Dynamic gradient porosity TPMS heat dissipation structure and control method

Abstract

The application discloses a dynamic gradient porosity TPMS heat dissipation structure and a control method, and relates to the technical field of liquid cooling heat dissipation control of data centers. The dynamic gradient porosity TPMS heat dissipation structure is composed of a cold plate body, a Gyroid type TPMS flow channel structure, a plurality of temperature sensors and a plurality of flow regulating elements. The cold plate body is a heat exchange base body for bearing the flow channel structure and being in contact with a chip to be cooled. The Gyroid type TPMS flow channel structure is arranged in the cold plate body and is divided into at least two independent functional sections along the flow direction of the cooling liquid. The plurality of temperature sensors are arranged in an array on the contact surface of the cold plate body and the chip to be cooled, correspond to different heating areas of the chip to be cooled respectively, and are used for collecting temperature signals of the heating areas in real time. The application upgrades the traditional static gradient design to a dynamic adaptive gradient design, can respond to the change of the chip thermal load in real time, and avoids the performance compromise of the fixed gradient flow channel under high and low load working conditions.

Description

Technical Field

[0001] This invention relates to the field of liquid cooling heat dissipation control technology for data centers, specifically to a dynamic gradient porosity TPMS heat dissipation structure and control method. Background Technology

[0002] In actual operation, the heat load of data center server CPUs exhibits dynamic fluctuations, typically ranging from 100W to 300W. Existing liquid cooling plates mostly employ uniform porosity channels or fixed gradient porosity channels, which cannot adaptively adjust to changes in actual heat load. Under high load conditions, fixed gradient channels may generate localized hot spots; under low load conditions, they can create unnecessary flow resistance, increasing pump power consumption.

[0003] Unlike existing static gradient designs, this invention proposes a dynamic gradient porosity (TPMS) heat dissipation structure and control method, which enables the porosity distribution of the cold plate to be optimized in real time according to the chip's operating conditions, fundamentally solving the limitations of fixed gradient designs. Summary of the Invention

[0004] Technical problems to be solved To address the shortcomings of existing technologies, this invention provides a dynamic gradient porosity TPMS heat dissipation structure and control method, which solves the problems mentioned in the background art.

[0005] Technical solution To achieve the above objectives, the present invention is implemented through the following technical solution: a dynamic gradient porosity TPMS heat dissipation cold plate structure, which consists of a cold plate body, a Gyroid-type TPMS flow channel structure, several temperature sensors and several flow regulation elements.

[0006] The core objective of this overall structural design is to address the issues of existing static gradient TPMS cold plates being unable to adapt to dynamic thermal load changes in chips, performance trade-offs under high and low load conditions, and high system pump power consumption. The goal is to construct a data center liquid cooling cold plate system that can match thermal load distribution in real time, has high heat dissipation efficiency, low energy consumption, and strong engineering applicability, and is suitable for the heat dissipation needs of high power density central processing units, graphics processing units, artificial intelligence chips, and power electronic devices.

[0007] The cold plate body is a heat exchange substrate that supports the internal flow channel structure and is in direct contact with the chip to be cooled. It undertakes the core heat exchange function of transferring heat from the chip surface to the coolant, and at the same time provides mounting support for temperature sensors and flow regulation components.

[0008] Preferably, the cold plate body is made of aluminum alloy or copper alloy. Aluminum alloy has low density, light weight, good processing performance, and low cost, making it suitable for the needs of large-scale data center mass applications. Copper alloy has high thermal conductivity and excellent heat exchange performance, making it suitable for the heat dissipation needs of ultra-high power density chips of more than 200 watts per square centimeter.

[0009] It also includes a controller, which is electrically connected to several temperature sensors and several flow regulating elements, and is used to receive temperature signals and output control signals. The controller is the control core of the entire dynamic regulation system. It is responsible for receiving temperature data collected by the temperature sensors, running control algorithms to calculate the target flow values ​​for each functional segment, and outputting corresponding control signals to the flow regulating elements to achieve coordinated operation of the entire system.

