Manufacturing method of artificial intelligence visual recognition module with three-dimensional heat dissipation structure

Abstract

The present invention discloses a method for manufacturing an artificial intelligence visual recognition module with a three-dimensional heat dissipation structure, which belongs to the field of semiconductor devices. First, a flexible insulating thermal conductive sheet is prepared by a hot pressing process, and a thermal conductive adhesive layer is coated on both sides thereof to form a three-dimensional heat dissipation structure. Secondly, the heat source of the FPGA chip on the middle-layer acceleration module is transferred to the heat dissipation end of the shell in a three-dimensional direction through the three-dimensional heat dissipation structure. This method solves the heat dissipation problem of high-power density chips in a compact space by optimizing the chip layout and heat dissipation path. It has the characteristics of lightweight, high heat dissipation efficiency, and high integration.

Description

Technical Field

[0001] The present invention belongs to the field of semiconductor devices and relates to a method for manufacturing an artificial intelligence visual recognition module with a three-dimensional heat dissipation structure. Background Art

[0002] As the computing power of AI visual recognition chips (such as GPUs, TPUs, and FPGAs) increases, their power density increases significantly, making it difficult for traditional cooling solutions to effectively dissipate heat within a compact space. These chips are subject to long-term, high-load operation in scenarios such as industrial quality inspection, autonomous driving, and medical imaging. Overheating has become a core bottleneck limiting system reliability and lifespan. Existing technologies often rely on indirect heat dissipation through thermal pads between module layers. However, the long distance between the chip and the heat sink and the large spatial span lead to excessive localized temperature rise.

[0003] Due to compact space constraints, cooling technologies like liquid cooling plates are less applicable within modules. An IDC 2024 report indicates that the failure rate of Artificial Intelligence Internet of Things (AIoT) devices due to insufficient cooling is increasing by an average of 12% annually. Therefore, a new cooling method that can directly target chip heat sources and adapt to three-dimensional heat transfer is urgently needed to overcome the conflict between high power density and miniaturization in co-design.

[0004] The three-dimensional heat dissipation structure fabricated by the present invention is insulative, flexible, and has high thermal conductivity. This invention proposes an integrated three-dimensional heat dissipation solution directly targeting chips. Through direct contact between the flexible heat dissipation structure and the chip, the heat conduction path is optimized and the chip operating temperature is reduced. The method provided by the present invention is characterized by simplicity, low cost, energy conservation, and safety. This invention also provides a method for fabricating an artificial intelligence visual recognition module with a three-dimensional heat dissipation structure. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to overcome the shortcomings of the existing technology, ensure the appropriate operating temperature of the chip inside the artificial intelligence visual recognition module, and realize the production of an artificial intelligence visual recognition module with good heat dissipation performance.

[0006] The present invention proposes a method for manufacturing an artificial intelligence visual recognition module with a three-dimensional heat dissipation structure. The detailed steps of the manufacturing process are as follows:

[0007] A method for manufacturing an artificial intelligence visual recognition module with a three-dimensional heat dissipation structure, characterized in that the main components include: a main control module, an acceleration module, a storage module, a box, a fixing column, a fastening device, and a three-dimensional heat dissipation structure, and the manufacturing steps include:

[0008] S01, preparing a three-dimensional heat dissipation structure: cutting a thermally conductive layer and an insulating layer with an adhesive layer on one side or on both sides so that the thermally conductive layer and the insulating layer with the adhesive layer have the same shape, and the area of the thermally conductive layer is smaller than the area of the insulating layer with the adhesive layer; setting the insulating layer as A, the thermally conductive layer as B, and the adhesive layer as C; stacking them in any order of ACBCA, ACBCABCCA, and ACBCABCACBCCA, with the edge of the thermally conductive layer not exceeding the insulating layer; obtaining an insulating thermally conductive sheet by hot pressing; silk-screening thermally conductive adhesive layers on both sides of the insulating thermally conductive sheet; and protecting the thermally conductive adhesive layers with release films or release paper on both sides; then heating and curing to obtain a three-dimensional heat dissipation structure;

[0009] S02, chip processing and integration: The chip is plastic-encapsulated, with a surface roughness Ra range of 1.0-2.0 μm. The main control chip, acceleration chip, and storage chip are then arranged on different modules.

