Space satellite AI computing payload device

Abstract

The application relates to the field of aerospace satellite AI computing technology, in particular to a kind of aerospace satellite AI computing load equipment, including AI computing board frame one, the inside of AI computing board frame one is fixed with PCB-AI computing board card one by screw, one side of AI computing board frame one is provided with AI computing board frame two, and the inside of AI computing board frame two is installed with PCB-AI computing board card two by screw, one side of AI computing board frame two is provided with AI computing board frame three, and the inside of AI computing board frame three is installed with PCB-AI computing board card three by screw, and one side of AI computing board frame three is covered with bottom cover. The application integrates multiple AI computing board cards in a unified device through a multi-layer embedded modular structure design, and has the characteristics of compact structure, good heat dissipation performance, strong impact resistance and flexible expansion.

Description

Technical Field

[0001] This invention relates to the field of aerospace satellite AI computing technology, specifically to an aerospace satellite AI computing payload device. Background Technology

[0002] With the rapid development of aerospace and information technology, satellite systems are playing an increasingly important role in remote sensing, space communication, space situational awareness, navigation and positioning, and deep space exploration. Modern space missions typically carry various types of payloads, such as optical remote sensing cameras, infrared detectors, laser communication terminals, synthetic aperture radar, and multispectral imaging equipment. These payloads continuously generate massive amounts of observational data during operation, placing increasingly higher demands on satellite information processing capabilities.

[0003] In traditional satellite systems, data acquired by the payload is typically stored in its raw form on internal satellite storage devices and then transmitted to ground stations for processing via data transmission links. Due to the limited downlink communication bandwidth of satellites, especially in high-resolution remote sensing missions where the amount of data acquired by the satellite far exceeds the amount that the communication link can transmit in real time, long-term storage and batch transmission are often required. This not only increases data processing latency but also limits the satellite's real-time response capability to unexpected events.

[0004] In recent years, with the development of artificial intelligence technology, especially deep learning technology, on-orbit intelligent processing has gradually become an important direction for the development of satellite systems. By deploying high-performance computing platforms on satellites, tasks such as image recognition, target detection, change detection, and data compression can be completed in orbit, thereby transmitting only valuable information or processing results to the ground, significantly reducing data transmission pressure and improving system response efficiency. For example, in remote sensing satellite applications, on-orbit AI processing can achieve ship identification, vehicle detection, disaster identification, cloud filtering, and target priority downlink. Therefore, spaceborne AI computing payloads are gradually becoming an important component of next-generation satellite systems.

[0005] However, space and power resources within spacecraft are extremely limited, while requirements for vibration resistance, radiation resistance, thermal control, and high reliability must also be met. Therefore, the structural design of spaceborne AI computing devices needs to possess ultra-high computing density, low power consumption, high reliability, modular structure, efficient heat dissipation capabilities, and good system scalability. Currently, spaceborne AI computing devices still suffer from problems such as insufficient system scalability, low device integration, inadequate thermal management capabilities, insufficient structural reliability, and difficulties in system maintenance and upgrades. Summary of the Invention

[0006] The purpose of this invention is to provide an AI computing payload device for aerospace satellites to solve the problems mentioned in the background art, such as insufficient scalability, low integration, poor thermal management, and insufficient structural reliability of existing spaceborne AI computing devices.

[0007] To achieve the above objectives, the present invention provides the following technical solution: an AI computing payload device for aerospace satellites, comprising an AI computing board frame one, wherein a PCB-AI computing board card one is fixedly connected to the interior of the AI ​​computing board frame one by screws; an AI computing board frame two is provided on one side of the AI ​​computing board frame one, and a PCB-AI computing board card two is installed inside the AI ​​computing board frame two by screws; an AI computing board frame three is provided on one side of the AI ​​computing board frame two, and a PCB-AI computing board card three is installed inside the AI ​​computing board frame three by screws; and a bottom cover is fitted onto one side of the AI ​​computing board frame three. Each of the four corners of the bottom cover surface is provided with a fastening bolt, which passes through the bottom cover and the AI ​​computing board frame three in sequence and is threadedly fixed to the AI ​​computing board frame two. The surface of the AI ​​computing board frame one is covered with a heat dissipation cover plate, which is fixed to the AI ​​computing board frame one by screws. Each of the four corners of the surface of the AI ​​computing board frame one is provided with a locking bolt, which passes through the AI ​​computing board frame one in sequence and is threadedly fixed to the AI ​​computing board frame two. The surface of the AI ​​computing board frame one is provided with an end cover, which covers the interface of the PCB-AI computing board card one.

