A manifold microchannel-based array metal retroreflector thermal management method

Abstract

This invention discloses a thermal management method for arrayed metal retroreflectors based on manifold microchannel heat dissipation, belonging to the field of thermal management technology for electronic devices. This method integrates the arrayed metal retroreflector with the manifold microchannel heat sink, deploys a distributed temperature-fluidity integrated sensor network, establishes a dynamic thermal-fluid coupling model based on digital twins, implements adaptive flow channel reconfiguration control based on hotspot region directional optimization, implements intelligent operating parameter regulation based on multi-objective optimization, and implements microchannel health status monitoring and early warning based on machine learning. This invention overcomes the shortcomings of existing technologies where the flow channel structure is fixed and cannot be adaptively adjusted. It achieves uniform control of the temperature field of the arrayed metal retroreflector, dynamic optimal balance between heat dissipation efficiency and energy consumption, and online monitoring of the long-term operational health status of the microchannel. It can be widely applied to the thermal management of high-power-density array devices such as phased array radar T / R components and high-power laser arrays.

Description

Technical Field

[0001] This invention belongs to the field of thermal management technology for electronic devices, specifically a thermal management method for array metal retroreflectors based on manifold microchannel heat dissipation. Background Technology

[0002] Manifold microchannel cooling technology introduces a jet effect through a manifold structure and shortens the heat dissipation channel length, reducing flow resistance while ensuring efficient heat dissipation. The topology-optimized flow channel structure features branching and converging characteristics, enabling direct delivery of the cooling medium along different flow paths to the corresponding heat sources, offering advantages such as good temperature uniformity and low pressure drop. Studies show that at a flow rate of 1.27 L / min, the manifold-jet-microchannel composite liquid cooling structure can achieve a heat dissipation capacity of 1221.6 W / cm².

[0003] Existing manifold microchannel cooling technology still has the following shortcomings:

[0004] When an array of metal retroreflectors is in operation, there are significant differences in the heat generation power and heat flux density in different regions—the central region usually bears the highest heat flux density, followed by the edge region. However, the flow channel structure of existing manifold microchannel heat sinks is mostly a fixed design, which cannot be adaptively adjusted according to the actual heat generation state of each unit in the retroreflector array, resulting in uneven temperature distribution and excessively high local hot spot temperatures.

[0005] Existing cooling systems typically operate under constant conditions, controlling parameters such as coolant flow rate, temperature, and pressure, lacking the ability to respond in real time to the dynamic heat load of the retroreflector array. When the retroreflector's operating mode changes or environmental conditions change, the cooling system cannot adjust its operating parameters in a timely manner, resulting in either excessive cooling capacity leading to energy waste or insufficient cooling capacity.

[0006] During long-term operation, air bubbles, particle deposits, or local blockages may occur in the microchannels of the array metal retroreflector, leading to increased flow resistance and reduced heat dissipation performance. Existing systems lack the ability to monitor and provide early warning of the health status of the microchannels online.

[0007] Therefore, the present invention provides a thermal management method for array metal retroreflectors based on manifold microchannel heat dissipation. Summary of the Invention

[0008] The purpose of this invention is to provide a thermal management method for array metal retroreflectors based on manifold microchannel heat dissipation. To solve the above-mentioned problems in the prior art, this invention is achieved through the following technical solution:

[0009] The technical solution adopted by this invention to solve its technical problem is: a thermal management method for an array metal retroreflector based on manifold microchannel heat dissipation, comprising the following steps:

[0010] Step 1: The array metal reflector and the manifold microchannel heat sink are integrally formed by diffusion bonding or brazing to eliminate interfacial thermal resistance. The manifold microchannel heat sink includes a cooling fluid inlet, a liquid distribution manifold layer, a microchannel array layer, a liquid collection manifold layer, and a cooling fluid outlet arranged in sequence. The microchannel array layer is provided with multiple microchannel units arranged in an array, and each microchannel unit corresponds spatially to at least one reflector unit of the array metal reflector. The liquid distribution manifold layer is provided with a hierarchical tree-shaped manifold channel to evenly distribute the cooling fluid from the inlet to each microchannel unit.