[0010] S03, fixing the main control module and the acceleration module: fix the lower main control module to the fixing column using the fastening device, then fix the middle acceleration module to the fixing column using the fastening device, and keep a safe distance from the lower main control module;

[0011] S04, installing a three-dimensional heat dissipation structure at the heat source end: peeling off the two layers of release film of the three-dimensional heat dissipation structure made in step S01, applying pressure and bonding it to the heat source chip of the middle-layer acceleration module;

[0012] S05, fixing the storage module: fixing the upper storage module to the fixing column by means of a fastening device, so that the upper storage module and the middle acceleration module maintain a safe distance;

[0013] S06, attaching the three-dimensional heat dissipation structure to the heat dissipation end: bend both sides of the three-dimensional heat dissipation structure upward by 180 degrees, apply pressure to adhere it to the upper storage module, and place a thermal pad on the bent three-dimensional heat dissipation structure to obtain an assembled visual recognition module;

[0014] S07, overall assembly: Assemble the four side surfaces and the bottom of the outer shell, place a thermal pad on the bottom inner side of the outer shell, and place the visual recognition module assembled in step S06 into the outer shell. Connect the outer shell to the fixed column, and connect the outer shell cover to the fixed column and the outer shell respectively, so that the outer shell cover is in close contact with the thermal pad in step S06, to obtain an artificial intelligence visual recognition module with a three-dimensional heat dissipation structure.

[0015] Furthermore, it is characterized in that the insulation strength of the insulation layer in step S01 is greater than 10kV / mm, and it contains one or more of polyester film, polyimide film, polypropylene film, and polycarbonate film, with a thickness ranging from 0.05 to 0.2mm, and the thickness of the adhesive layer coated on the single / double side surface of the insulation layer in step S01 is 0.02 to 0.05mm.

[0016] Furthermore, it is characterized in that in the S01 step, the thermal conductive layer is a graphite film, and the graphite film contains one or more of metal powder, silicon powder, boron nitride powder, aluminum nitride powder, silicon nitride powder, silicon carbide powder, and diamond powder. The shape of the above powder is spherical or ellipsoidal, the particle size is 5 to 20 μm, the longitudinal thermal conductivity is greater than 6 W / (m·K), the transverse thermal conductivity is greater than 600 W / (m·K), the thickness is 0.01 mm to 3 mm, and there are a large number of through holes on the surface, the diameter of the through holes is less than 1 mm, and the density of the through holes is in the range of 5 to 10 / cm 2 The hot pressing temperature of the step S01 is 80 to 120°C, the hot pressing pressure is 0.2 to 2 MPa, and the hot pressing time is 120 to 300 s;

[0017] Furthermore, it is characterized in that the thermal conductivity of the thermally conductive adhesive layer in step S01 is 3-8 W / (m·K), the thickness is 0.5-1 mm, and the peel strength is 4-8 N / mm; the thermally conductive adhesive layer in step S01 contains one or more of a resin matrix and a thermally conductive filler, the resin matrix includes one or more of an acrylic resin and a silicone resin, and the thermally conductive filler includes one or more of silica, aluminum oxide, magnesium oxide, zinc oxide, silicon, boron nitride, aluminum nitride, silicon nitride, silicon carbide, diamond, carbon nanotubes, graphene, graphite nanosheets, and metal particles, and the particle size of the thermally conductive filler ranges from 0.01 to 50 μm. The heating and curing temperature in step S02 is 60-120°C, and the heating time is 10-200 minutes. The three-dimensional heat dissipation structure in step S01 is flexible and can be bent more than 180° with a thermal conductivity drop of less than 5% after bending.

[0018] Furthermore, it is characterized in that the main control module in the step S03 has a main control chip, including one or more of a graphics processing unit (GPU), a tensor processing unit (TPU), a digital signal processor (DSP), a system-on-chip (SoC), a microcontroller unit (MCU), and a vision processing unit (VPU), and the power density of the main control chip is 5-50W / cm 2 ; The safety distance in the S03 step is 10-40mm.

[0019] Furthermore, it is characterized in that the acceleration module in step S03 has a field programmable gate array (FPGA) chip to accelerate image preprocessing and real-time reasoning; the high-power units (such as DSP and high-speed transceiver) inside the FPGA chip are designed to be distributed at the edge of the chip to accelerate heat transfer; the power density range of the FPGA chip is 5-30W / cm 2 .

