significantly reduced, causing some areas to overheat and other areas to be overcooled. Secondly, traditional systems mostly adopt an equal diversion strategy, that is, the cold air flow rate of each branch remains consistent. However, due to the significant differences in the heat load distribution of electronic components, a constant flow rate will lead to insufficient heat dissipation in high-load areas, while causing resource waste in low-load areas, reducing the heat dissipation efficiency of the manifold. Summary of the Invention

[0004] The present invention provides a manifold shunt system and a liquid separation method with adaptive heat distribution of electronic components, the main purpose of which is to solve the problem in the prior art that the heat dissipation efficiency of the manifold is reduced when the manifold shunt layout is used to dissipate heat for circuit chips.

[0005] To achieve the above objectives, the present invention provides a manifold shunting system for adaptive heat distribution of electronic components, characterized in that the system includes a temperature acquisition module, a cluster analysis module, a target heat exchange function construction module, a target space layout determination module, a cold air demand determination module, and a heat dissipation execution module, wherein:

[0006] The temperature acquisition module is used to obtain the real-time temperature of the electronic components in the integrated circuit chip;

[0007] The cluster analysis module is configured to calculate the difference between the real-time temperature and the real-time temperature of adjacent components of the electronic component to obtain a temperature gradient of the electronic component, and perform cluster analysis on the electronic component based on the temperature gradient and the real-time temperature to obtain a plurality of thermal load characteristic clusters of the integrated circuit chip;

[0008] The target heat exchange function construction module is configured to obtain the cross-sectional area and cold air flow rate of each branch pipe in the manifold, obtain the heat dissipation parameters of the electronic component, calculate the target heat load of the heat load characteristic cluster based on the heat dissipation parameters, construct the target heat exchange function of the integrated circuit chip based on the cross-sectional area, the cold air flow rate, the target heat load, and the position information of the heat load characteristic cluster, and generate the constraint conditions of the manifold according to the physical limitation parameters of the branch pipe;

[0009] The target spatial layout determination module is configured to optimize the parameters of the target heat exchange function based on a gradient descent algorithm under the constraints to obtain a minimum value of the target heat exchange function, determine the optimal parameters of the target heat exchange function based on the minimum value, determine the target position of the branch pipe based on the two-dimensional coordinates in the optimal parameters, determine the target inclination angle of the branch pipe based on the pitch angle in the optimal parameters, and determine the target spatial layout of the manifold based on the target position and the target inclination angle;

[0010] The cold air demand determination module is configured to calculate the branch cold air demand of the branch pipe based on the target heat load;

[0011] The heat dissipation execution module is configured to dissipate heat for the electronic components corresponding to the heat load characteristic cluster based on the target space layout and the branch cold air demand.

[0012] Preferably, when the cluster analysis module performs cluster analysis on the electronic components based on the temperature gradient and the real-time temperature to obtain a plurality of thermal load characteristic clusters of the integrated circuit chip, the cluster analysis module includes:

[0013] Performing feature fusion on the temperature gradient and the real-time temperature to obtain a two-dimensional feature vector of the electronic component, and aggregating the two-dimensional feature vectors into a vector data set of the integrated circuit chip;

[0014] Cluster analysis is performed on the vector data set based on a K-means clustering algorithm to obtain multiple thermal load feature clusters of the integrated circuit chip.

[0015] Preferably, when the target heat exchange function construction module calculates the target heat load of the heat load characteristic cluster based on the heat dissipation parameters, it includes:

[0016] approximating the power consumption data of the heat dissipation parameter to the thermal energy data of the electronic component;

[0017] Calculating a cluster total thermal energy of the heat load characteristic cluster based on the thermal energy data;

[0018] Determining the intra-cluster heat dissipation efficiency of the heat load characteristic cluster according to the heat dissipation efficiency of the component of the heat dissipation parameter;

[0019] The target heat load of the heat load characteristic cluster is calculated based on the heat dissipation efficiency within the cluster and the total heat energy of the cluster, wherein the calculation formula of the target heat load is as follows: Where, Indicates the The total cluster thermal energy of the heat load characteristic cluster, Indicates the The heat dissipation efficiency within the cluster of the heat load characteristic cluster, Indicates the The target heat load of each heat load characteristic cluster, Indicates the ID of the heat load signature cluster.

[0020] Preferably, the target heat exchange function is as follows: Where, represents the target heat exchange function, represents the parameters of the target heat exchange function, Indicates the identifier of the heat load characteristic cluster, represents the number of heat load characteristic clusters, Indicates the The target heat load of each heat load characteristic cluster, Indicates the branch pipe identifier. Indicates the number of branches, represents the heat transfer efficiency coefficient, represents the cold air flow of the i-th branch, Indicates the The cross-sectional area of ​​each branch pipe, Indicates the The pitch angle of each branch pipe, Indicates the The horizontal coordinate of the two-dimensional coordinates of the branch, Indicates the The vertical coordinate of the two-dimensional coordinates of the branch, Indicates the The horizontal coordinate in the position information of the target heat load is Indicates the The vertical coordinate in the location information of the target heat load.

[0021] Preferably, the constraints are as follows: Where, Indicates the The horizontal coordinate of the two-dimensional coordinates of the branch, Indicates the The vertical coordinate of the two-dimensional coordinates of the branch, Indicates the branch pipe identifier. Indicates the The horizontal coordinate of the two-dimensional coordinates of the branch, Indicates the The vertical coordinate of the two-dimensional coordinates of the branch, Indicates the minimum spacing constraint between branches. Indicates the branch pipe identifier. Indicates the The pitch angle of each branch pipe, Indicates the lower limit of the pitch angle, Indicates the upper limit of the pitch angle.

[0022] Preferably, when the target space layout determination module optimizes the parameters of the target heat exchange function based on the gradient descent algorithm to obtain the minimum value of the target heat exchange function, it includes:

[0023] The parameters of the target heat exchange function are optimized using the update formula of the gradient descent algorithm, wherein the update formula is as follows: Where, Indicates the The parameters at the iteration, Indicates the The parameters at the iteration, represents the learning rate, Indicates that the target heat exchange function is The gradient at The identifier of the iteration, represents the target heat exchange function;

[0024] When the change of the target heat exchange function is less than a preset change threshold, a minimum value of the target heat exchange function is obtained.