[0010] The Gyroid-type TPMS flow channel structure serves as the flow channel for coolant. Its 3D continuous interconnected topology features a large specific surface area, strong fluid turbulence, high heat exchange efficiency, and excellent mechanical properties, making it a core component for achieving efficient liquid cooling.

[0011] The Gyroid-type TPMS flow channel structure is located inside the cold plate body and is divided into at least two independent functional sections along the coolant flow direction. Dividing the flow channel into independent functional sections is the basis for achieving dynamic porosity adjustment. By independently controlling the flow rate of each functional section, the heat exchange capacity of different areas can be precisely matched with the heat load of the corresponding areas of the chip, avoiding the local overheating or overcooling problems that exist in traditional integral flow channels.

[0012] Preferably, the wall thickness of the Gyroid-type TPMS flow channel structure remains constant, ranging from 0.2mm to 0.5mm, and the total length of the flow channel is equal to the effective heat exchange area length of the cold plate, ranging from 80mm to 120mm. A constant wall thickness ensures uniform mechanical strength of the flow channel structure, avoiding deformation or cracking caused by localized stress concentration. A wall thickness of 0.2mm to 0.5mm ensures the flow channel structure's pressure resistance meets the working pressure requirements of liquid cooling systems above one MPa while minimizing solid thermal resistance and improving heat exchange efficiency. A wall thickness less than 0.2mm leads to insufficient structural strength, making it prone to collapse and deformation defects during 3D printing. A wall thickness greater than 0.5mm significantly increases solid thermal resistance and reduces heat dissipation performance. The total flow channel length of 80mm to 120mm perfectly matches the dimensions of current mainstream data center server chips, such as the effective heat exchange area length of AI chips like Intel Xeon, AMD EPYC, and NVIDIA H100, ensuring full coverage heat exchange on the chip surface. A flow path that is too short will result in insufficient residence time of the coolant within the cold plate, leading to inadequate heat exchange. A flow path that is too long will significantly increase the overall system pressure drop, increasing pump power consumption.

[0013] Several temperature sensors are used to collect temperature signals from different heat-generating areas on the chip surface in real time, providing feedback for dynamic flow regulation and serving as the core sensing element for achieving closed-loop control.

[0014] The array of temperature sensors is arranged on the contact surface between the cold plate body and the chip to be cooled, each corresponding to a different heat-generating area of ​​the chip to be cooled, and is used to collect the temperature signal of each heat-generating area in real time.

[0015] Preferably, the plurality of temperature sensors are arranged in a matrix uniformly on the bottom surface of the cold plate body, with the spacing between adjacent temperature sensors not exceeding 20 mm, and each functional segment corresponding to the acquisition area of ​​at least one temperature sensor. The matrix uniform arrangement enables high-precision gridded acquisition of the chip surface temperature distribution, accurately identifying hot spots and heat load distribution on the chip. The spacing between adjacent sensors not exceeding 20 mm ensures spatial resolution of temperature acquisition, avoiding missed detection of local hot spots. Each functional segment corresponding to at least one temperature sensor ensures that the flow regulation of each functional segment has an independent temperature feedback signal, achieving precise zoned control.

[0016] Several flow regulation elements are used to independently regulate the coolant flow rate in each functional segment. By changing the flow distribution of each functional segment, the porosity of each functional segment can be dynamically and independently adjusted, which is the core of the execution of dynamic gradient porosity.

[0017] Each functional segment is equipped with an independent flow regulating element at its inlet or outlet. This flow regulating element can be a piezoelectric microvalve or a miniature solenoid valve. Piezoelectric microvalve features fast response, low power consumption, small size, and no electromagnetic interference, making it suitable for rapid response to millisecond-level heat load fluctuations. Miniature solenoid valves are low-cost, highly reliable, and easy to control, making them suitable for general heat dissipation scenarios with relatively low response speed requirements.