[0008] Preferably, the AI ​​computing board frame is an integrated frame structure formed by CNC machining of magnesium alloy or aluminum alloy, with reinforcing ribs around its perimeter, and threaded mounting posts and positioning countersunk holes inside for fixing the PCB-AI computing board card; the upper surface of the AI ​​computing board frame has a flat heat-conducting plane for fitting the heat dissipation cover, and through holes for locking bolts are opened at the four corners of the AI ​​computing board frame.

[0009] Preferably, the heat dissipation cover is a flat plate structure formed by CNC machining of magnesium alloy or aluminum alloy, and its outer surface is provided with multiple parallel heat dissipation fins, with heat dissipation channels formed between adjacent heat dissipation fins; the bottom of the heat dissipation cover is provided with a flat heat-conducting surface that is in close contact with the upper surface of the AI ​​computing board frame, and countersunk holes for screws to pass through are provided at the four corners of the heat dissipation cover.

[0010] Preferably, the end cap is a thin sheet structure integrally formed of metal or plastic, and its shape matches the interface outline of the PCB-AI computing board. The end cap is provided with buckles or screw holes around its periphery for detachably fixing the end cap to the interface opening of the AI ​​computing board frame.

[0011] Preferably, the second AI computing board frame is a rectangular frame structure formed by CNC machining of magnesium alloy or aluminum alloy, and both its upper and lower surfaces are provided with boss fitting structures; the boss fitting structure includes multiple upwardly protruding positioning bosses and downwardly recessed positioning grooves. The positioning bosses fit into the corresponding grooves at the bottom of the first AI computing board frame, and the positioning grooves fit into the corresponding bosses at the top of the third AI computing board frame, so as to achieve precise alignment and positioning of the multi-layer board frame; the four corners of the second AI computing board frame are provided with internal threaded holes for fastening bolts to pass through.

[0012] Preferably, the AI ​​computing board frame three is a frame structure formed by CNC machining of magnesium alloy or aluminum alloy. Its top is provided with a positioning boss that matches the bottom groove of the AI ​​computing board frame two, and its bottom is provided with a flat surface for fitting the bottom cover. The four corners of the AI ​​computing board frame three are provided with through holes for fastening bolts to pass through, and the upper end of the through hole is provided with a countersunk hole.

[0013] Preferably, the bottom cover is a flat plate structure formed by CNC machining of magnesium alloy or aluminum alloy, and its bottom is provided with multiple mounting interfaces for fixed connection with the satellite platform. The mounting interfaces are threaded holes or through holes. The inner upper surface of the bottom cover is provided with positioning grooves for accommodating heat-conducting pads, and countersunk holes are provided at the four corners of the bottom cover for fastening bolts to pass through. The periphery of the bottom cover is provided with upward-curving positioning flanges.

[0014] Preferably, the PCB-AI computing board is a multilayer printed circuit board structure, with its substrate being FR-4 or polyimide material, and AI computing chips, power modules, and rectangular connectors of various specifications soldered on its upper and lower surfaces respectively; positioning holes for screws to pass through are provided at the four corners of the PCB-AI computing board, and positioning notches that mate with the inner wall of the AI ​​computing board frame are provided on its edges.

[0015] Preferably, the second PCB-AI computing board has a multilayer printed circuit board structure, with its substrate being FR-4 or polyimide material, and its surface integrated with a data exchange chip and a high-speed connector; the four corners of the second PCB-AI computing board have positioning holes for screws to pass through, and its upper and lower surfaces are respectively fixedly connected to the second AI computing board frame by screws; the second PCB-AI computing board is electrically connected to the first PCB-AI computing board and the third PCB-AI computing board through a board-to-board connector.