[0020] Furthermore, it is characterized in that the storage module in the S05 step includes one or more of dynamic random access memory (DRAM), static random access memory (SRAM), Flash memory, programmable ROM (PROM), erasable PROM (EPROM), and electrically erasable PROM (EEPROM), and the safety distance is 10-40 mm.

[0021] Furthermore, it is characterized in that the material of the fixing column in steps S03, S05, and S07 is one or more of aluminum and its alloys, titanium and its alloys, and carbon fiber composite materials; the material of the outer shell and the outer shell cover in step S07 is one or more of aluminum and its alloys, titanium and its alloys, and carbon fiber composite materials, and the outer shell and the outer shell cover are provided with air cooling or liquid cooling measures.

[0022] Furthermore, it is characterized in that the thermal conductivity of the thermal pad in steps S06 and S07 is 3.5-8 W / (m·K), the Shore hardness is 30-50, and the thickness is 2-5 mm.

[0023] An artificial intelligence visual recognition module with a three-dimensional heat dissipation structure is characterized in that the artificial intelligence visual recognition module with a three-dimensional heat dissipation structure is prepared by the above method and has the function of transferring heat from the middle-layer acceleration module chip in a three-dimensional direction to reduce the operating temperature of the chip. Compared with the existing technology, the advantages of the present invention are:

[0024] (1) The heat of the FPGA acceleration chip on the middle acceleration module of the artificial intelligence visual recognition module prepared by this method can be transferred to the cold end of the shell through the three-dimensional heat dissipation structure, thereby reducing the operating temperature of the FPGA chip and improving the image processing efficiency;

[0025] (2) The heat of the main control chip on the lower main control module of the artificial intelligence visual recognition module prepared by this method can be directly transferred to the cold end of the shell through the thermal pad, thereby reducing the operating temperature of the main control chip;

[0026] (3) The method provided by the present invention has the characteristics of simple process and low cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Attachment Figure 1 It is the overall manufacturing flow chart of the present invention.

[0028] Attachment Figure 2 It is a structural explosion diagram of the present invention.

[0029] Attachment Figure 3 It is a schematic diagram of the three-dimensional structure of the present invention in an application scenario.

[0030] Attachment Figure 4 It is an exploded view of the three-dimensional heat dissipation structure of the present invention.

[0031] Attachment Figure 5 This is a comparison chart of the operating temperatures of the middle-layer FPGA chips of the embodiment and the comparative example at the same power. DETAILED DESCRIPTION

[0032] In the following description, only certain exemplary embodiments are briefly described. As will be appreciated by those skilled in the art, the described embodiments may be modified in various ways without departing from the spirit or scope of the present invention. Therefore, the drawings and description are to be regarded as illustrative in nature and not restrictive.

[0033] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0034] In the description of the present invention, it should be understood that the terms "center," "longitudinal," "transverse," "length," "width," "thickness," "up," "down," "front," "back," "left," "right," "vertical," "horizontal," "top," "bottom," "inside," "outside," "clockwise," "counterclockwise," and the like, indicating positions or relationships, are based on the positions or relationships shown in the accompanying drawings and are intended solely for the purpose of facilitating the description of the present invention and simplifying the description. They do not indicate or imply that the devices or components referred to must have a specific orientation, be constructed, or operate in a specific orientation, and therefore should not be construed as limiting the present invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features referred to. Thus, features defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, "plurality" means two or more, unless otherwise specifically defined.

[0035] The disclosure below provides many different embodiments or examples for implementing the present invention. In order to simplify the disclosure of the present invention, specific examples are described below. Of course, they are merely examples and are not intended to limit the present invention. In addition, the present invention may repeat reference numerals and / or reference letters in different examples. Such repetition is for the purpose of simplicity and clarity and does not in itself indicate the relationship between the various embodiments and / or settings discussed. In addition, the present invention provides examples of various specific processes and materials, but those skilled in the art will recognize the application of other processes and / or the use of other materials.

[0036] Example

[0037] The embodiment specifically describes the application of an artificial intelligence visual recognition module with a three-dimensional heat dissipation structure.