[0011] Step 2: Embed distributed temperature sensors on the back or adjacent positions of each retroreflector unit of the array metal retroreflector to collect the temperature data Tij of each retroreflector unit in real time (i=1,2,…,m; j=1,2,…,n, where m and n are the number of rows and columns of the array, respectively); install micro flow sensors and pressure sensors in each branch channel of the liquid inlet distribution manifold layer and each branch channel of the liquid outlet collection manifold layer to collect the inlet flow rate Qin_ij, outlet flow rate Qout_ij, inlet pressure Pin_ij, and outlet pressure Pout_ij of each microchannel unit in real time.

[0012] Step 3: Based on the real-time sensing data obtained in Step 2, and combined with the material thermal properties of the array metal retroreflector, the geometric parameters of the microchannels, and the properties of the cooling medium, a digital twin model of the array metal retroreflector-manifold microchannel heat dissipation system is established. The digital twin model calculates the temperature field distribution T(x,y,z,t), the flow field distribution V(x,y,z,t), and the pressure field distribution P(x,y,z,t) of each retroreflector unit, in real time by solving the three-dimensional conjugate heat transfer control equations. The digital twin model adopts a reduced-order modeling technique to improve the solution speed of the full-order model to the millisecond level, meeting the real-time computing requirements.

[0013] Step 4: Based on the calculation results of the digital twin model in Step 3, identify hotspot regions in the array of metal retroreflectors—defining retroreflector units with temperatures exceeding a preset threshold T_threshold as hotspot units, and defining the spatial distribution of hotspot units in the array as a hotspot distribution map; based on the hotspot distribution map, implement adaptive flow channel reconfiguration control in the following manner:

[0014] (4a) When the hot spot units are distributed in a concentrated manner (the hot spot units are clustered in a certain continuous area of ​​the array), the adjustable throttle valve in the hierarchical tree-shaped manifold channel corresponding to the hot spot area in the liquid inlet distribution manifold layer is controlled to increase the opening degree, thereby increasing the cooling medium distribution ratio in the hot spot area. At the same time, the throttle valve in the non-hot spot area is controlled to decrease the opening degree, thereby reducing the cooling medium distribution ratio.

[0015] (4b) When the hot spot units are discretely distributed (the hot spot units are scattered in multiple discontinuous positions in the array), the micro regulating valve at the inlet of the microchannel unit corresponding to each hot spot unit is controlled to increase the opening independently, thereby increasing the cooling fluid flow rate of the corresponding microchannel unit;

[0016] (4c) When the hot spot units are periodically distributed in the array, the total flow rate of the cooling working fluid is controlled to be pulsated at the same frequency as the hot spot distribution period, and the heat transfer effect is enhanced by pulsating flow to improve the heat dissipation efficiency of the hot spot area.

[0017] Step 5: Based on Step 4, construct a multi-objective optimization function with the dual objectives of "maximizing array temperature uniformity" and "minimizing cooling fluid pump work":

[0018] Where σ(T) is the standard deviation of the temperature of each retroreflector unit in the array metal retroreflector, characterizing the temperature uniformity; W_pump is the power consumption of the cooling working fluid circulation pump; α and β are weighting coefficients, which are dynamically adjusted according to the actual working conditions;

[0019] The above multi-objective optimization function is solved online using a multi-objective particle swarm optimization algorithm or a Bayesian optimization algorithm, and the optimal combination of operating parameters is output, including: total flow rate of cooling medium Q_total, inlet temperature of cooling medium T_in, inlet pressure of cooling medium P_in, opening combination of each adjustable throttle valve and opening combination of each micro-regulating valve;

[0020] The optimal combination of operating parameters is sent to the actuator to achieve closed-loop intelligent control of the heat dissipation system.