[0016] Preferably, the PCB-AI computing board three is a multilayer printed circuit board structure, with its substrate being FR-4 or polyimide material, and its surface integrating redundant control chips and fault detection circuits; the four corners of the PCB-AI computing board three are provided with positioning holes for screws to pass through, and its edges are provided with positioning structures that cooperate with the inner wall of the AI ​​computing board frame three; the PCB-AI computing board three is electrically connected to the PCB-AI computing board two through a board-to-board connector, and fault takeover is achieved through redundant design.

[0017] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention integrates multiple AI computing boards into a unified device through a multi-layered, modular design, which solves the problem of insufficient scalability of traditional spaceborne computing platforms and enables the system to flexibly configure computing resources according to mission requirements.

[0018] 2. The invention uses a shell structure made of magnesium alloy or aluminum alloy by CNC machining, which can withstand large and severe vibrations and impacts. The internal multi-layer shock absorption structure effectively absorbs the impact force and protects various high-reliability computing chips.

[0019] 3. This invention constructs a hierarchical modular structure, which separates the functions of the AI ​​computing module, the data storage module, and the power management module and integrates them structurally. This enables the system to flexibly configure computing and storage resources according to mission requirements, meeting the high standards of computing power required by different aerospace satellite missions.

[0020] 4. This invention integrates the AI ​​computing module, storage module, and power management module into a unified device structure through a multi-layered, interlocking structural design, and optimizes the module layout, thereby effectively reducing the device size and weight, improving system integration, and meeting the stringent requirements of spacecraft for size, weight, and power consumption.

[0021] 5. This invention uses a metal heat conduction structure design to enable the heat generated by the computing module to be quickly conducted to the satellite structure through the device housing. At the same time, the combination of structural heat conduction design and shell heat dissipation structure improves the overall heat dissipation capacity of the system, thereby ensuring the long-term stable operation of the AI ​​computing payload in complex space environment.

[0022] 6. This invention improves the structural rigidity and vibration resistance of the equipment through an integrated overall structural design and modular integrated assembly installation method, enabling the equipment to maintain stable operation in aerospace launch and on-orbit operation mission environments, thereby improving the overall reliability of the system. Attached Figure Description

[0023] Figure 1 This is a three-dimensional structural diagram of the present invention; Figure 2This is a schematic diagram of the main appearance structure of the present invention; Figure 3 This is a top view of the external structure of the present invention; Figure 4 This is a schematic diagram of the rear-view partially exploded structure of the present invention; Figure 5 This is a schematic diagram of the front-view partially exploded structure of the present invention; Figure 6 This is a front-view exploded magnification diagram of the present invention; Figure 7 This is a rear-view exploded structure diagram of the present invention; Figure 8 This is a front-view exploded structure diagram of the present invention.

[0024] In the diagram: 1. AI computing board frame one; 2. Heat sink cover; 3. End cover; 4. AI computing board frame two; 5. AI computing board frame three; 6. Bottom cover; 7. PCB-AI computing board card one; 8. PCB-AI computing board card two; 9. PCB-AI computing board card three; 10. Fastening bolt; 11. Locking bolt. Detailed Implementation

[0025] 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, not all embodiments. In addition, the terms "first," "second," "third," "upper," "lower," "left," "right," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance. At the same time, in the description of the present invention, unless otherwise explicitly specified and limited, the terms "connected" and "linked" should be interpreted broadly. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium. 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.

[0026] The structure of an AI computing payload device for aerospace satellites provided by this invention is as follows: Figure 1 as well as Figure 2 As shown, the system includes an AI computing board frame 1, which serves as the outer shell of the aerospace satellite's AI computing payload equipment, protecting the internal electronic components and effectively resisting vibrations, impacts, and external physical damage that may be encountered during flight. A heat dissipation cover 2 is fitted onto the surface of the AI ​​computing board frame 1, and this cover 2 is fixed to the AI ​​computing board frame 1 with screws. The heat dissipation cover 2 is the core of the aerospace AI computing equipment's heat outlet and temperature regulation, and is crucial to determining whether the chip can operate stably for a long time in extreme environments.