[0025] Preferably, when determining the target spatial layout of the manifold according to the target position and the target inclination angle, the target spatial layout determining module includes:

[0026] Installing the branch pipes according to the target positions to obtain a preliminary spatial layout of the branch pipes;

[0027] Under the preliminary spatial layout, the initial inclination angles of the branch pipes are adjusted according to the target inclination angles to obtain the target spatial layout of the manifolds.

[0028] Preferably, the calculation formula for the branch cold air demand is as follows: Where, Indicates the The branch cooling air demand of each branch pipe, Indicates the branch pipe identifier. Indicates the The heat transfer efficiency coefficient of each branch pipe is Indicates the The target heat load of each heat load characteristic cluster, Indicates the identifier of the heat load characteristic cluster, Indicates the The pitch angle of each branch pipe, Indicates the The horizontal coordinate of the two-dimensional coordinates of the branch, Indicates the The vertical coordinate of the two-dimensional coordinates of the branch, Indicates the The horizontal coordinate in the position information of the target heat load is Indicates the The vertical coordinate in the location information of the target heat load.

[0029] Preferably, when the heat dissipation execution module performs heat dissipation for the electronic components corresponding to the heat load characteristic cluster based on the target space layout and the branch cold air demand, it includes:

[0030] calculating a total cold air demand of the manifold based on the branch cold air demands;

[0031] delivering cold air to the manifold according to the total cold air demand;

[0032] Calculating the branch cold air injection time of the manifold according to the branch cold air demand and the cold air flow rate;

[0033] Under the target spatial layout, the branch pipes dissipate heat for the electronic components corresponding to the heat load characteristic cluster according to the branch cold air injection time.

[0034] In order to solve the above problems, the present invention further provides a manifold liquid distribution method for electronic component thermal distribution self-adaptation, the method comprising:

[0035] S1. Obtaining the real-time temperature of electronic components in the integrated circuit chip;

[0036] S2. Calculating the difference between the real-time temperature and the real-time temperatures of adjacent components of the electronic component to obtain a temperature gradient of the electronic component, and performing cluster analysis on the electronic components based on the temperature gradient and the real-time temperature to obtain multiple thermal load characteristic clusters of the integrated circuit chip;

[0037] S3. Obtaining the cross-sectional area and cold air flow rate of each branch pipe in the manifold, obtaining the heat dissipation parameters of the electronic component, calculating the target heat load of the heat load characteristic cluster based on the heat dissipation parameters, constructing the target heat exchange function of the integrated circuit chip based on the cross-sectional area, the cold air flow rate, the target heat load, and the position information of the heat load characteristic cluster, and generating the constraint conditions of the manifold according to the physical limitation parameters of the branch pipe;

[0038] S4. Under the constraints, optimize the parameters of the target heat exchange function based on a gradient descent algorithm to obtain a minimum value of the target heat exchange function, determine the optimal parameters of the target heat exchange function based on the minimum value, determine the target position of the branch pipe based on the two-dimensional coordinates of the optimal parameters, determine the target inclination angle of the branch pipe based on the pitch angle of the optimal parameters, and determine the target spatial layout of the manifold based on the target position and the target inclination angle;

[0039] S5. Calculating the branch cold air demand of the branch pipe based on the target heat load;

[0040] S6. Cooling the electronic components corresponding to the heat load characteristic cluster based on the target space layout and the branch cold air demand.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] 1. This invention analyzes the temperature data of electronic components in an integrated circuit chip, taking into account temperature gradients and temperature differences between adjacent components, and divides the electronic components into multiple heat load characteristic clusters. This allows accurate identification of the heat load distribution in different areas of the integrated circuit chip. Based on the physical properties of the manifold branches and the heat dissipation requirements of the electronic components, a target heat exchange function is constructed, optimizing heat exchange efficiency. This target heat exchange function is solved using an optimization algorithm to determine the optimal position and tilt angle of the manifold branches, ensuring that the branch layout can best adapt to the heat load distribution on the chip and improving the manifold's heat dissipation efficiency.

[0043] 2. The present invention calculates the branch cold air demand of each branch pipe by the target heat load of the heat load characteristic cluster, accurately reflecting the actual heat that the electronic components need to dissipate; the heat dissipation execution module installs the branch pipe according to the target spatial layout to ensure that the cold air can efficiently exchange heat with the heat load characteristic cluster. The target spatial layout takes into account the optimal position and inclination angle of the branch pipe to maximize the heat dissipation efficiency, and dynamically adjusts the cold air release amount and injection time according to the real-time temperature changes and the heat dissipation conditions of the heat load characteristic cluster to achieve continuous dynamic heat dissipation adjustment; through this collaborative working mode, the system can dissipate heat according to the actual situation of the heat load group distribution in the integrated circuit chip, avoiding the uneven heat dissipation problem caused by the equal diversion strategy in the traditional manifold heat dissipation system, and improving the heat dissipation efficiency of the manifold. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 A system architecture diagram of a manifold shunt system for adaptive heat distribution of electronic components provided by one embodiment of the present invention;

[0045] Figure 2 A schematic flow chart of a manifold liquid distribution method for adaptive thermal distribution of electronic components provided by one embodiment of the present invention.

[0046] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0047] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments belong to some embodiments of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0048] The terms used in the embodiments of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The singular forms "said" and "the" used in the embodiments of the present invention and the appended claims are also intended to include plural forms, unless the context clearly indicates otherwise, and "a plurality" generally includes at least two.

[0049] As used herein, the words “if” or “when” may be interpreted as “at the time of” or “when” or “in response to determining” or “in response to detecting,” depending on the context. Similarly, the phrases “if it is determined” or “if (stated condition or event) is detected” may be interpreted as “when it is determined” or “in response to the determination” or “when detecting (stated condition or event)” or “in response to detecting (stated condition or event),” depending on the context.

[0050] In addition, the step sequence in the following method embodiments is only an example and not a strict limitation.