[0021] Step Six: Based on the inlet flow rate Qin_ij and outlet flow rate Qout_ij of each microchannel unit collected in real time in Step Two, calculate the flow loss coefficient of each microchannel unit:

[0022] When the flow loss coefficient ξ_ij of a certain microchannel unit exceeds the preset warning threshold ξ_warning, it is determined that the microchannel unit is at risk of blockage or bubble retention, and a warning signal is generated.

[0023] Meanwhile, based on the deviation Δ between the calculation results of the digital twin model in step three and the actual sensing data, a microchannel health status assessment model is constructed; when the deviation Δ continuously exceeds the preset deviation threshold Δ_threshold and the duration exceeds the preset time threshold τ, it is determined that the microchannel heat dissipation performance has irreversibly degraded, and a maintenance reminder signal is issued.

[0024] The beneficial effects of this invention are as follows:

[0025] 1. This invention integrates the arrayed metal retroreflector with the manifold microchannel heat sink, eliminating the interfacial thermal resistance inherent in traditional discrete designs and significantly improving the efficiency of heat transfer from the retroreflector to the cooling medium. Simultaneously, a one-to-one or one-to-many corresponding heat dissipation design ensures that each microchannel unit specifically serves its corresponding retroreflector unit, achieving precise heat dissipation. By deploying a distributed temperature-fluidity integrated sensor network, real-time sensing of the temperature of each unit in the arrayed metal retroreflector and the flow rate and pressure of each microchannel unit are achieved. A dynamic thermal-fluid coupling model based on digital twins is established, enabling real-time calculation and prediction of the temperature field of the arrayed metal retroreflector and the flow field of the manifold microchannels. This provides accurate decision-making basis for adaptive flow channel reconfiguration and intelligent operating parameter control.

[0026] 2. This invention implements adaptive flow channel reconfiguration control based on hotspot region-oriented optimization. According to the different spatial distribution patterns of hotspot units, it adopts differentiated flow channel reconfiguration strategies to achieve precise suppression of local hotspots and uniform control of the temperature field, overcoming the defects of fixed flow channel structure and inability to adaptively adjust in the prior art. By implementing intelligent operating parameter regulation based on multi-objective optimization, a dynamic optimal balance is achieved between "maximizing temperature uniformity" and "minimizing pump power", which not only ensures heat dissipation effect but also reduces energy consumption, realizing efficient and energy-saving operation of the heat dissipation system.

[0027] 3. This invention implements machine learning-driven microchannel health status monitoring and early warning, achieving online monitoring and early warning of abnormal states such as microchannel blockage and bubble retention. This effectively avoids heat dissipation failures caused by microchannel performance degradation, improving the long-term operational reliability of the system. A comprehensive intelligent thermal management closed-loop system of "perception-modeling-decision-execution-early warning" is constructed, realizing intelligent management of the entire process of array metal retroreflector thermal management. It can be widely applied to efficient heat dissipation and precise temperature control of high-power-density array devices such as phased array radar T / R components, high-power laser arrays, and high-performance AI computing chips. Attached Figure Description

[0028] The invention will now be further described with reference to the accompanying drawings.

[0029] Figure 1 This is a flowchart of a thermal management method for an array metal retroreflector based on manifold microchannel heat dissipation according to the present invention.

[0030] Figure 2 This is a system block diagram of Embodiment 2 of the present invention. Detailed Implementation

[0031] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0032] Example 1: As Figure 1 As shown in the embodiment of the present invention, a thermal management method for an array of metal retroreflectors based on manifold microchannel heat dissipation includes the following steps:

[0033] The array metal reflector and the manifold microchannel heat sink are integrally formed. The manifold microchannel heat sink includes a cooling fluid inlet, a liquid distribution manifold layer, a microchannel array layer, a liquid collection manifold layer, and a cooling fluid outlet. The microchannel array layer is provided with multiple microchannel units arranged in an array. Each microchannel unit corresponds in spatial position to at least one reflector unit of the array metal reflector.