[0027] Among them, the AI ​​computing board frame 1 can be assembled with the heat sink cover and other structural components of the module layer. It is not only the bracket of the PCB computing board, but also the structural skeleton, heat dissipation channel and precision assembly platform of the entire AI payload, which directly determines the reliability and performance of the equipment. The heat sink cover 2 is part of the shell and serves as the heat outlet and temperature regulation core of the aerospace AI computing equipment. It is crucial to whether the chip can work stably for a long time in extreme environments.

[0028] Specifically, the AI ​​computing board frame 1 is an integrated frame structure formed by CNC machining of magnesium alloy or aluminum alloy. It is equipped with reinforcing ribs around its perimeter and has threaded mounting posts and positioning countersunk holes inside for fixing the PCB-AI computing board card 7. The upper surface of the AI ​​computing board frame 1 is provided with a flat heat-conducting plane for fitting the heat dissipation cover 2. The four corners of the AI ​​computing board frame 1 are provided with through holes through which locking bolts 11 pass.

[0029] The heat dissipation cover 2 is a flat plate structure formed by CNC machining of magnesium alloy or aluminum alloy. Multiple parallel heat dissipation fins are provided on its outer surface, and heat dissipation channels are formed between adjacent heat dissipation fins. The bottom of the heat dissipation cover 2 is provided with a flat heat-conducting surface that is in close contact with the upper surface of the AI ​​computing board frame 1, and countersunk holes for screws to pass through are provided at the four corners of the heat dissipation cover 2.

[0030] Furthermore, such as Figure 3 as well as Figure 4 As shown, PCB-AI computing board 7 is fixed to the inside of AI computing board frame 1 by screws. AI computing board frame 2 4 is set on one side of AI computing board frame 1, and PCB-AI computing board 8 is installed inside AI computing board frame 2 4 by screws. AI computing board frame 2 4 adopts a boss-fitting structure design, which facilitates precise and accurate positioning. It is the core hub of the entire AI load structure, and at the same time undertakes structural force transmission, heat relay and multi-board collaborative positioning.

[0031] Among them, the AI ​​computing board frame 2 4 adopts a boss-fitting structure design, which facilitates precise and accurate positioning. It is the core hub of the entire AI load structure, and at the same time undertakes structural force transmission, heat relay and multi-board collaborative positioning. Preferably, the PCB-AI computing board 1 7 is a multi-layer printed circuit board structure. Its substrate is made of FR-4 or polyimide material. The AI ​​computing chip, power module and various sizes of rectangular connectors are soldered on its upper and lower surfaces, respectively. The four corners of the PCB-AI computing board 1 7 are provided with positioning holes for screws to pass through, and its edges are provided with positioning notches that mate with the inner wall of the AI ​​computing board frame 1.

[0032] Specifically, the AI ​​computing board frame 2 4 is a rectangular frame structure formed by CNC machining of magnesium alloy or aluminum alloy, and both its upper and lower surfaces are provided with boss fitting structures; the boss fitting structure includes multiple upwardly protruding positioning bosses and downwardly recessed positioning grooves. The positioning bosses fit into the corresponding grooves at the bottom of the AI ​​computing board frame 1, and the positioning grooves fit into the corresponding bosses at the top of the AI ​​computing board frame 3 5, so as to achieve precise alignment and positioning of the multi-layer board frame; the four corners of the AI ​​computing board frame 2 4 are provided with internal threaded holes for fastening bolts to pass through.

[0033] Furthermore, such as Figure 5 as well as Figure 6 As shown, AI computing board frame 3 (5) is located on one side of AI computing board frame 2 (4), and PCB-AI computing board card 3 (9) is installed inside AI computing board frame 3 (5) with screws. AI computing board frame 3 (5) is the bottom support and finishing node of the entire structure, which not only bears the end point of force transmission, but also provides an auxiliary heat dissipation path.