[0038] Refer to the attached Figure 2 With attached Figure 3 The artificial intelligence visual recognition module with a three-dimensional heat dissipation structure mainly includes the following parts: a shell cover 101, a thermal pad 102, a three-dimensional heat dissipation structure 103, an upper storage module 104, a middle acceleration module 105, a lower main control module 106, a fixing column 107, a fastening device 108, and an outer shell 109;

[0039] Refer to the attached Figure 4 , the manufactured three-dimensional heat dissipation structure mainly includes the following parts: thermal conductive adhesive layer 201, insulating layer 202, adhesive layer 203, thermal conductive layer 204;

[0040] S01, prepare a three-dimensional heat dissipation structure: select a polyimide film with a thickness of 0.2mm and a dielectric strength of 12kV / mm as the insulating layer 202, and coat its surface with a 0.02mm epoxy resin adhesive layer 203. Select a graphite film with a longitudinal thermal conductivity of 85W / (m•K) and a transverse thermal conductivity of 800W / (m•K) as the heat conducting layer 204. Aluminum nitride spheres with a particle size of 10μm are added to the inside of the graphite film. The through-hole diameter of the surface is 0.8mm and the density is 8 / cm 2 , the thermal conductive layer 204 is cut into a rectangle of 35×170 mm, and the insulating layer 202 with the adhesive layer 203 is cut into a rectangle of 40×180 mm, and the insulating layer 202 is set as A, the thermal conductive layer 204 is set as B, and the adhesive layer 203 is set as C. They are stacked in the order of ACBCA, and the edge of the thermal conductive layer 204 does not exceed the insulating layer 202. They are hot-pressed at 120°C and 1 MPa pressure for 300 seconds to obtain an insulating thermal conductive sheet; 5 μm particle size aluminum oxide powder and 10 μm particle size silicon carbide powder are added to the silicone resin to prepare a thermal conductive adhesive, and a thermal conductive adhesive layer 201 with a thickness of 0.5 mm, a thermal conductivity of 5 W / (m·K), and a peel strength of 6 N / mm is coated on both sides of the insulating thermal conductive sheet prepared in step S01 by screen printing, and the thermal conductive adhesive layer is protected with a PET release film on both sides. Then, it is heated and cured at 100°C for 30 minutes to obtain a three-dimensional heat dissipation structure 103;

[0041] S02, production of each module: Resistors, capacitors and other components are soldered to different modules through the reflow process, and the upper surface of the plastic-encapsulated FPGA chip with Ra of 1.4~1.6μm, which is in direct contact with the three-dimensional heat dissipation structure 103, is processed so that the main control chip based on the ARM architecture is reflow-welded on the lower main control module 106, the FPGA chip is reflow-welded on the middle acceleration module 105, and the EEPROM and DRAM memory are reflow-welded on the upper storage module 104. The main control chip and the FPGA chip are interconnected using AXI. The size of the three modules is 100×100mm. A circular limit hole is punched at each corner of the module. The center of the circle is located on the diagonal of the square, 5mm away from the boundary line, and the radius of the circle is 2.5mm.

[0042] S03, Main control and acceleration module fixing: The lower-layer main control module 106 with the ARM architecture-based main control chip is fixed to the surrounding fixing columns 107 using the fastening device 108. Then, the middle-layer acceleration module 105 with the FPGA chip is fixed to the fixing columns 107 using the fastening device 108, maintaining a safe distance of 20mm from the lower-layer module;

[0043] S04: Install a three-dimensional heat dissipation structure at the heat source end: Peel off the two layers of PET release film of the three-dimensional heat dissipation structure 103 made in step S01, apply a force of 0.1 MPa and adhere it to the heat source FPGA chip in the middle of the middle acceleration module 105. The FPGA chip size is 35×35 mm, and the heat power density is 10 W / cm 2 ;

[0044] S05, upper module fixation: The upper storage module 104 with EEPROM and DRAM is fixed to the fixing column 107 by the fastening device 108, so that the upper storage module 104 and the middle acceleration module 105 maintain a safe distance of 20 mm;

[0045] S06, attaching the three-dimensional heat dissipation structure to the heat dissipation end: Bend both sides of the three-dimensional heat dissipation structure 103 upward by 180 degrees, apply a force of 0.1 MPa to adhere it to the upper storage module 104, and place a thermal pad 102 with a thermal conductivity of 5 W / (m•K), a Shore hardness of 45, and a thickness of 5 mm on the bent three-dimensional heat dissipation structure 103 to obtain an assembled visual recognition module.