[0051] In practice, the server-side device deployed by the electronic component thermal distribution adaptive manifold and diversion system may be composed of one or more devices. The aforementioned electronic component thermal distribution adaptive manifold and diversion system can be implemented as a service instance, a virtual machine, or a hardware device. For example, the electronic component thermal distribution adaptive manifold and diversion system can be implemented as a service instance deployed on one or more devices in a cloud node. Simply put, the electronic component thermal distribution adaptive manifold and diversion system can be understood as software deployed on a cloud node, used to provide the electronic component thermal distribution adaptive manifold and diversion system to each client. Alternatively, the electronic component thermal distribution adaptive manifold and diversion system can be implemented as a virtual machine deployed on one or more devices in a cloud node. Application software for managing each client is installed in the virtual machine. Alternatively, the electronic component thermal distribution adaptive manifold and diversion system can be implemented as a server-side device composed of multiple hardware devices of the same or different types, with one or more hardware devices configured to provide the electronic component thermal distribution adaptive manifold and diversion system to each client.

[0052] In terms of implementation, the electronic component thermal distribution adaptive manifold and diversion system and the user end are mutually adapted. Specifically, the electronic component thermal distribution adaptive manifold and diversion system is an application installed on a cloud service platform, and the user end is a client that establishes a communication connection with the application. Alternatively, the electronic component thermal distribution adaptive manifold and diversion system is implemented as a website, and the user end is implemented as a webpage. Alternatively, the electronic component thermal distribution adaptive manifold and diversion system is implemented as a cloud service platform, and the user end is implemented as a mini-program within an instant messaging application.

[0053] like Figure 1 1 is a system architecture diagram of a manifold shunt system for adaptive heat distribution of electronic components provided by an embodiment of the present invention.

[0054] The electronic component thermal distribution adaptive manifold shunt system 100 of the present invention can be set in a cloud server. In terms of implementation, it can be used as one or more service devices, or as an application installed on the cloud (such as a mobile service operator's server, server cluster, etc.), or it can be developed as a website. According to the functions implemented, the electronic component thermal distribution adaptive manifold shunt system 100 can include an information extraction module 101, a product verification module 102, a verification failure module 103, a verification success module 104, a product settlement module 105, and a settlement success module 106. The module of the present invention can also be called a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0055] In an embodiment of the present invention, in the manifold shunt system with adaptive thermal distribution of electronic components, each of the above modules can be implemented independently and called with other modules. The call here can be understood as that a certain module can connect to multiple modules of another type and provide corresponding services to the multiple modules connected to it. In the manifold shunt system with adaptive thermal distribution of electronic components provided by an embodiment of the present invention, the scope of application of the manifold shunt system architecture with adaptive thermal distribution of electronic components can be adjusted by adding modules and directly calling them without modifying the program code, thereby realizing cluster-type horizontal expansion, so as to achieve the purpose of quickly and flexibly expanding the manifold shunt system with adaptive thermal distribution of electronic components. In actual applications, the above modules can be set in the same device or different devices, or they can be set in a virtual device, such as a service instance in a cloud server.

[0056] The following describes the various components and specific workflow of the manifold shunting system for adaptive heat distribution of electronic components in conjunction with specific embodiments.

[0057] The temperature acquisition module is used to obtain the real-time temperature of the electronic components in the integrated circuit chip;

[0058] In detail, an integrated circuit chip is a miniaturized electronic device or module with specific circuit functions that is formed by integrating a large number of electronic components such as transistors, resistors, capacitors, and wiring on a semiconductor substrate through semiconductor manufacturing processes; electronic components are the basic units that make up the circuit, including active components (such as transistors, integrated circuits) and passive components (such as resistors, capacitors, inductors), which are used to realize the circuit's conductivity, insulation, energy storage, amplification and other characteristics; the real-time temperature of the electronic components is obtained by integrating temperature sensors on the surface of the integrated circuit chip.

[0059] The cluster analysis module is configured to calculate the difference between the real-time temperature and the real-time temperature of adjacent components of the electronic component to obtain a temperature gradient of the electronic component, and perform cluster analysis on the electronic component based on the temperature gradient and the real-time temperature to obtain a plurality of thermal load characteristic clusters of the integrated circuit chip;

[0060] In detail, the "adjacent components" in "adjacent components of the electronic component" refer to other electronic components that have a direct physical relationship with the "said electronic component"; the temperature gradient refers to the real-time temperature difference between adjacent electronic components, reflecting the uneven distribution of heat in the chip. The larger the temperature gradient, the more significant the accumulation of heat in the local area.

[0061] In an embodiment of the present invention, when the cluster analysis module performs cluster analysis on the electronic components based on the temperature gradient and the real-time temperature to obtain a plurality of thermal load characteristic clusters of the integrated circuit chip, the cluster analysis module includes:

[0062] Performing feature fusion on the temperature gradient and the real-time temperature to obtain a two-dimensional feature vector of the electronic component, and aggregating the two-dimensional feature vectors into a vector data set of the integrated circuit chip;

[0063] In detail, the real-time temperature reflects the thermal load of the electronic component itself (such as whether it is overheated), and the temperature gradient reflects the intensity of thermal interaction between the electronic component and the surrounding area (such as whether there is heat concentration); feature fusion refers to the process of merging multiple independent features (temperature gradient and real-time temperature) into a composite feature. After feature fusion, each two-dimensional feature vector contains both "absolute thermal state" and "relative thermal state" information; a two-dimensional feature vector is a vector composed of two eigenvalues, which is used to characterize the thermal state of a single electronic component. The thermal state of each electronic component is converted into a point in space, which facilitates the mining of the overall thermal distribution law of the integrated circuit chip through clustering algorithms; the two-dimensional feature vectors of all electronic components are integrated into a vector data set. The vector data set is a collection of the two-dimensional feature vectors of all electronic components, forming a thermal state data set of the circuit chip.

[0064] Cluster analysis is performed on the vector data set based on a K-means clustering algorithm to obtain multiple thermal load feature clusters of the integrated circuit chip.