[0018] Preferably, the flow regulating element is a piezoelectric microvalve with a flow regulation range of 0 L / min to 5 L / min. This range covers the flow requirements of a single cooling plate in current mainstream data center liquid cooling systems, typically ranging from one to three liters per minute, while also providing sufficient adjustment margin to handle extreme high-load conditions. The lower limit of 0 L / min allows for complete flow interruption in low-load areas, minimizing the total system flow and pump power consumption. The upper limit of 5 L / min ensures adequate heat dissipation for hotspot areas of ultra-high power density chips.

[0019] The core principle of this invention, which achieves dynamic porosity adjustment through flow distribution regulation, is that for a given TPMS flow channel structure, its equivalent porosity is positively correlated with the coolant flow rate through that channel per unit time. Under the same inlet pressure conditions, increasing the flow rate of a certain functional segment is equivalent to increasing the equivalent porosity of that functional segment, thereby enhancing the heat transfer capacity of that region. Conversely, decreasing the flow rate of a certain functional segment is equivalent to decreasing the equivalent porosity of that functional segment, thereby reducing the heat transfer capacity of that region. By independently adjusting the flow rate of each functional segment, the porosity distribution inside the cold plate can be matched in real time with the heat load distribution on the chip surface. Regions with high heat load correspond to high porosity and high flow rate, while regions with low heat load correspond to low porosity and low flow rate. This minimizes the total system flow rate and total pressure drop, and reduces pump power consumption, while ensuring the overall temperature uniformity of the chip and preventing hot spot temperatures from exceeding limits.

[0020] The present invention also provides a control method for a dynamic gradient porosity TPMS heat dissipation cold plate structure, for controlling the above-mentioned dynamic gradient porosity TPMS heat dissipation cold plate structure, comprising the following steps: Step 1: Using several temperature sensors arranged on the contact surface between the cold plate body and the chip to be cooled, the temperature distribution data of each heat-generating area on the surface of the chip to be cooled is collected in real time.

[0021] The core purpose of this step is to obtain real-time heat load distribution information on the chip surface, providing accurate feedback for subsequent flow regulation. The accuracy and real-time performance of temperature data acquisition directly determine the control accuracy and response speed of the entire dynamic regulation system.

[0022] Step 2: Based on the preset temperature and porosity mapping model and combined with the real-time collected temperature distribution data, calculate the target porosity value required for each functional segment, and determine the target flow rate value corresponding to each functional segment accordingly.

[0023] The core purpose of this step is to convert the collected temperature distribution data into executable flow control commands, which is the core algorithm of the entire control method.

[0024] Preferably, when determining the target flow rate for each functional segment, the optimization objective is to minimize the total voltage drop of the cold plate system, provided that the hotspot temperature of the chip to be cooled does not exceed a preset safety threshold. The safe operating temperature of the chip is the primary constraint; the hotspot temperature must always be below the preset safety threshold, typically between 85 and 95 degrees Celsius, depending on the chip model. Minimizing the total system voltage drop, while meeting safety constraints, directly reduces the pump power consumption of the liquid cooling system, improving the energy efficiency of the data center.

[0025] Preferably, the method for establishing the temperature-porosity mapping model involves pre-establishing a database of correspondences between the temperature distribution of the chip to be cooled and the optimal porosity gradient distribution of the cold plate under different heat load distribution conditions through computational fluid dynamics (CFD) simulation. During actual operation, based on real-time collected temperature distribution data, matching searches or interpolation calculations are performed in the database to obtain the optimal porosity distribution parameters under the current operating condition. Computational fluid dynamics simulation can comprehensively simulate various possible heat load distribution conditions during the product design phase, quickly establishing a mapping database covering all operating conditions without the need for extensive physical testing, significantly reducing R&D costs and time. The database matching and interpolation calculation method features high calculation speed and good real-time performance, meeting the requirements of millisecond-level control cycles.