[0046] S07, overall assembly: Assemble the four side surfaces and the bottom of the outer shell 109, place a thermal pad with a thermal conductivity of 8 W / (m•K), a Shore hardness of 45, and a thickness of 3 mm on the inner bottom of the outer shell 109, and place the visual recognition module assembled in step S07 into the outer shell 109, connect the outer shell 109 to the fixed column 107, and connect the outer shell cover 101 to the fixed column 107 and the outer shell 109 respectively, so that the outer shell cover 101 and the thermal pad 102 in step S07 are in close contact, thereby obtaining an artificial intelligence visual recognition module with a three-dimensional heat dissipation structure.

[0047] When the AI visual recognition module is operating, heat generated by the FPGA chip is transferred longitudinally to the thermally conductive layer of the three-dimensional heat dissipation structure. Heat is then transferred along this layer in three-dimensional space to the upper thermal pad, which then transfers it to the outer cover, achieving effective heat dissipation. The main control module is close to the cold end of the outer shell, allowing direct heat transfer using the highly conductive pad. This system effectively dissipates heat in three dimensions, effectively reducing the temperature of the main control and accelerator chips.

[0048] Although the present invention has been described with reference to exemplary embodiments, it should be understood that the terms used are illustrative and exemplary rather than restrictive. Since the present invention can be embodied in many forms without departing from the spirit or essence of the invention, it should be understood that the above embodiments are not limited to any of the foregoing details.

[0049] Comparative Example

[0050] Except that the three-dimensional heat dissipation structure and the thermal pad on the upper storage module are not installed, other structures and processes are the same as those in the embodiment.

Claims

1. A method for manufacturing an artificial intelligence visual recognition module with a three-dimensional heat dissipation structure, characterized in that: Composition includes: The manufacturing steps of the main control module, acceleration module, storage module, housing, fixing column, fastening device and three-dimensional heat dissipation structure include: S01, preparing a three-dimensional heat dissipation structure: cutting a thermally conductive layer and an insulating layer with an adhesive layer on one side or on both sides so that the thermally conductive layer and the insulating layer with the adhesive layer have the same shape, and the area of the thermally conductive layer is smaller than the area of the insulating layer with the adhesive layer; setting the insulating layer as A, the thermally conductive layer as B, and the adhesive layer as C; stacking them in any order of ACBCA, ACBCABCCA, and ACBCABCACBCCA, with the edge of the thermally conductive layer not exceeding the insulating layer; obtaining an insulating thermally conductive sheet by hot pressing; silk-screening thermally conductive adhesive layers on both sides of the insulating thermally conductive sheet; and protecting the thermally conductive adhesive layers with release films or release paper on both sides; then heating and curing to obtain a three-dimensional heat dissipation structure; S02, chip processing and integration: The chip is plastic-encapsulated, with a surface roughness Ra range of 1.0-2.0 μm. The main control chip, acceleration chip, and storage chip are then placed on the main control module, acceleration module, and storage module respectively; S03, fixing the main control module and the acceleration module: fix the lower main control module to the fixing column using the fastening device, then fix the middle acceleration module to the fixing column using the fastening device, and keep a safe distance from the lower main control module; S04, installing a three-dimensional heat dissipation structure at the heat source end: peeling off the two layers of release film or release paper of the three-dimensional heat dissipation structure made in step S01, applying pressure and bonding it to the heat source chip of the middle-layer acceleration module; S05, fixing the storage module: fixing the upper storage module to the fixing column by means of a fastening device, so that the upper storage module and the middle acceleration module maintain a safe distance; S06, attaching the three-dimensional heat dissipation structure to the heat dissipation end: bend both sides of the three-dimensional heat dissipation structure upward by 180 degrees, apply pressure to adhere it to the upper storage module, and place a thermal pad on the bent three-dimensional heat dissipation structure to obtain an assembled visual recognition module; S07, overall assembly: Assemble the four side surfaces and the bottom of the outer shell, place a thermal pad on the bottom inner side of the outer shell, and place the visual recognition module assembled in step S06 into the outer shell. Connect the outer shell to the fixed column, and connect the outer shell cover to the fixed column and the outer shell respectively, so that the outer shell cover is in close contact with the thermal pad in step S06, to obtain an artificial intelligence visual recognition module with a three-dimensional heat dissipation structure.