[0065] In detail, the vector data set is first preprocessed, and the feature dimension differences are eliminated by normalization. Then, the elbow rule is used to determine the optimal number of clusters K, and K cluster centers are randomly initialized. The Euclidean distance between each two-dimensional feature vector and each center is calculated and assigned to the nearest cluster. The cluster center is updated based on the mean of all two-dimensional feature vectors in the cluster. This process is repeated until the center position reaches the preset number of iterations. According to the converged clustering results, the statistical characteristics of each cluster, such as the real-time temperature mean and temperature gradient mean, are analyzed. Different clusters are defined as high-load heat island clusters, medium-load balanced clusters, and low-load cold zone clusters, providing data support for integrated circuit chip thermal design optimization and reliability evaluation.

[0066] In general: Using the K-means algorithm to perform cluster analysis on a vector data set that integrates the real-time temperature and temperature gradient of electronic components can quickly and efficiently divide electronic components into multiple heat load characteristic clusters. By identifying different heat load characteristic clusters, the cold air flow distribution can be adjusted in a targeted manner to achieve focused heat dissipation in high heat load areas, improve heat dissipation efficiency, and solve the problem that traditional manifold cooling systems cannot adapt to dynamic heat distribution changes.

[0067] The target heat exchange function construction module is configured to obtain the cross-sectional area and cold air flow rate of each branch pipe in the manifold, obtain the heat dissipation parameters of the electronic component, calculate the target heat load of the heat load characteristic cluster based on the heat dissipation parameters, construct the target heat exchange function of the integrated circuit chip based on the cross-sectional area, the cold air flow rate, the target heat load, and the position information of the heat load characteristic cluster, and generate the constraint conditions of the manifold according to the physical limitation parameters of the branch pipe;

[0068] Specifically, a manifold is a fluid distribution device consisting of a main pipe and multiple branch pipes (branches), used to distribute fluid from the main pipe to the various branches. Cross-sectional area refers to the area of ​​the internal cross section of a branch pipe, which is circular in cross section. Cooling air flow rate refers to the volume or mass of cooling air passing through the branch pipe per unit time. The branch pipe cross-sectional area and cooling air flow rate are directly read based on the branch pipe's geometric design drawings or specifications. The heat dissipation parameters of each electronic component are read based on the integrated circuit chip's design specifications. These heat dissipation parameters include power consumption data and component heat dissipation efficiency.

[0069] In an embodiment of the present invention, when the target heat exchange function construction module calculates the target heat load of the heat load characteristic cluster based on the heat dissipation parameters, the target heat exchange function construction module includes:

[0070] approximating the power consumption data of the heat dissipation parameter to the thermal energy data of the electronic component;

[0071] In detail, power consumption data refers to the electric power consumed by electronic components per unit time, and thermal energy data refers to the data related to the heat actually generated or accumulated by electronic components during operation. "Approximating the power consumption data of the heat dissipation parameters as the thermal energy data of the electronic components" means that it is assumed that the electric power consumed by electronic components is almost entirely converted into thermal energy, that is, it is assumed that the power consumption data and the thermal energy data are approximately equal in numerical value.

[0072] The total thermal energy of the thermal load characteristic cluster is calculated based on the thermal energy data. That is, the total thermal energy of the thermal load characteristic cluster is calculated by accumulating the thermal energy data of each electronic component in the thermal load characteristic cluster. The calculation formula of the total thermal energy in the cluster is as follows: Where, Indicates the first Thermal energy data of electronic components, Indicates the identification of the electronic component Indicates the The total cluster thermal energy of the heat load characteristic cluster, Indicates the identifier of the heat load feature cluster.

[0073] The intra-cluster heat dissipation efficiency of the heat load characteristic cluster is determined according to the heat dissipation efficiency of the component of the heat dissipation parameter, wherein the calculation formula of the intra-cluster heat dissipation efficiency is as follows: Where, Indicates the The heat dissipation efficiency within a cluster of heat load characteristics, Indicates the The heat dissipation efficiency of each electronic component, Indicates the Thermal energy data of electronic components, Indicates the total number of electronic components in the heat load characteristic cluster;

[0074] Specifically, component heat dissipation efficiency refers to the ability of a single electronic component to dissipate heat to the surrounding environment, and cluster heat dissipation efficiency refers to the heat dissipation efficiency of all components in a heat load characteristic cluster as a whole, reflecting the degree to which the heat in the cluster is effectively dissipated.

[0075] In detail, the calculation formula of the heat dissipation efficiency within the cluster adopts the weighted average method to calculate the heat dissipation efficiency within the cluster. , which is to multiply the power consumption of each electronic component by its heat dissipation efficiency and then add them up. This is equivalent to considering the contribution weight of the power consumption of each component to the overall heat dissipation efficiency. The electronic components with large power consumption account for a larger proportion in the calculation; the denominator part The sum of the power consumption of all electronic components within a cluster is calculated by dividing the numerator by the denominator to obtain the heat dissipation efficiency of the entire heat load characteristic cluster, which serves as a measure of the cluster's overall heat dissipation capacity. This calculation is based on the principle that components with higher power consumption have a greater impact on overall heat dissipation. Therefore, when calculating the heat dissipation efficiency within a cluster, the heat dissipation efficiency of each component should be weighted accordingly based on its power consumption.

[0076] The target heat load of the heat load characteristic cluster is calculated based on the heat dissipation efficiency within the cluster and the total heat energy of the cluster, wherein the calculation formula of the target heat load is as follows: Where, Indicates the The total cluster thermal energy of the heat load characteristic cluster, Indicates the The heat dissipation efficiency within the cluster of the heat load characteristic cluster, Indicates the The target heat load of each heat load characteristic cluster, Indicates the ID of the heat load signature cluster.

[0077] Specifically, the total cluster heat energy is the total amount of heat generated by the heat load characteristic cluster, while the heat dissipation efficiency within the cluster reflects the proportion of heat that can be effectively dissipated. Dividing the total heat by the heat dissipation efficiency gives the actual heat load that the heat load characteristic cluster needs to handle under the current heat dissipation capacity.