[0026] Step 3: Send control signals to the flow regulation elements corresponding to each functional segment to adjust the coolant flow rate of each functional segment to the target flow rate value, thereby achieving a dynamic gradient distribution of porosity inside the cold plate.

[0027] The core purpose of this step is to execute the flow regulation command output by the control algorithm, convert the calculated target flow value into the actual coolant flow distribution, and ultimately achieve the dynamic gradient distribution of porosity.

[0028] Preferably, in step three, the flow regulation response time of the flow regulation element is no greater than 100ms. A response time of no more than 100ms ensures that the system can respond quickly to millisecond-level thermal load fluctuations in the chip, such as sudden computing tasks of artificial intelligence chips, avoiding chip performance degradation or damage caused by instantaneous overheating of hotspots. An excessively long response time will lead to control lag and will not effectively suppress temperature fluctuations caused by dynamic thermal loads.

[0029] Preferably, when the hotspot temperature of the chip to be cooled exceeds a preset safety threshold, the coolant flow rate of the corresponding functional segment of the hotspot is increased first, until the hotspot temperature drops below the safety threshold. This is a priority control strategy; when extreme high-load conditions cause the hotspot temperature to exceed the limit, the system will temporarily abandon the optimization objective of minimizing the total voltage drop and prioritize ensuring the safe operation of the chip. By concentrating the flow rate to the hotspot area, the hotspot temperature can be quickly reduced, preventing the chip from triggering protection mechanisms due to overheating or causing permanent damage.

[0030] This invention constructs a liquid-cooled cold plate system with dynamically adjustable porosity distribution by dividing the Gyroid-type TPMS flow channel into independent functional sections and combining a temperature sensor array with an independent flow regulation element.

[0031] The Gyroid-type TPMS flow channel itself possesses excellent heat transfer and mechanical properties, laying the foundation for efficient heat dissipation. Independent functional segment division enables zoned control. A temperature sensor array provides high-precision real-time temperature feedback. Flow regulation elements achieve millisecond-level precise flow regulation. A temperature-porosity mapping model ensures the accuracy and real-time performance of the control algorithm.

[0032] The various components work together to fundamentally solve the technical problem that existing static gradient TPMS cold plates cannot adapt to dynamic heat load changes. This allows the porosity distribution of the cold plate to be optimized in real time according to the chip's heat load, avoiding the performance trade-off of insufficient heat dissipation under high load conditions and excessive energy consumption under low load conditions inherent in fixed gradient designs. Simultaneously, by minimizing the total voltage drop, the pump power consumption of the liquid cooling system can be significantly reduced while ensuring heat dissipation requirements are met, thereby improving the energy efficiency of the data center.

[0033] The cold plate structure of this invention is compatible with the standard interface of existing data center liquid cooling pipelines, without the need for large-scale modification of existing liquid cooling systems. It has good engineering applicability and promotion value, and can be widely used for heat dissipation of high power density electronic devices in data centers, high-performance computing, artificial intelligence, new energy vehicles and other fields.

[0034] Beneficial effects The present invention has the following beneficial effects: (1) The dynamic gradient porosity TPMS heat dissipation structure and control method upgrades the traditional static gradient design to a dynamic adaptive gradient design, which can respond to the changes in chip thermal load in real time and avoid the performance trade-off of fixed gradient channels under high and low load conditions.

[0035] (2) The dynamic gradient porosity TPMS heat dissipation structure and control method can minimize the system pressure drop and reduce the pump power consumption of the data center liquid cooling system by means of closed-loop feedback control, while ensuring heat dissipation requirements.

[0036] (3) The dynamic gradient porosity TPMS heat dissipation structure and control method can be compatible with existing data center liquid cooling pipelines through the set cold plate structure, and has good engineering applicability.