[0034] Specifically, the AI ​​computing board frame 35 is a frame structure formed by CNC machining of magnesium alloy or aluminum alloy. Its top has a positioning boss that mates with the bottom groove of the AI ​​computing board frame 24, and its bottom has a flat surface for fitting the bottom cover 6. The four corners of the AI ​​computing board frame 35 have through holes for fastening bolts 10 to pass through, and the upper end of each through hole has a countersunk hole. The PCB-AI computing board card 39 is a multilayer printed circuit board structure. Its substrate is FR-4 or polyimide material, and its surface integrates redundant control chips and fault detection circuits. The four corners of the PCB-AI computing board card 39 have positioning holes for screws to pass through, and its edges have positioning structures that mate with the inner wall of the AI ​​computing board frame 35. The PCB-AI computing board card 39 is electrically connected to the PCB-AI computing board card 28 via a board-to-board connector, and fault management is achieved through redundant design.

[0035] Furthermore, such as Figure 7 as well as Figure 8 As shown, a bottom cover 6 is fitted onto one side of the AI ​​computing board frame 3 5. The bottom cover 6 is not only a sealing component, but also the basic platform for the entire AI computing load, undertaking key functions such as structural load-bearing, external installation, auxiliary heat dissipation, and shielding closure. Fastening bolts 10 are provided at the four corners of the surface of the bottom cover 6. These fastening bolts 10 pass through the bottom cover 6 and the AI ​​computing board frame 3 5 in sequence and are threadedly fixed to the AI ​​computing board frame 2 4.

[0036] Specifically, the bottom cover 6 is a flat plate structure formed by CNC machining of magnesium alloy or aluminum alloy. Its bottom is provided with multiple mounting interfaces for fixed connection with the satellite platform. The mounting interfaces are threaded holes or through holes. The inner upper surface of the bottom cover 6 is provided with positioning grooves for accommodating heat-conducting pads, and countersunk holes are provided at the four corners of the bottom cover 6 for fastening bolts 10 to pass through. The bottom cover 6 is provided with upward-folded positioning flanges around its perimeter.

[0037] Furthermore, such as Figures 6 to 8 Locking bolts 11 are provided at the four corners of the surface of the AI ​​computing board frame 1 shown. The locking bolts 11 pass through the AI ​​computing board frame 1 and are threadedly fixed to the AI ​​computing board frame 4.

[0038] Furthermore, such as Figure 8 As shown, an end cap 3 is provided on the surface of the AI ​​computing board frame 1, which covers the interface of the PCB-AI computing board 7. The end cap 3 is installed at the connector port of the AI ​​computing board and is mainly used to seal the connector when it is not in use, to isolate the connector from the harsh external environment, prevent external particles or contaminants from entering, and suppress external electromagnetic interference.

[0039] Specifically, the end cover 3 is a thin sheet structure integrally formed from metal or plastic, and its shape matches the interface outline of the PCB-AI computing board 7. The end cover 3 is provided with buckles or screw holes around its perimeter for detachably fixing the end cover 3 to the interface opening of the AI ​​computing board frame 1.

[0040] Furthermore, such as Figure 7 as well as Figure 8 As shown, the PCB-AI computing board 7 integrates chips YULONG810A, LTM4644, LTM4630, DDR3, SoC packaging, power modules, and various rectangular connectors of different specifications. This computing board mainly provides computing and control for aerospace missions and is a core functional board integrating AI computing power, system control, data storage, and power management.

[0041] PCB-AI Computing Board 2 (8) primarily works in conjunction with PCB-AI Computing Board 1 (7), serving as the central hub for data exchange and collaborative control in the entire multi-board system. It occupies a critical position in terms of computing, communication, and thermal management. PCB-AI Computing Board 3 (9) is the system's security unit, employing redundancy design to achieve fault takeover, task continuity, and improved system reliability.

[0042] Specifically, PCB-AI computing board 28 has a multilayer printed circuit board structure, with its substrate being FR-4 or polyimide material, and its surface integrated with data exchange chips and high-speed connectors; positioning holes for screws to pass through are provided at the four corners of PCB-AI computing board 28, and its upper and lower surfaces are respectively fixedly connected to AI computing board frame 24 by screws; PCB-AI computing board 28 is electrically connected to PCB-AI computing board 1 7 and PCB-AI computing board 3 9 through board-to-board connectors.