[0078] In an embodiment of the present invention, the target heat exchange function is as follows:

[0079] Where, represents the target heat exchange function, represents the parameters of the target heat exchange function, Indicates the identifier of the heat load characteristic cluster, represents the number of heat load characteristic clusters, Indicates the The target heat load of each heat load characteristic cluster, Indicates the branch pipe identifier. Indicates the number of branches, represents the heat transfer efficiency coefficient, represents the cold air flow of the i-th branch, Indicates the The cross-sectional area of ​​each branch pipe, Indicates the The pitch angle of each branch pipe, Indicates the The horizontal coordinate of the two-dimensional coordinates of the branch, Indicates the The vertical coordinate of the two-dimensional coordinates of the branch, Indicates the The horizontal coordinate in the position information of the target heat load is Indicates the The vertical coordinate in the location information of the target heat load.

[0080] In detail, for each heat load feature cluster, the inner layer sums The calculation is the sum of the heat exchange contributions of all branches to the heat load characteristic cluster; the numerator The heat transfer efficiency coefficient, the cold air flow rate of the branch pipe, the cross-sectional area, and the pitch angle are comprehensively considered to reflect the heat exchange capacity of each branch pipe. The cold air flow rate and cross-sectional area determine the branch pipe's ability to transport cold air for heat exchange in terms of flow rate and channel size. The effect of the branch pipe installation angle on the heat exchange effect is taken into account, and the heat transfer efficiency coefficient is used to correct the entire heat exchange capacity; the denominator This is the distance calculated based on the 2D coordinates of the branch pipe and the heat load signature cluster, reflecting the attenuation of heat exchange efficiency due to distance. This fractional calculation determines the actual heat exchange contribution of each branch pipe to a specific heat load signature cluster, and the total heat exchange contribution of all branches to that heat load signature cluster is then calculated by summing them.

[0081] In detail, Subtract the target heat load of each heat load characteristic cluster from the sum of the heat exchange contributions of all branches to the heat load characteristic cluster to obtain the difference between the heat load demand and the actual heat exchange capacity of each heat load characteristic cluster. This difference reflects the degree of heat exchange matching of the current cooling system at that heat load characteristic cluster. The smaller the difference, the closer the heat exchange matching and the better the cooling effect.

[0082] Specifically, the above differences of all heat load feature clusters are squared and then summed, i.e. The square operation is to avoid the cancellation of positive and negative differences and highlight the heat exchange matching differences of all heat load characteristic clusters; the summation operation comprehensively considers the heat exchange conditions of all heat load characteristic clusters in the entire heat dissipation system, and finally obtains the target heat exchange function. By minimizing the target heat exchange function, the heat exchange relationship between the heat load characteristic clusters and the branch pipes in the heat dissipation system can be optimized to achieve better heat dissipation effect.

[0083] In detail, based on the physical limitation parameters of the branch pipe, a series of restriction rules and boundary conditions are set for the overall operation and performance optimization of the manifold to ensure that the manifold operates within a reasonable and safe range.

[0084] In the embodiment of the present invention, the constraints are as follows: Where, Indicates the The horizontal coordinate of the two-dimensional coordinates of the branch, Indicates the The vertical coordinate of the two-dimensional coordinates of the branch, Indicates the branch pipe identifier. Indicates the The horizontal coordinate of the two-dimensional coordinates of the branch, Indicates the The vertical coordinate of the two-dimensional coordinates of the branch, Indicates the minimum spacing constraint between branches. Indicates the branch pipe identifier. Indicates the The pitch angle of each branch pipe, Indicates the lower limit of the pitch angle, Indicates the upper limit of the pitch angle.

[0085] In detail, the above constraints are imposed on the branch pipes in terms of their spatial position and pitch angle, with the aim of ensuring the rationality, safety and effectiveness of the manifold system in terms of physical layout and operation. The branch and the adjacent Branch pipes ( for or ) must be greater than or equal to the minimum spacing The square of . The pitch angle of each branch pipe must be within the specified lower limit and upper limit If the temperature exceeds this range, the constraint conditions will not be met, which will affect the heat exchange efficiency or cause problems such as poor airflow.

[0086] The target spatial layout determination module is configured to optimize the parameters of the target heat exchange function based on a gradient descent algorithm under the constraints to obtain a minimum value of the target heat exchange function, determine the optimal parameters of the target heat exchange function based on the minimum value, determine the target position of the branch pipe based on the two-dimensional coordinates in the optimal parameters, determine the target inclination angle of the branch pipe based on the pitch angle in the optimal parameters, and determine the target spatial layout of the manifold based on the target position and the target inclination angle;

[0087] In general, the target heat exchange function quantifies the relationship between cooling resource allocation and heat load. Constraints are combined to ensure the design is physically feasible, addressing the inability of traditional cooling systems to adapt to the dynamic thermal distribution of integrated circuit chips while also avoiding resource waste. This precise modeling allows for dynamic adjustment of coolant distribution based on actual heat load requirements, improving cooling efficiency, reducing local hotspot temperatures on the chip, enhancing chip performance and reliability, and extending its service life. This approach, while meeting the physical limitations of engineering implementation, provides a scientific and feasible design solution for adaptive manifold shunt systems for the thermal distribution of electronic components.

[0088] In an embodiment of the present invention, when the target space layout determination module optimizes the parameters of the target heat exchange function based on the gradient descent algorithm to obtain the minimum value of the target heat exchange function, it includes:

[0089] The parameters of the target heat exchange function are optimized using the update formula of the gradient descent algorithm, wherein the update formula is as follows:

[0090] Where, Indicates the The parameters at the iteration, Indicates the The parameters at the iteration, represents the learning rate, Indicates that the target heat exchange function is The gradient at The identifier of the iteration, represents the target heat exchange function;

[0091] When the change of the target heat exchange function is less than a preset change threshold, a minimum value of the target heat exchange function is obtained.

[0092] In detail, first initialize the parameters of the target heat exchange function (Branch coordinates, angles and other related parameters), set the initial value , and determine the learning rate (control parameter update step) and the preset change threshold, then enter the iterative process, in the first In the iteration, calculate the current parameters The gradient of the target heat transfer function at , according to the update formula Adjust the parameters and then check the target heat transfer function in and If the function value at the position changes and the change is less than the preset change threshold, the algorithm is considered to have converged. The optimized parameters correspond to the minimum value of the target heat exchange function. If the conditions are not met, the next iteration is continued. By repeatedly repeating the process of "calculating gradients - updating parameters - judging convergence", the parameters are gradually adjusted in the negative direction of the gradient, eventually approaching the minimum point of the target heat exchange function, thus achieving parameter optimization.