[0037] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0038] Figure 1 This is a flowchart of the dynamic gradient porosity TPMS heat dissipation structure control method of the present invention. Detailed Implementation

[0039] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0040] Example 1

[0041] This embodiment provides a dynamic gradient porosity TPMS heat dissipation cold plate structure and its control method. The dynamic gradient porosity TPMS heat dissipation cold plate structure consists of a cold plate body, a Gyroid-type TPMS flow channel structure, several temperature sensors, several flow regulation elements, and a controller.

[0042] The cold plate body is made of aluminum alloy and has dimensions of 80mm × 80mm × 8mm. The Gyroid-type TPMS flow channel structure has a wall thickness of 0.2mm and a total length of 80mm, divided into three independent functional sections along the coolant flow direction, each section being 26.7mm long. Several temperature sensors are evenly arranged in a 3×3 matrix on the bottom surface of the cold plate body, with an adjacent sensor spacing of 15mm. Each functional section corresponds to the acquisition area of ​​three temperature sensors. The temperature sensors have a measurement accuracy of 0.5 degrees Celsius and a sampling frequency of 100Hz. Each functional section has an independently installed piezoelectric microvalve at its inlet, with a flow rate adjustment range of 0L / min to 5L / min. The controller uses an ARM microcontroller, electrically connected to all temperature sensors and piezoelectric microvalves.

[0043] Methods for controlling the dynamic gradient porosity TPMS heat dissipation plate structure, such as Figure 1 As shown, it includes the following steps: Step 1: Using nine temperature sensors arranged on the bottom surface of the cold plate, the temperature distribution data of each heat-generating area on the surface of the central processing unit chip to be cooled is collected in real time.

[0044] Step 2: Based on the preset temperature and porosity mapping model and combined with the real-time collected temperature distribution data, calculate the target porosity value required for each functional segment. Under the premise that the hot spot temperature of the chip to be cooled does not exceed 85 degrees Celsius, and with minimizing the total pressure drop of the cold plate system as the optimization objective, determine the target flow rate value corresponding to each functional segment.

[0045] Step 3: Send control signals to the piezoelectric microvalves corresponding to each functional segment to adjust the coolant flow rate of each functional segment to the target flow rate value, thereby achieving a dynamic gradient distribution of porosity inside the cold plate.

[0046] The method for establishing the temperature-porosity mapping model is as follows: A database of correspondences between the temperature distribution of the chip to be cooled and the optimal porosity gradient distribution of the cold plate under one hundred different heat load distribution conditions is established in advance through computational fluid dynamics simulation. The heat load distribution conditions cover all typical conditions such as uniform chip heating, single-hotspot heating, and multi-hotspot heating, with a heat load range of 50W to 300W. During actual operation, based on the real-time collected temperature distribution data, matching searches or linear interpolation calculations are performed in the database to obtain the optimal porosity distribution parameters under the current condition.

[0047] It is worth noting that this database covers all typical operating conditions of data center central processing units, ensuring accurate optimal porosity distribution parameters under any actual operating state. The linear interpolation method can obtain sufficiently accurate results for operating conditions not covered by the database, while requiring minimal computation and meeting the requirements of real-time control.

[0048] The flow regulation response time of the flow regulation element is no more than 100ms.

[0049] The linear interpolation calculation is obtained as follows: In the formula, Indicates the current temperature The corresponding optimal porosity. and This represents two temperature points in the database that are adjacent to the current temperature T. and They represent and The corresponding optimal porosity. This represents the average temperature of the chip surface as collected in real time.

[0050] Example 2

[0051] This embodiment provides a dynamic gradient porosity TPMS heat dissipation cold plate structure and its control method. The dynamic gradient porosity TPMS heat dissipation cold plate structure consists of a cold plate body, a Gyroid-type TPMS flow channel structure, several temperature sensors, several flow regulation elements, and a controller.