[0034] Distributed temperature sensors are embedded on the back of each retroreflector unit of the array metal retroreflector. Miniature flow sensors and pressure sensors are installed in each branch channel of the liquid inlet distribution manifold layer and the liquid outlet collection manifold layer to collect temperature data of each retroreflector unit, inlet flow rate, outlet flow rate, inlet pressure and outlet pressure of each microchannel unit in real time.

[0035] A digital twin model of the array metal retroreflector-manifold microchannel heat dissipation system is established based on real-time acquired sensor data, and the temperature field distribution of each retroreflector unit and the flow field distribution and pressure field distribution within each microchannel unit are calculated in real time.

[0036] Based on the calculation results of the digital twin model, hotspot areas are identified, and adaptive flow channel reconfiguration control is implemented according to the spatial distribution pattern of the hotspot units.

[0037] A multi-objective optimization function with the dual objectives of maximizing array temperature uniformity and minimizing cooling fluid pump work is constructed. An optimization algorithm is used to solve the optimal combination of operating parameters online and the solution is sent to the actuator.

[0038] The flow loss coefficient is calculated based on the inlet and outlet flow of each microchannel unit. When the flow loss coefficient exceeds the warning threshold, a warning signal is generated.

[0039] It should be noted that: firstly, the heat sink and the heating element are bonded together using semiconductor microfabrication technology; secondly, three types of sensors are deployed on the physical entity for high-frequency data acquisition; thirdly, the acquired data is input into a local embedded edge computing unit to run a reduced-order digital twin model; finally, the control signal (PWM wave or analog voltage) output by the model drives the pump and valve, while the monitoring module continuously compares the "model predicted value" with the "sensor measured value" to determine the health status. Addressing the problems of existing technologies where heat sinks can only passively and uniformly dissipate heat, cannot sense internal flow states, and lack system-level self-healing capabilities, a complete closed loop from perception to execution is established. This achieves full automation and intelligence of the array retroreflector thermal management process. The cooling system is no longer isolated hardware, but an intelligent entity with self-perception, self-decision-making, and self-diagnosis capabilities, significantly reducing reliance on manual intervention.

[0040] The liquid inlet distribution manifold layer is provided with a hierarchical tree-shaped manifold flow channel, which includes a primary main flow channel, a secondary branch flow channel, and a tertiary micro-distribution flow channel. The cross-sectional area of ​​each flow channel decreases progressively according to Murray's law to achieve uniform distribution of the cooling working fluid among the microchannel units.

[0041] It's important to note that Murray's Law states that in a bifurcation network, to achieve minimum flow resistance (pump power loss), the cube of the main pipe diameter (in laminar flow) equals the sum of the cubes of the daughter pipe diameters. In manifold design, this ensures that the fluid velocity doesn't change drastically from the inlet to the inlets of each microchannel, resulting in extremely uniform pressure distribution. When designing the photolithographic mask for the inlet manifold, the hydraulic diameters of each flow channel are calculated strictly according to the mathematical formula of Murray's Law (in laminar flow, the cube of the main pipe diameter equals the sum of the cubes of all daughter pipe diameters). For example, if the diameter of the primary main flow channel is D, then the diameter of the symmetrically branched secondary flow channels is precisely set to... The same principle applies to the three-stage flow channel. From a fluid dynamics perspective, this ensures the natural uniformity of the cooling medium distribution in each microchannel unit, significantly reducing local hot spots caused by uneven flow distribution. Furthermore, the low flow resistance characteristic of this structure significantly reduces the energy consumption of the cooling pump.

[0042] Adaptive flow channel reconfiguration control is implemented based on the spatial distribution pattern of hotspot units, specifically including:

[0043] When the hot spot units are centrally distributed, the adjustable throttle valve in the hierarchical tree-shaped manifold channel corresponding to the hot spot area in the liquid inlet distribution manifold layer increases the opening degree and increases the distribution ratio of cooling working fluid in the hot spot area.

[0044] When the hot spot units are discretely distributed, the micro regulating valve at the inlet of the microchannel unit corresponding to each hot spot unit is controlled to independently increase the opening degree, thereby increasing the cooling fluid flow rate of the corresponding microchannel unit.