[0093] Specifically, the gradient descent algorithm used to solve the target heat exchange function yields a set of parameters that minimize the target heat exchange function. This set of parameters is known as the optimal parameters. These parameters contain various attributes related to the branch pipe. Parameter values ​​representing the two-dimensional coordinates of the branch pipe are extracted from the optimal parameters; these values ​​determine the target position of the branch pipe in the planar layout. The optimal parameters also include a parameter representing the branch pipe's pitch angle, which describes its tilt in space. Based on this optimal pitch angle parameter, the target tilt angle for the branch pipe during installation can be determined.

[0094] In an embodiment of the present invention, when determining the target spatial layout of the manifold according to the target position and the target inclination angle, the target spatial layout determining module includes:

[0095] Installing the branch pipes according to the target positions to obtain a preliminary spatial layout of the branch pipes;

[0096] Under the preliminary spatial layout, the initial inclination angles of the branch pipes are adjusted according to the target inclination angles to obtain the target spatial layout of the manifolds.

[0097] Specifically, the optimal planar positions of the branches, obtained through the previous optimization process, are determined by two-dimensional coordinates (abscissa and ordinate). Based on this target position information, the branches are installed in the corresponding positions, thus obtaining a preliminary spatial layout of the branches. For example, in a cooling manifold system, the branches are installed at pre-calculated planar positions to ensure they are optimally distributed horizontally. Based on this preliminary spatial layout of the branches, the initial tilt angles of the branches are adjusted according to the target tilt angles, ensuring that the branches are both optimally positioned and tilted in space, ultimately achieving the target spatial layout of the manifold. For example, in a cooling manifold system, adjusting the tilt angles of the branches can ensure more efficient heat exchange between cool air and the heat load characteristic clusters, improving heat dissipation efficiency.

[0098] In general, the gradient descent algorithm can quickly find the optimal parameters that minimize the target heat transfer function while satisfying the physical constraints of the manifolds. These optimal parameters correspond to the optimal state of chip thermal management, namely, maximum heat dissipation efficiency and the most uniform temperature distribution. By mapping the optimization results to the target positions and tilt angles of the manifolds, the spatial layout of the manifolds can be precisely designed, ensuring that the cool air flow direction closely matches the actual heat load distribution of the integrated circuit chips. This overcomes the limitations of traditional fixed-layout cooling systems and adaptively adjusts cooling resource allocation based on the dynamic thermal characteristics of the chip, effectively reducing local hotspot temperatures and improving overall chip performance and reliability.

[0099] The cold air demand determination module is configured to calculate the branch cold air demand of the branch pipe based on the target heat load;

[0100] The heat dissipation execution module is configured to dissipate heat for the electronic components corresponding to the heat load characteristic cluster based on the target space layout and the branch cold air demand.

[0101] In an embodiment of the present invention, the calculation formula for the branch cold air demand is as follows: Where, Indicates the The branch cooling air demand of each branch pipe, Indicates the branch pipe identifier. Indicates the The heat transfer efficiency coefficient of each branch pipe is Indicates the The target heat load of each heat load characteristic cluster, Indicates the identifier of the heat load characteristic cluster, Indicates the The pitch angle of each branch pipe, Indicates the The horizontal coordinate of the two-dimensional coordinates of the branch, Indicates the The vertical coordinate of the two-dimensional coordinates of the branch, Indicates the The horizontal coordinate in the position information of the target heat load is Indicates the The vertical coordinate in the location information of the target heat load.

[0102] In detail, It is the distance between the branch pipe and the heat load characteristic cluster calculated based on their two-dimensional coordinates. The longer the distance, the lower the heat exchange efficiency will be, and the required cooling air flow may increase accordingly. In the formula, the distance is in the denominator, which means that the greater the distance, the greater the cooling air demand of the branch pipe. The larger the heat load The bigger it is, the more cold air is needed to dissipate the heat. It reflects the effect of angle on heat exchange efficiency. The angle is appropriate ( When the value is large, the heat exchange efficiency is high and the amount of cooling air required can be relatively reduced. The product of the two reflects the combined impact of heat load and angle on the cooling air demand. As the heat transfer efficiency coefficient, the entire calculation results are corrected.

[0103] Specifically, the branch pipe heat transfer efficiency coefficient is determined based on expert rules. Experts first determine the operating parameters of the branch pipe, including density, viscosity, velocity, flow rate, inlet and outlet temperatures, pressure, and the heat load characteristic cluster temperature. They also analyze the branch pipe's geometry, including diameter, length, and shape; the material properties of aluminum and copper (including thermal conductivity, temperature resistance, and corrosion resistance); and internal surface roughness and contamination. Drawing on their deep theoretical knowledge and extensive project experience in the field of heat transfer, they comprehensively assess the impact of operating conditions and branch pipe characteristics on heat transfer, ultimately determining a reasonable branch pipe heat transfer efficiency coefficient.

[0104] In general, by using the target heat load of the heat load signature cluster as the basis for calculation, the required cold air flow rate for each branch can be derived, dynamically matching the cold air distribution with the actual heat distribution of the integrated circuit chip. This method solves the fixed flow distribution problem of traditional manifold distribution systems, avoiding performance degradation due to insufficient cooling in high heat load areas and resource waste caused by excessive cooling in low heat load areas.

[0105] In an embodiment of the present invention, when the heat dissipation execution module performs heat dissipation for the electronic components corresponding to the heat load characteristic cluster based on the target space layout and the branch cold air demand, the heat dissipation execution module includes:

[0106] The total cold air demand of the manifold is calculated based on the branch cold air demand, that is, the total cold air demand of the entire manifold is calculated by summing the branch cold air demand of each branch obtained previously.

[0107] Cold air is delivered to the manifold according to the total cold air demand: a corresponding amount of cold air is delivered to the manifold according to the calculated total cold air demand, to ensure that the manifold has enough cold air to perform subsequent heat dissipation work.