[0052] The cold plate body is made of copper alloy and has dimensions of 120mm × 80mm × 10mm. The Gyroid-type TPMS flow channel structure has a wall thickness of 0.5mm and a total length of 120mm, divided into four independent functional sections along the coolant flow direction, each section being 30mm long. Several temperature sensors are evenly arranged in a 2×4 matrix on the bottom surface of the cold plate body, with an adjacent sensor spacing of 20mm. Each functional section corresponds to the acquisition area of ​​two temperature sensors. The temperature sensors have a measurement accuracy of 0.3 degrees Celsius and a sampling frequency of 200Hz. Each functional section outlet is independently equipped with a miniature solenoid valve, with a flow rate adjustment range of 0L / min to 5L / min. The controller uses a DSP digital signal processor, electrically connected to all temperature sensors and miniature solenoid valves.

[0053] A method for controlling the dynamic gradient porosity TPMS heat dissipation cold plate structure includes the following steps: Step 1: Using eight temperature sensors arranged on the bottom surface of the cold plate, the temperature distribution data of each heat-generating area on the surface of the IGBT power module of the new energy vehicle to be cooled is collected in real time.

[0054] Step 2: Based on the preset temperature and porosity mapping model and combined with the real-time collected temperature distribution data, calculate the target porosity value required for each functional segment. Under the premise that the hot spot temperature of the power module to be cooled does not exceed 95 degrees Celsius, take minimizing the total pressure drop of the cold plate system as the optimization objective and determine the target flow rate value corresponding to each functional segment.

[0055] Step 3: Send control signals to the miniature solenoid valves corresponding to each functional segment to adjust the coolant flow rate of each functional segment to the target flow rate value, thereby achieving a dynamic gradient distribution of porosity inside the cold plate.

[0056] The method for establishing the temperature-porosity mapping model is as follows: A database of correspondences between the temperature distribution of the power module to be cooled and the optimal porosity gradient distribution of the cold plate under 150 different heat load distribution conditions is established in advance through computational fluid dynamics simulation. The heat load distribution conditions cover all typical operating conditions of the power module, including startup, acceleration, constant speed, and deceleration, with a heat load range of 100W to 500W. During actual operation, based on the real-time collected temperature distribution data, a matching search or linear interpolation calculation is performed in the database to obtain the optimal porosity distribution parameters under the current operating condition.

[0057] It is worth noting that the thermal load fluctuation frequency of the power module in new energy vehicles is much higher than that of data center chips. Therefore, this embodiment improves the sampling frequency of the temperature sensor and the response speed of the flow regulation element to ensure that the system can respond promptly to rapidly changing thermal loads. Under extreme high-load conditions, the flow rate of the functional segment corresponding to the hotspot can be adjusted to 5L / min to quickly reduce the hotspot temperature.

[0058] The flow regulation response time of the flow regulation element shall not exceed 80ms.

[0059] Experiment Example 1: Heat Dissipation Performance Test The core objective of this test is to verify the advantages of the present invention's technical solution in terms of heat dissipation performance and temperature uniformity across the entire operating range.

[0060] Test items: Hot spot temperature and temperature uniformity tests under different heat load conditions, performed according to standard GB / T40663-2021 "Test Method for Liquid Cooling Performance of Electronic Equipment". Test indicators were: hot spot temperature at 100W, hot spot temperature at 300W, maximum temperature difference at 300W, hot spot temperature at 500W, and maximum temperature difference at 500W. Test samples were cold plate samples prepared in Examples 1, 2, Comparative Example 1, and Comparative Example 2. Test conditions were: ambient temperature 25°C, coolant inlet temperature 25°C, and system inlet pressure 0.2 MPa. Heat loads were set to 100W, 300W, and 500W respectively. Test data were recorded after each heat load condition had been running stably for 30 minutes. Test results are shown in the table below: Test Result Analysis: The hot spot temperatures of Examples 1 and 2 were significantly lower than those of Comparative Examples 1 and 2 under all test conditions. Under a low load of 100W, the hot spot temperature of the present invention was more than 3 degrees Celsius lower than that of Comparative Example 1 and more than 1.5 degrees Celsius lower than that of Comparative Example 2. Under a typical load of 300W, the present invention not only had a lower hot spot temperature, but also controlled the maximum temperature difference within 5 degrees Celsius, which is far superior to 8.9 degrees Celsius in Comparative Example 1 and 7.2 degrees Celsius in Comparative Example 2. Under extreme conditions of 500W, the hot spot temperature of Comparative Example 1 exceeded the safety threshold of 95 degrees Celsius, failing to meet the requirements for normal chip operation; the hot spot temperature of Comparative Example 2 was close to 90 degrees Celsius, and the maximum temperature difference was close to 10 degrees Celsius, posing a risk of local overheating; while the hot spot temperatures of Example 1 and Example 2 were still controlled at 82.1 degrees Celsius and 79.5 degrees Celsius respectively, with a maximum temperature difference of no more than 6 degrees Celsius, fully meeting the requirements for safe chip operation. Example 2, due to the use of copper alloy materials and a higher sensor sampling frequency, had slightly better heat dissipation performance than Example 1.