[0045] When the hot spots are distributed periodically, the total flow rate of the cooling medium is pulsated and adjusted at the same frequency as the hot spot distribution period.

[0046] After the digital twin model outputs the temperature matrix, the control algorithm first performs cluster analysis in image processing (such as the DBSCAN algorithm) to determine the spatial distribution of the overheating units. If the coordinates of the overheating units are continuous and concentrated (clustered as a single cluster), the main control chip sends a large current to the MEMS thermally driven throttle valve on the primary / secondary manifold main channel corresponding to the hot spot area, pushing the valve core upward to increase the opening. If the coordinates of the overheating units are randomly dispersed (multiple clusters), only the piezoelectric ceramic micro-valve at the microchannel inlet corresponding to each hot spot is adjusted individually. If the overheating units are distributed in an equally spaced stripe / checkerboard pattern, the algorithm sends a sinusoidal drive signal to the variable frequency pump, causing the total flow rate to oscillate at this spatial frequency. This solves the problem that a single control method cannot adapt to complex and variable heat source distribution patterns, resulting in "slow temperature control response" or "over-control interfering with normal units." It avoids the pump power waste caused by indiscriminately increasing the total flow rate. It can process quickly and accurately, and can also utilize the resonance effect to improve the heat transfer limit of periodic hot spots.

[0047] The multi-objective optimization function is:

[0048] minF = [α·σ(T) + β·W_pump]

[0049] Where σ(T) is the standard deviation of the temperature of each reflector unit in the array metal reflector, W_pump is the power consumption of the cooling medium circulation pump, and α and β are weighting coefficients dynamically adjusted according to actual operating conditions; the optimization algorithm is a multi-objective particle swarm optimization algorithm or a Bayesian optimization algorithm; the optimal combination of operating parameters includes the total flow rate of the cooling medium, the inlet temperature of the cooling medium, the inlet pressure of the cooling medium, the opening combination of each adjustable throttle valve, and the opening combination of each micro-regulating valve. The control objective is quantified into a mathematical function, achieving a dynamic intelligent balance between heat dissipation performance and operating energy consumption. Under light load conditions, the system automatically reduces the total flow rate to save energy; under heavy load conditions or when uneven heat flow occurs, the system automatically increases input to maintain performance, thereby achieving optimal energy efficiency ratio throughout the entire life cycle of the equipment.

[0050] The digital twin model calculates the temperature field, flow field, and pressure field in real time by solving the three-dimensional conjugate heat transfer control equation. The digital twin model adopts the order reduction modeling technique of intrinsic orthogonal decomposition or dynamic mode decomposition, which improves the solution speed of the full-order model to the millisecond level.

[0051] The flow loss coefficient is calculated using the following formula:

[0052] The flow loss coefficient ξ_ij = (Qin_ij - Qout_ij) / Qin_ij, where Qin_ij is the inlet flow rate of the microchannel unit in the i-th row and j-th column, and Qout_ij is the outlet flow rate of the microchannel unit in the i-th row and j-th column. When ξ_ij exceeds the preset warning threshold ξ_warning, it is determined that the microchannel unit is at risk of blockage or bubble retention, and a warning signal is generated. At the same time, a microchannel health status assessment model is constructed based on the deviation Δ = |T_DT - T_sensor| between the calculation results of the digital twin model and the actual sensor data. When the deviation Δ continuously exceeds the preset deviation threshold Δ_threshold and the duration exceeds the preset time threshold τ, it is determined that the heat dissipation performance of the microchannel has irreversibly degraded, and a maintenance reminder signal is issued.