[0108] Calculate the branch cold air injection time of the manifold according to the branch cold air demand and the cold air flow rate: Assuming that the branch cold air demand of a branch is , the cold air flow is , then the branch cold air injection time of the branch pipe is By calculating the injection time, the duration for each branch to release cold air can be reasonably controlled to accurately meet the heat dissipation needs of the heat load characteristic cluster.

[0109] Under the target spatial layout (i.e., the branch pipes are arranged according to the target positions and target inclination angles), the branch pipes dissipate heat for the electronic components corresponding to the heat load characteristic cluster according to the branch cold air injection time, thereby ensuring that the cold air is injected at the appropriate position and for an appropriate duration, thereby efficiently dissipating heat for the electronic components and maintaining their normal operating temperature.

[0110] In summary: By combining the optimized manifold spatial layout with the precise cooling air requirements of each branch, cooling can be customized based on the thermal load characteristics of different areas within the integrated circuit chip. This precise matching effectively solves the local hotspot problem caused by the fixed layout and extensive flow distribution of traditional cooling systems, allowing high-heat-load areas to obtain more cooling resources and avoiding overcooling in low-heat-load areas, thereby significantly improving the overall temperature uniformity of the chip.

[0111] Reference Figure 2 FIG. 1 is a flow chart of a method for self-adapting the thermal distribution of electronic components provided by an embodiment of the present invention. In this embodiment, the method for self-adapting the thermal distribution of electronic components comprises:

[0112] S1. Obtaining the real-time temperature of electronic components in the integrated circuit chip;

[0113] S2. Calculating the difference between the real-time temperature and the real-time temperatures of adjacent components of the electronic component to obtain a temperature gradient of the electronic component, and performing cluster analysis on the electronic components based on the temperature gradient and the real-time temperature to obtain multiple thermal load characteristic clusters of the integrated circuit chip;

[0114] S3. Obtaining the cross-sectional area and cold air flow rate of each branch pipe in the manifold, obtaining the heat dissipation parameters of the electronic component, calculating the target heat load of the heat load characteristic cluster based on the heat dissipation parameters, constructing the target heat exchange function of the integrated circuit chip based on the cross-sectional area, the cold air flow rate, the target heat load, and the position information of the heat load characteristic cluster, and generating the constraint conditions of the manifold according to the physical limitation parameters of the branch pipe;

[0115] S4. Under the constraints, optimize the parameters of the target heat exchange function based on a gradient descent algorithm to obtain a minimum value of the target heat exchange function, determine the optimal parameters of the target heat exchange function based on the minimum value, determine the target position of the branch pipe based on the two-dimensional coordinates of the optimal parameters, determine the target inclination angle of the branch pipe based on the pitch angle of the optimal parameters, and determine the target spatial layout of the manifold based on the target position and the target inclination angle;

[0116] S5. Calculating the branch cold air demand of the branch pipe based on the target heat load;

[0117] S6. Cooling the electronic components corresponding to the heat load characteristic cluster based on the target space layout and the branch cold air demand.

[0118] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0119] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence refers to the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to achieve optimal results.

[0120] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A manifold shunt system with adaptive heat distribution for electronic components, characterized in that: The system includes a temperature acquisition module, a cluster analysis module, a target heat exchange function construction module, a target space layout determination module, a cold air demand determination module and a heat dissipation execution module, wherein: The temperature acquisition module is used to obtain the real-time temperature of the electronic components in the integrated circuit chip; The cluster analysis module is configured to calculate the difference between the real-time temperature and the real-time temperature of adjacent components of the electronic component to obtain a temperature gradient of the electronic component, and perform cluster analysis on the electronic component based on the temperature gradient and the real-time temperature to obtain a plurality of thermal load characteristic clusters of the integrated circuit chip; The target heat exchange function construction module is configured to obtain the cross-sectional area and cold air flow rate of each branch pipe in the manifold, obtain the heat dissipation parameters of the electronic component, calculate the target heat load of the heat load characteristic cluster based on the heat dissipation parameters, construct the target heat exchange function of the integrated circuit chip based on the cross-sectional area, the cold air flow rate, the target heat load, and the position information of the heat load characteristic cluster, and generate the constraint conditions of the manifold according to the physical limitation parameters of the branch pipe; The target spatial layout determination module is configured to optimize the parameters of the target heat exchange function based on a gradient descent algorithm under the constraints to obtain a minimum value of the target heat exchange function, determine the optimal parameters of the target heat exchange function based on the minimum value, determine the target position of the branch pipe based on the two-dimensional coordinates in the optimal parameters, determine the target inclination angle of the branch pipe based on the pitch angle in the optimal parameters, and determine the target spatial layout of the manifold based on the target position and the target inclination angle; The cold air demand determination module is configured to calculate the branch cold air demand of the branch pipe based on the target heat load; The heat dissipation execution module is configured to dissipate heat for the electronic components corresponding to the heat load characteristic cluster based on the target space layout and the branch cold air demand.

2. The electronic component heat distribution adaptive manifold shunt system according to claim 1, characterized in that: When the cluster analysis module performs cluster analysis on the electronic components based on the temperature gradient and the real-time temperature to obtain a plurality of thermal load characteristic clusters of the integrated circuit chip, the cluster analysis module includes: Performing feature fusion on the temperature gradient and the real-time temperature to obtain a two-dimensional feature vector of the electronic component, and aggregating the two-dimensional feature vectors into a vector data set of the integrated circuit chip; Cluster analysis is performed on the vector data set based on a K-means clustering algorithm to obtain multiple thermal load feature clusters of the integrated circuit chip.