[0061] Under low-load conditions, the temperature uniformity of the samples was relatively small. Therefore, this test focused on the temperature uniformity performance under typical and extreme conditions. The test results show that the technical solution of this invention, by dynamically adjusting the flow distribution of each functional segment, can accurately match the porosity distribution with the real-time heat load distribution, achieving excellent heat dissipation performance and temperature uniformity across the entire operating range. This effectively solves the problems of insufficient heat dissipation and temperature unevenness in existing static gradient designs and integrated flow regulation designs.

[0062] Experiment Example 2: System Energy Consumption Test The core purpose of this test is to verify the energy-saving effect of the technical solution of this invention.

[0063] Test items: System pump power consumption test under the same heat dissipation effect, the test standard is GB / T40663-2021 "Test method for liquid cooling performance of electronic equipment", and the test indicators are total system flow, total system pressure drop, pump power consumption and data center energy utilization efficiency.

[0064] The test samples were cold plate samples prepared in Examples 1, 2, 1, and 2.

[0065] The test conditions were: ambient temperature (25°C), coolant inlet temperature (25°C), and heat load (300W). The hot spot temperature of the chip for all samples was controlled to not exceed 70°C. The test results are shown in the table below: It is worth noting that the calculation of data center energy efficiency only considers the pump power consumption of the cold plate system and does not include the energy consumption of other auxiliary equipment.

[0066] The energy efficiency is obtained in the following ways: In the formula, This indicates the energy efficiency of the data center. This indicates the pump power consumption of the cold plate system. This indicates the thermal load of the chip.

[0067] Test Result Analysis: Under the same heat dissipation effect, the total system flow, total pressure drop, and pump power consumption of Examples 1 and 2 are significantly lower than those of Comparative Examples 1 and 2. Comparative Example 1 adopts a static gradient design, which requires maintaining a high system flow to ensure heat dissipation under high load conditions, resulting in excessive energy consumption under low load conditions. Comparative Example 2 adopts an overall flow regulation, which requires increasing the overall system flow to control the temperature of local hot spots, leading to unnecessary energy loss. Examples 1 and 2, by dynamically adjusting the flow of each functional segment, provide high flow only in areas with high heat load and reduce flow in areas with low heat load. This significantly reduces the total system flow and total pressure drop while ensuring heat dissipation requirements, and reduces pump power consumption by more than 50%, effectively improving the energy utilization efficiency of the data center.

[0068] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0069] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A dynamic gradient porosity TPMS heat dissipation cold plate structure, characterized in that, It consists of a cold plate body, a Gyroid-type TPMS flow channel structure, several temperature sensors, and several flow regulation elements; The cold plate body is a heat exchange substrate that supports the flow channel structure and is in contact with the chip to be cooled; The Gyroid-type TPMS flow channel structure is located inside the cold plate body and is divided into at least two independent functional sections along the coolant flow direction. The array of temperature sensors is arranged on the contact surface between the cold plate body and the chip to be cooled, and corresponds to different heat-generating areas of the chip to be cooled, for real-time acquisition of temperature signals of each heat-generating area. Each functional segment is equipped with an independent flow regulating element at its inlet or outlet. The flow regulating element is a piezoelectric microvalve or a miniature solenoid valve, which is used to independently regulate the coolant flow rate in the corresponding functional segment. By changing the flow distribution of each functional segment, the porosity of each functional segment can be dynamically and independently adjusted.