[0053] The arrayed metal reflector and the manifold microchannel heat sink are integrally formed using diffusion bonding or active brazing processes, without the traditional application of thermal grease or soldering. Instead, the pre-fabricated silicon-based / copper-based microchannel heat sink cover plate is stacked with the metal reflector base plate and placed in a vacuum autoclave. Under high temperature (e.g., 900-1000℃ for copper-copper diffusion bonding) and high pressure (e.g., 5-10 MPa) conditions for several hours, the atoms at the contact surface diffuse to form grain connections. During factory inspection, an infrared thermal imager is used in conjunction with a heat flux density meter to ensure that the equivalent contact thermal resistance of the interface is no greater than 1×10⁻⁶. − 6 m 2 The fracture occurred at ⋅ K / W, and destructive tensile testing showed that the fracture occurred in the base material rather than at the interface. This solves the thermal bottleneck problem in split designs where microscopic air gaps still exist, leading to interfacial temperature differences of 5-15°C, even when filled with the best thermally conductive medium. The interfacial temperature difference is eliminated from the physical structure. Heat is conducted almost unimpeded directly from the reflector junction to the bottom wall of the microchannel.

[0054] Example 2: Figure 2 As shown, this embodiment proposes a thermal management system for an array of metal retroreflectors based on manifold microchannel heat dissipation, used to implement the method described in Embodiment 1 of the present invention, including:

[0055] Integrated metal retroreflector array with manifold microchannel heat dissipation;

[0056] The distributed temperature-influenza integrated sensor network includes distributed temperature sensors embedded on the back of each retroreflector unit of the array metal retroreflector, and miniature flow sensors and pressure sensors set in each branch channel of the inlet distribution manifold layer and the outlet collection manifold layer.

[0057] The data acquisition and preprocessing module is used to acquire the raw signals from each sensor and perform filtering, amplification, and analog-to-digital conversion preprocessing.

[0058] The digital twin computing module is used to build and solve the thermal-fluid coupling digital twin model based on the preprocessed sensor data;

[0059] The hotspot identification and flow channel reconfiguration decision module is used to identify hotspot areas based on the calculation results of the digital twin model and generate adaptive flow channel reconfiguration control commands.

[0060] The multi-objective optimization decision module is used to solve multi-objective optimization functions online and generate optimal operating parameters.

[0061] The actuator drive module is used to convert control commands and operating parameters into drive signals and send them to each adjustable throttle valve, micro regulating valve and cooling working fluid circulation pump;

[0062] The health status monitoring and early warning module is used to calculate the flow loss coefficient and monitor the deviation between the digital twin model and the actual sensor data, and generate early warning signals or maintenance reminder signals when abnormalities occur.

[0063] The inner wall of each microchannel unit of the microchannel array layer is provided with a hydrophobic coating or a hydrophilic coating. The hydrophobic coating or hydrophilic coating is prepared by chemical vapor deposition or atomic layer deposition process, and the coating thickness is 10nm to 100nm. It is used to regulate the wettability and flow state of the cooling working fluid in the microchannel and inhibit bubble nucleation and growth.

[0064] It needs to be explained that after the microchannel etching is completed and before bonding and capping, an atomic layer deposition device is used to uniformly deposit a layer of alumina or polytetrafluoroethylene film with a precise and controllable thickness on the sidewalls and bottom of the microchannel with a very high aspect ratio in the form of precursor gas pulses.

[0065] Hydrophilic treatment: used for water-based working fluids, making water easier to spread, increasing the effective heat exchange area, and delaying the boiling critical point (CHF).

[0066] Hydrophobic treatment: Used for easily vaporized working fluids or applications requiring extremely low pressure drop, it makes it difficult for bubbles to adhere to the wall surface and allows them to be quickly carried away by the water flow, avoiding bubble blockage in dead zones.

[0067] This addresses the fluid instability problem caused by the easy formation of vaporization nuclei on the walls of high-temperature zones inside microchannels, which can lead to air embolisms that obstruct flow and cause localized burnout.