3. The electronic component heat distribution adaptive manifold shunt system according to claim 1, characterized in that: When the target heat exchange function construction module calculates the target heat load of the heat load characteristic cluster based on the heat dissipation parameters, the target heat exchange function construction module includes: approximating the power consumption data of the heat dissipation parameter to the thermal energy data of the electronic component; Calculating a cluster total thermal energy of the heat load characteristic cluster based on the thermal energy data; Determining the intra-cluster heat dissipation efficiency of the heat load characteristic cluster according to the heat dissipation efficiency of the component of the heat dissipation parameter; The target heat load of the heat load characteristic cluster is calculated based on the heat dissipation efficiency within the cluster and the total heat energy of the cluster, wherein the calculation formula of the target heat load is as follows: Where, Indicates the The total cluster thermal energy of the heat load characteristic cluster, Indicates the The heat dissipation efficiency within the cluster of the heat load characteristic cluster, Indicates the The target heat load of each heat load characteristic cluster, Indicates the ID of the heat load signature cluster.

4. The electronic component heat distribution adaptive manifold shunt system according to claim 1, characterized in that: The target heat exchange function is as follows: Where, represents the target heat exchange function, represents the parameters of the target heat exchange function, Indicates the identifier of the heat load characteristic cluster, represents the number of heat load characteristic clusters, Indicates the The target heat load of each heat load characteristic cluster, Indicates the branch pipe identifier. Indicates the number of branches, represents the heat transfer efficiency coefficient, represents the cold air flow of the i-th branch, Indicates the The cross-sectional area of ​​each branch pipe, Indicates the The pitch angle of each branch pipe, Indicates the The horizontal coordinate of the two-dimensional coordinates of the branch, Indicates the The vertical coordinate of the two-dimensional coordinates of the branch, Indicates the The horizontal coordinate in the position information of the target heat load is Indicates the The vertical coordinate in the location information of the target heat load.

5. The electronic component heat distribution adaptive manifold shunt system according to claim 1, characterized in that: The constraints are as follows: Where, Indicates the The horizontal coordinate of the two-dimensional coordinates of the branch, Indicates the The vertical coordinate of the two-dimensional coordinates of the branch, Indicates the branch pipe identifier. Indicates the The horizontal coordinate of the two-dimensional coordinates of the branch, Indicates the The vertical coordinate of the two-dimensional coordinates of the branch, Indicates the minimum spacing constraint between branches. Indicates the branch pipe identifier. Indicates the The pitch angle of each branch pipe, Indicates the lower limit of the pitch angle, Indicates the upper limit of the pitch angle.

6. The electronic component heat distribution adaptive manifold shunt system according to claim 1, characterized in that: When the target space layout determination module optimizes the parameters of the target heat exchange function based on the gradient descent algorithm to obtain the minimum value of the target heat exchange function, it includes: The parameters of the target heat exchange function are optimized using the update formula of the gradient descent algorithm, wherein the update formula is as follows: Where, Indicates the The parameters at the iteration, Indicates the The parameters at the iteration, represents the learning rate, Indicates that the target heat exchange function is The gradient at The identifier of the iteration, represents the target heat exchange function; When the change of the target heat exchange function is less than a preset change threshold, a minimum value of the target heat exchange function is obtained.

7. The electronic component heat distribution adaptive manifold shunt system according to claim 1, characterized in that: When the target spatial layout determination module determines the target spatial layout of the manifold according to the target position and the target inclination angle, the module includes: Installing the branch pipes according to the target positions to obtain a preliminary spatial layout of the branch pipes; Under the preliminary spatial layout, the initial inclination angles of the branch pipes are adjusted according to the target inclination angles to obtain the target spatial layout of the manifolds.

8. The electronic component heat distribution adaptive manifold shunt system according to claim 1, characterized in that: The calculation formula for the branch cold air demand is as follows: Where, Indicates the The branch cooling air demand of each branch pipe, Indicates the branch pipe identifier. Indicates the The heat transfer efficiency coefficient of each branch pipe is Indicates the The target heat load of each heat load characteristic cluster, Indicates the identifier of the heat load characteristic cluster, Indicates the The pitch angle of each branch pipe, Indicates the The horizontal coordinate of the two-dimensional coordinates of the branch, Indicates the The vertical coordinate of the two-dimensional coordinates of the branch, Indicates the The horizontal coordinate in the position information of the target heat load is Indicates the The vertical coordinate in the location information of the target heat load.

9. The electronic component heat distribution adaptive manifold shunt system according to claim 1, characterized in that: When the heat dissipation execution module performs heat dissipation on the electronic components corresponding to the heat load characteristic cluster based on the target space layout and the branch cold air demand, the heat dissipation execution module includes: calculating a total cold air demand of the manifold based on the branch cold air demands; delivering cold air to the manifold according to the total cold air demand; Calculating the branch cold air injection time of the manifold according to the branch cold air demand and the cold air flow rate; Under the target spatial layout, the branch pipes dissipate heat for the electronic components corresponding to the heat load characteristic cluster according to the branch cold air injection time.

10. A manifold liquid distribution method with adaptive thermal distribution of electronic components, characterized in that: The method comprises: S1. Obtaining the real-time temperature of electronic components in the integrated circuit chip; S2. Calculating the difference between the real-time temperature and the real-time temperatures of adjacent components of the electronic component to obtain a temperature gradient of the electronic component, and performing cluster analysis on the electronic components based on the temperature gradient and the real-time temperature to obtain multiple thermal load characteristic clusters of the integrated circuit chip; S3. Obtaining the cross-sectional area and cold air flow rate of each branch pipe in the manifold, obtaining the heat dissipation parameters of the electronic component, calculating the target heat load of the heat load characteristic cluster based on the heat dissipation parameters, constructing the target heat exchange function of the integrated circuit chip based on the cross-sectional area, the cold air flow rate, the target heat load, and the position information of the heat load characteristic cluster, and generating the constraint conditions of the manifold according to the physical limitation parameters of the branch pipe; S4. Under the constraints, optimize the parameters of the target heat exchange function based on a gradient descent algorithm to obtain a minimum value of the target heat exchange function, determine the optimal parameters of the target heat exchange function based on the minimum value, determine the target position of the branch pipe based on the two-dimensional coordinates of the optimal parameters, determine the target inclination angle of the branch pipe based on the pitch angle of the optimal parameters, and determine the target spatial layout of the manifold based on the target position and the target inclination angle; S5. Calculating the branch cold air demand of the branch pipe based on the target heat load; S6. Cooling the electronic components corresponding to the heat load characteristic cluster based on the target space layout and the branch cold air demand.

Images