2. The production method according to claim 1, characterized in that The insulating layer of step S01 has a dielectric strength greater than 10 kV / mm, and contains one or more of polyester film, polyimide film, polypropylene film, and polycarbonate film, with a thickness ranging from 0.05 to 0.2 mm. The thickness of the adhesive layer coated on one or both sides of the insulating layer of step S01 is 0.02 to 0.05 mm.

3. The production method according to claim 1, characterized in that In step S01, the thermal conductive layer is a graphite film, and the graphite film contains one or more of metal powder, silicon powder, boron nitride powder, aluminum nitride powder, silicon nitride powder, silicon carbide powder, and diamond powder. The above powders are spherical or ellipsoidal in shape, with a particle size of 5 to 20 μm, a longitudinal thermal conductivity greater than 6 W / (m·K), a transverse thermal conductivity greater than 600 W / (m·K), a thickness of 0.01 mm to 3 mm, and a large number of through holes on the surface. The diameter of the through holes is less than 1 mm, and the density of the through holes is in the range of 5 to 10 / cm 2 The hot pressing temperature of the S01 step is 80 to 120°C, the hot pressing pressure is 0.2 to 2 MPa, and the hot pressing time is 120 to 300 seconds.

4. The production method according to claim 1, characterized in that The thermal conductivity of the thermally conductive adhesive layer in step S01 is 3 to 8 W / (m·K), the thickness is 0.5 to 1 mm, and the peel strength is 4 to 8 N / mm; the thermally conductive adhesive layer in step S01 contains one or more of a resin matrix and a thermally conductive filler, the resin matrix includes one or more of an acrylic resin and a silicone resin, the thermally conductive filler includes one or more of silica, aluminum oxide, magnesium oxide, zinc oxide, silicon, boron nitride, aluminum nitride, silicon nitride, silicon carbide, diamond, carbon nanotubes, graphene, graphite nanosheets, and metal particles, the particle size range of the thermally conductive filler is 0.01 to 50 μm, the heating and curing temperature in step S02 is 60 to 120°C, and the heating time is 10 to 200 min; the three-dimensional heat dissipation structure in step S01 is flexible and can withstand 180° bending, and the thermal conductivity drops by less than 5% after 100 bends.

5. The production method according to claim 1, characterized in that: The main control module in step S03 has a main control chip, including one or more of a graphics processing unit (GPU), a tensor processing unit (TPU), a digital signal processor (DSP), a system-on-chip (SoC), a microcontroller unit (MCU), and a visual processing unit (VPU); the power density of the main control chip is 5~50W / cm 2 ; The safety distance in the S03 step is 10-40mm.

6. The production method according to claim 1, characterized in that: The acceleration module in step S03 has a field programmable gate array (FPGA) chip to accelerate image preprocessing and real-time reasoning; the high-power consumption units inside the FPGA chip are distributed at the edge of the chip to accelerate heat dissipation. The power density of the FPGA chip ranges from 5 to 30W / cm 2 .

7. The production method according to claim 1, characterized in that: The storage module in step S05 includes one or more of dynamic random access memory (DRAM), static random access memory (SRAM), Flash memory, programmable ROM (PROM), erasable PROM (EPROM), and electrically erasable PROM (EEPROM), and the safety distance is 10-40 mm.

8. The production method according to claim 1, characterized in that: The material of the fixing column in steps S03, S05, and S07 is one or more of aluminum and its alloys, titanium and its alloys, and carbon fiber composite materials; the material of the outer shell and the outer shell cover in step S07 is one or more of aluminum and its alloys, titanium and its alloys, and carbon fiber composite materials, and the outer shell and the outer shell cover are provided with air cooling or liquid cooling measures.

9. The production method according to claim 1, characterized in that: The thermal conductivity of the thermal pad in steps S06 and S07 is 3.5-8 W / (m·K), the Shore hardness is 30-50, and the thickness is 2-5 mm.

10. An artificial intelligence visual recognition module with a three-dimensional heat dissipation structure, characterized in that: The artificial intelligence visual recognition module with a three-dimensional heat dissipation structure is prepared by the method described in any one of claims 1 to 9, and has the function of transferring the heat of the middle-layer acceleration module chip in a three-dimensional direction to reduce the operating temperature of the chip.

Images