[0043] The system workflow of this invention is as follows: After the aerospace satellite AI computing payload is powered on, PCB-AI computing board 1 (7) serves as the main computing core, responsible for receiving remote sensing data from the satellite payload and executing AI inference computing tasks. PCB-AI computing board 2 (8) serves as the data exchange and collaborative control center, coordinating data flow and control signals between the boards. PCB-AI computing board 3 (9) serves as a redundant backup unit, monitoring the main system's operating status in real time and automatically taking over tasks when the main system fails, ensuring continuous system operation. The three boards are tightly connected to the board frame structure via fastening bolts 10. Heat is conducted through the board frame to the heat dissipation cover 2 and the outer shell, and then to the satellite structure for dissipation. End caps 3 seal and protect unused interfaces, and the bottom cover 6 serves as the mounting base for connection to the satellite platform.

[0044] Working Principle: In use, the aerospace satellite AI computing payload is first installed on the satellite platform via the bottom cover 6. Remote sensing data collected by the satellite payload is transmitted to PCB-AI computing board 1 7 via an interface. PCB-AI computing board 1 7 performs AI inference calculations on the data, such as target detection and image recognition. The processing results are transmitted back to the satellite data transmission system via a data interface. PCB-AI computing board 2 8 coordinates the data flow of each board, and PCB-AI computing board 3 9 serves as a redundant backup, ready for immediate use. Throughout the operation, AI computing board frame 1, AI computing board frame 2 4, AI computing board frame 3 5, and the heat sink 2 form an efficient heat dissipation channel, conducting and dissipating the heat generated by the chip. Fastening bolts 10 ensure that each layer of the structure remains secure under launch vibration and impact. End caps 3 seal and protect unused interfaces. Finally, the on-orbit intelligent computing work of the aerospace satellite AI computing payload is completed.

[0045] 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 invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. An AI computing payload device for aerospace satellites, comprising an AI computing board frame (1), characterized in that: The AI ​​computing board frame 1 (1) is internally fixed with a PCB-AI computing board card 1 (7) by screws. An AI computing board frame 2 (4) is provided on one side of the AI ​​computing board frame 1 (1), and a PCB-AI computing board card 2 (8) is installed inside the AI ​​computing board frame 2 (4) by screws. An AI computing board frame 3 (5) is provided on one side of the AI ​​computing board frame 2 (4), and a PCB-AI computing board card 3 (9) is installed inside the AI ​​computing board frame 3 (5) by screws. A bottom cover (6) is fitted on one side of the AI ​​computing board frame 3 (5). Fastening bolts (10) are provided at the four corners of the surface of the bottom cover (6). The bolt (10) passes through the bottom cover (6) and the third AI computing board frame (5) in sequence and is threadedly fixed to the second AI computing board frame (4); the surface of the first AI computing board frame (1) is covered with a heat dissipation cover (2), which is fixed to the first AI computing board frame (1) by screws. There are locking bolts (11) at the four corners of the surface of the first AI computing board frame (1), which pass through the first AI computing board frame (1) in sequence and are threadedly fixed to the second AI computing board frame (4). The surface of the first AI computing board frame (1) is provided with an end cover (3), which covers the interface of the first PCB-AI computing board card (7).

2. The aerospace satellite AI computing payload device according to claim 1, characterized in that: The AI ​​computing board frame (1) is an integrated frame structure formed by CNC machining of magnesium alloy or aluminum alloy. It is equipped with reinforcing ribs around its perimeter and has threaded mounting posts and positioning countersunk holes inside for fixing the PCB-AI computing board card (7). The upper surface of the AI ​​computing board frame (1) is provided with a flat heat-conducting plane for fitting the heat dissipation cover plate (2). The four corners of the AI ​​computing board frame (1) are provided with through holes for the locking bolts (11) to pass through.