[0068] The cooling medium is any one of deionized water, nanofluid, liquid metal, or phase change microcapsule suspension; when nanofluid is used, the mass fraction of nanoparticles is 0.1% to 5.0%, and the material of the nanoparticles is any one or a combination of alumina, copper oxide, carbon nanotubes, or graphene; when liquid metal is used, the liquid metal is a gallium-based alloy or a bismuth-based alloy; when phase change microcapsule suspension is used, the core material of the phase change microcapsules is paraffin or fatty acid, the wall material is urea-formaldehyde resin or melamine-formaldehyde resin, and the phase change temperature is 40℃ to 80℃.

[0069] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the present invention should still fall within the scope of the present invention.

Claims

1. A method of thermal management of an array of metallic retro-reflectors based on manifold micro-channel heat spreading, characterized by: Includes the following steps: The array metal reflector and the manifold microchannel heat sink are integrally formed. The manifold microchannel heat sink includes a cooling fluid inlet, a liquid distribution manifold layer, a microchannel array layer, a liquid collection manifold layer, and a cooling fluid outlet. The microchannel array layer is provided with multiple microchannel units arranged in an array. Each microchannel unit corresponds in spatial position to at least one reflector unit of the array metal reflector. Distributed temperature sensors are embedded on the back of each retroreflector unit of the array metal retroreflector. Miniature flow sensors and pressure sensors are installed in each branch channel of the liquid inlet distribution manifold layer and the liquid outlet collection manifold layer to collect temperature data of each retroreflector unit, inlet flow rate, outlet flow rate, inlet pressure and outlet pressure of each microchannel unit in real time. A digital twin model of the array metal retroreflector-manifold microchannel heat dissipation system is established based on real-time acquired sensor data, and the temperature field distribution of each retroreflector unit and the flow field distribution and pressure field distribution within each microchannel unit are calculated in real time. Based on the calculation results of the digital twin model, hotspot areas are identified, and adaptive flow channel reconfiguration control is implemented according to the spatial distribution pattern of the hotspot units. A multi-objective optimization function with the dual objectives of maximizing array temperature uniformity and minimizing cooling fluid pump work is constructed. An optimization algorithm is used to solve the optimal combination of operating parameters online and the solution is sent to the actuator. The flow loss coefficient is calculated based on the inlet and outlet flow of each microchannel unit. When the flow loss coefficient exceeds the warning threshold, a warning signal is generated.

2. The manifold microchannel heat sink based array metal retroreflector thermal management method of claim 1, wherein, The liquid inlet distribution manifold layer is provided with a hierarchical tree-shaped manifold flow channel, which includes a primary main flow channel, a secondary branch flow channel, and a tertiary micro-distribution flow channel. The cross-sectional area of ​​each flow channel decreases progressively according to Murray's law to achieve uniform distribution of the cooling working fluid among the microchannel units.

3. The manifold microchannel heat sink based array metal retroreflector thermal management method of claim 1, wherein, Adaptive flow channel reconfiguration control is implemented based on the spatial distribution pattern of hotspot units, specifically including: When the hot spot units are centrally distributed, the adjustable throttle valve in the hierarchical tree-shaped manifold channel corresponding to the hot spot area in the liquid inlet distribution manifold layer increases the opening degree and increases the distribution ratio of cooling working fluid in the hot spot area. When the hot spot units are discretely distributed, the micro regulating valve at the inlet of the microchannel unit corresponding to each hot spot unit is controlled to independently increase the opening degree, thereby increasing the cooling fluid flow rate of the corresponding microchannel unit. When the hot spots are distributed periodically, the total flow rate of the cooling medium is pulsated and adjusted at the same frequency as the hot spot distribution period.

4. The manifold microchannel heat sink based array metal retroreflector thermal management method of claim 1, wherein, The multi-objective optimization function is: minF = [α·σ(T) + β·W_pump] Where σ(T) is the standard deviation of the temperature of each retroreflector unit in the array metal retroreflector, W_pump is the power consumption of the cooling medium circulation pump, and α and β are weighting coefficients dynamically adjusted according to the actual operating conditions; the optimization algorithm is a multi-objective particle swarm optimization algorithm or a Bayesian optimization algorithm; the optimal combination of operating parameters includes the total flow rate of the cooling medium, the inlet temperature of the cooling medium, the inlet pressure of the cooling medium, the opening combination of each adjustable throttle valve, and the opening combination of each micro-regulating valve.