2. The dynamic gradient porosity TPMS heat dissipation cold plate structure according to claim 1, characterized in that, The wall thickness of the Gyroid-type TPMS flow channel structure remains constant, ranging from 0.2 mm to 0.5 mm. The total length of the flow channel is equal to the length of the effective heat exchange area of ​​the cold plate, ranging from 80 mm to 120 mm.

3. The dynamic gradient porosity TPMS heat dissipation cold plate structure according to claim 1, characterized in that, The plurality of temperature sensors are arranged in a matrix uniformly on the bottom surface of the cold plate body, with the spacing between adjacent temperature sensors not exceeding 20mm, and each functional segment corresponding to the acquisition area of ​​at least one temperature sensor.

4. The dynamic gradient porosity TPMS heat dissipation cold plate structure according to claim 1, characterized in that, The flow regulating element is a piezoelectric microvalve, and its flow regulating range is 0L / min to 5L / min.

5. The dynamic gradient porosity TPMS heat dissipation cold plate structure according to claim 1, characterized in that, The cold plate body is made of aluminum alloy or copper alloy and also includes a controller. The controller is electrically connected to several temperature sensors and several flow regulating elements to receive temperature signals and output control signals.

6. A method for controlling a dynamic gradient porosity TPMS heat dissipation cold plate structure, characterized in that, To control the dynamic gradient porosity TPMS heat dissipation cold plate structure according to any one of claims 1-5, the following steps are included: Step 1: By using several temperature sensors arranged on the contact surface between the cold plate body and the chip to be cooled, the temperature distribution data of each heat-generating area on the surface of the chip to be cooled is collected in real time. Step 2: Based on the preset temperature and porosity mapping model and combined with the real-time collected temperature distribution data, calculate the target porosity value required for each functional segment, and determine the target flow rate value corresponding to each functional segment accordingly. Step 3: Send control signals to the flow regulation elements corresponding to each functional segment to adjust the coolant flow rate of each functional segment to the target flow rate value, thereby achieving a dynamic gradient distribution of porosity inside the cold plate.

7. The control method for a dynamic gradient porosity TPMS heat dissipation cold plate structure according to claim 6, characterized in that, In step two, the method for establishing the temperature-porosity mapping model is as follows: a database of correspondences between the temperature distribution of the chip to be cooled and the optimal porosity gradient distribution of the cold plate under different heat load distribution conditions is established in advance through computational fluid dynamics (CFD) simulation. In actual operation, based on the real-time collected temperature distribution data, matching or interpolation calculations are performed in the database to obtain the optimal porosity distribution parameters under the current operating conditions.

8. The control method for a dynamic gradient porosity TPMS heat dissipation cold plate structure according to claim 6, characterized in that, In step three, the flow regulation response time of the flow regulation element is ≤100ms.

9. The control method for a dynamic gradient porosity TPMS heat dissipation cold plate structure according to claim 6, characterized in that, In step two, when determining the target flow rate value corresponding to each functional segment, the optimization objective is to minimize the total voltage drop of the cold plate system, provided that the hot spot temperature of the chip to be cooled does not exceed the preset safety threshold.

10. The control method for a dynamic gradient porosity TPMS heat dissipation cold plate structure according to claim 6, characterized in that, When the temperature of a hot spot on a chip to be cooled exceeds a preset safety threshold, the coolant flow rate of the corresponding functional segment of the hot spot is increased first until the temperature of the hot spot drops below the safety threshold.

Images