3. The aerospace satellite AI computing payload device according to claim 1, characterized in that: The heat dissipation cover (2) is a flat plate structure formed by CNC machining of magnesium alloy or aluminum alloy. Multiple parallel heat dissipation fins are provided on its outer surface, and heat dissipation channels are formed between adjacent heat dissipation fins. The bottom of the heat dissipation cover (2) is provided with a flat heat-conducting surface that is closely attached to the upper surface of the AI ​​computing board frame (1), and countersunk holes for screws to pass through are provided at the four corners of the heat dissipation cover (2).

4. The aerospace satellite AI computing payload device according to claim 1, characterized in that: The end cap (3) is a thin sheet structure integrally formed of metal or plastic material, and its shape matches the interface outline of the PCB-AI computing board (7). The end cap (3) is provided with buckles or screw holes around its periphery, which are used to detachably fix the end cap (3) to the interface opening of the AI ​​computing board frame (1).

5. The aerospace satellite AI computing payload device according to claim 1, characterized in that: The upper and lower surfaces of the AI ​​computing board frame 2 (4) are provided with boss fitting structures; the boss fitting structure includes multiple upward protruding positioning bosses and downward recessed positioning grooves. The positioning bosses fit into the corresponding grooves at the bottom of the AI ​​computing board frame 1 (1), and the positioning grooves fit into the corresponding bosses at the top of the AI ​​computing board frame 3 (5), so as to realize the precise alignment and positioning of the multi-layer board frame; the four corners of the AI ​​computing board frame 2 (4) are provided with internal threaded holes for fastening bolts to pass through.

6. The aerospace satellite AI computing payload device according to claim 1, characterized in that: The top of the AI ​​computing board frame three (5) is provided with a positioning boss that matches the bottom groove of the AI ​​computing board frame two (4), and the bottom is provided with a flat surface for fitting the bottom cover (6); the four corners of the AI ​​computing board frame three (5) are provided with through holes for fastening bolts (10) to pass through, and the upper end of the through hole is provided with a countersunk hole.

7. The aerospace satellite AI computing payload device according to claim 1, characterized in that: The bottom of the bottom cover (6) is provided with multiple mounting interfaces for fixed connection with the satellite platform. The mounting interfaces are threaded holes or through holes. The upper inner surface of the bottom cover (6) is provided with positioning grooves for accommodating heat-conducting pads. The bottom cover (6) has countersunk holes at its four corners for fastening bolts (10) to pass through. The bottom cover (6) has upward-folded positioning folds around its perimeter.

8. The aerospace satellite AI computing payload device according to claim 1, characterized in that: The PCB-AI computing board (7) is a multilayer printed circuit board structure. Its substrate is FR-4 or polyimide material. AI computing chips, power modules and rectangular connectors of various specifications are welded on its upper and lower surfaces respectively. The PCB-AI computing board (7) has positioning holes for screws to pass through at its four corners, and its edges are provided with positioning notches that cooperate with the inner wall of the AI ​​computing board frame (1).

9. The aerospace satellite AI computing payload device according to claim 1, characterized in that: The PCB-AI computing board card two (8) is a multilayer printed circuit board structure. Its substrate is FR-4 or polyimide material, and its surface is integrated with data exchange chip and high-speed connector. The four corners of the PCB-AI computing board card two (8) are provided with positioning holes for screws to pass through. Its upper and lower surfaces are respectively fixedly connected to the AI ​​computing board frame two (4) by screws. The PCB-AI computing board card two (8) is electrically connected to the PCB-AI computing board card one (7) and the PCB-AI computing board card three (9) through board-to-board connectors.

10. The aerospace satellite AI computing payload device according to claim 1, characterized in that: The PCB-AI computing board card three (9) is a multilayer printed circuit board structure, and its substrate is FR-4 or polyimide material; the four corners of the PCB-AI computing board card three (9) are provided with positioning holes for screws to pass through, and its edge is provided with positioning structure that cooperates with the inner wall of the AI ​​computing board frame three (5); the PCB-AI computing board card three (9) is electrically connected to the PCB-AI computing board card two (8) through a board-to-board connector.

Images