5. The manifold microchannel heat sink based array metal retroreflector thermal management method of claim 1, wherein, The digital twin model calculates the temperature field, flow field, and pressure field in real time by solving the three-dimensional conjugate heat transfer control equations.

6. The manifold microchannel heat sink based array metal retroreflector thermal management method of claim 1, wherein, The flow loss coefficient is calculated using the following formula: The flow loss coefficient ξ_ij = (Qin_ij - Qout_ij) / Qin_ij, where Qin_ij is the inlet flow rate of the microchannel unit in the i-th row and j-th column, and Qout_ij is the outlet flow rate of the microchannel unit in the i-th row and j-th column. When ξ_ij exceeds the preset warning threshold ξ_warning, it is determined that the microchannel unit is at risk of blockage or bubble retention, and a warning signal is generated. At the same time, a microchannel health status assessment model is constructed based on the deviation Δ between the calculation results of the digital twin model and the actual sensing data. When the deviation Δ continuously exceeds the preset deviation threshold Δ_threshold and the duration exceeds the preset time threshold τ, it is determined that the heat dissipation performance of the microchannel has irreversibly degraded, and a maintenance reminder signal is issued.

7. The manifold microchannel heat sink based array metal retroreflector thermal management method of claim 1, wherein, The array metal retroreflector and the manifold microchannel heat sink are integrally formed by diffusion bonding or active brazing.

8. An array of metal retro-reflectors heat management system based on manifold microchannel heat spreading for implementing the method of any one of claims 1-7, characterized in that, include: Integrated metal retroreflector array with manifold microchannel heat dissipation; The distributed temperature-influenza integrated sensor network includes distributed temperature sensors embedded on the back of each retroreflector unit of the array metal retroreflector, and miniature flow sensors and pressure sensors set in each branch channel of the inlet distribution manifold layer and the outlet collection manifold layer. The data acquisition and preprocessing module is used to acquire the raw signals from each sensor and perform filtering, amplification, and analog-to-digital conversion preprocessing. The digital twin computing module is used to build and solve the thermal-fluid coupling digital twin model based on the preprocessed sensor data; The hotspot identification and flow channel reconfiguration decision module is used to identify hotspot areas based on the calculation results of the digital twin model and generate adaptive flow channel reconfiguration control commands. The multi-objective optimization decision module is used to solve multi-objective optimization functions online and generate optimal operating parameters. The actuator drive module is used to convert control commands and operating parameters into drive signals and send them to each adjustable throttle valve, micro regulating valve and cooling working fluid circulation pump; The health status monitoring and early warning module is used to calculate the flow loss coefficient and monitor the deviation between the digital twin model and the actual sensor data, and generate early warning signals or maintenance reminder signals when abnormalities occur.

9. The manifold microchannel heat sink based array metal retroreflector thermal management system of claim 8, wherein, The inner wall of each microchannel unit in the microchannel array layer is provided with a hydrophobic coating or a hydrophilic coating. The hydrophobic coating or hydrophilic coating is prepared by chemical vapor deposition or atomic layer deposition process and is used to regulate the wettability and flow state of the cooling working fluid in the microchannel and inhibit bubble nucleation and growth.

10. The manifold microchannel heat sink based array metal retroreflector thermal management system of claim 8, wherein, The cooling medium is any one of deionized water, nanofluid, liquid metal, or phase change microcapsule suspension; when nanofluid is used, the mass fraction of nanoparticles is 0.1% to 5.0%, and the material of the nanoparticles is any one or a combination of alumina, copper oxide, carbon nanotubes, or graphene; when liquid metal is used, the liquid metal is a gallium-based alloy or a bismuth-based alloy; when phase change microcapsule suspension is used, the core material of the phase change microcapsules is paraffin or fatty acid, and the wall material is urea-formaldehyde resin or melamine-formaldehyde resin.

Images