Notebook computer waste heat recycling system and method based on thermoelectric conversion effect

Abstract

The invention discloses a notebook computer waste heat recycling system and method based on a thermoelectric conversion effect, and relates to the technical field of thermoelectric materials and devices, and the system comprises a distributed thermoelectric array module which is composed of a plurality of miniature TEG units, is distributed in a high heating area of a notebook computer, and directly generates power through temperature difference; the self-adaptive heat management module comprises a heat-conducting silica gel layer and micro heat dissipation fins and is used for optimizing the temperature difference between the hot end and the cold end; the multi-stage electric energy management module comprises a rectifying circuit, a DC-DC boosting chip and a recycling energy storage battery device, converts the unstable low-voltage direct current output by the TEG into stable electric energy, stores the stable electric energy and supplies the stable electric energy to target equipment; and the intelligent control module dynamically adjusts the working area of the micro thermoelectric power generation unit according to the data of the temperature sensor, preferentially selects the position with the maximum temperature difference to generate power, and matches the power supply demand through a PMIC chip. Efficient conversion and utilization of waste heat are realized through innovative design, the problem of low thermoelectric conversion efficiency in miniaturized equipment is solved, and energy conservation and environmental protection are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of thermoelectric materials and devices, and particularly to a waste heat recovery and utilization system and method for laptop computers based on the thermoelectric conversion effect. Background Art

[0002] With the continuous improvement of the performance of laptop computers, the problem of their thermal power consumption has become increasingly prominent: the TDP (thermal design power) of modern CPUs / GPUs continues to grow, and that of high-performance laptops can reach more than 45W. At present, about 60% - 70% of the electrical energy of laptop computers is finally converted into waste heat and dissipated into the environment. Traditional cooling methods (heat pipes + fans) only focus on heat discharge and do not consider energy recovery. In addition, the number of electronic devices globally has increased rapidly, and energy recovery and utilization has become an important topic for sustainable development. At present, the progress of battery technology is slow, and the battery life can be extended by means of waste heat recovery. Due to its characteristics such as no moving parts in the solid state, high reliability, and suitability for small temperature difference applications, the thermoelectric conversion technology has become an ideal choice for laptop waste heat recovery.

[0003] The problems faced by the current laptop waste heat recovery and utilization technology based on the thermoelectric conversion effect include: during the operation of laptop computers, components such as CPUs, GPUs, and power modules will generate a large amount of waste heat, and the thermal energy is not effectively utilized, which aggravates greenhouse gas emissions and thermal pollution, further increasing the environmental burden and wasting resources; the long-term use of low-power modules such as built-in camera fill lights and sensors in laptop computers will affect the battery life. Summary of the Invention

[0004] To solve the above technical problems, a waste heat recovery and utilization system and method for laptop computers based on the thermoelectric conversion effect are provided. The technical solution solves the problems that during the operation of the above laptop computers, components such as CPUs, GPUs, and power modules will generate a large amount of waste heat and the waste heat is not effectively utilized; the low-power modules such as built-in camera fill lights and sensors in laptop computers need to rely on battery power supply, and the long-term use may affect the battery life.

[0005] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] A waste heat recovery and utilization system for laptop computers based on the thermoelectric conversion effect, comprising:

[0007] A distributed thermoelectric array module: composed of multiple micro-TEG units, distributed in the high-heat generation areas of the laptop, including the CPU heat dissipation module substrate, the middle bending area of the heat pipe, and the air outlet guide fins, directly generating electricity through the temperature difference and supplying electric energy to the device function module;

[0008] Adaptive Thermal Management Module: It includes a thermal conductive silicone layer and micro heat dissipation fins to optimize the temperature difference between the hot end and the cold end; the hot end forms a directional heat channel with the CPU top cover through laser micro-welding, and the fin spacing at the cold end is designed with a gradient;

[0009] Multi-level Electric Energy Management Module: It includes a rectifier circuit, a DC-DC boost chip, and a recycled energy storage battery device to convert the unstable low-voltage direct current output by the TEG into stable electric energy, store it, and supply it to the target device;

[0010] Intelligent Control Module: Dynamically adjust the working area of the micro thermoelectric power generation unit according to the temperature sensor data, preferentially select the position with the largest temperature difference for power generation, and match the power supply demand through the PMIC chip.

[0011] Preferably, the distributed thermoelectric array module specifically includes:

[0012] Dynamic Thermal Impedance Matching Unit: A micro heat flux sensor is integrated at the bottom of each TEG unit to monitor the heat flux distribution in real time; the inclination of the unit is dynamically adjusted through a piezoelectric ceramic actuator to align the heat flow direction with the optimal orientation of the thermoelectric lattice;

[0013] Three-dimensional Interconnected Power Supply Unit: Laser-induced graphene wiring is used to directly form a circuit on the surface of the heat dissipation module. The TEG units in each area achieve series-parallel adaptive recombination through a topology optimization algorithm to maintain the stability of the output voltage within a given temperature difference range;

[0014] Failure Fault Tolerance Mechanism Unit: A TEG unit health assessment model is built in, and the remaining life is predicted through the decay rate of the Seebeck coefficient; when the efficiency of a certain unit drops to a given threshold, the system automatically switches it to a pure heat conduction channel, and the total power output is maintained through the current redistribution of adjacent units.

[0015] Preferably, the distributed thermoelectric array module specifically includes:

[0016] CPU Heat Dissipation Module Substrate Area: Multiple micro TEG units are deployed in an asymmetric star-shaped arrangement, and the center point is offset from the geometric center of the CPU top cover to match the actual hot spot distribution; a thick boron nitride thermal conductive insulating film is filled between the TEG units to achieve electrical isolation while maintaining the lateral thermal conductivity;

[0017] Heat Pipe Middle Bending Area: Multiple flexible TEG units are deployed, and the substrate uses a copper-graphene composite foil to fit the curved shape of the heat pipe; a micro shape memory alloy spring is integrated to automatically increase the contact pressure when the temperature of the heat pipe exceeds a given temperature;

[0018] Air Outlet Deflector Fin Area: Multiple air flow-driven TEG units are deployed, and a micron-level turbulent structure is processed on the fin surface to reduce the cold end temperature through the convection of the exhaust air flow.

[0019] Preferably, the adaptive thermal management module specifically includes:

[0020] Hot-end conduction unit: The thermal conductive silicone layer uses graphene / boron nitride hybrid fillers and adds phase change microcapsules, and the critical temperature of the phase change microcapsules is defined.

[0021] Cold-end heat dissipation unit: The micro heat dissipation fins are designed with gradient arrangement according to the near heat source area, far heat source area, and transition area, and the transition area adopts a bionic spiral gradient design.

[0022] Heat flux sensor unit: Through an embedded micro heat flux sensor, the heat flux distribution is monitored in real time. Combined with a shape memory alloy actuator, the fin inclination angle is dynamically adjusted. Computational fluid dynamics simulation is used to guide the optimization of the fin inclination angle, so that the heat dissipation efficiency is automatically optimized with the change of the load. Phonon engineering is introduced to optimize the silicone filler network and optimize the intermediate frequency phonon transmission efficiency.

[0023] Preferably, the multi-stage power management module specifically includes:

[0024] Front-end signal processing unit: A wide input range rectifier circuit uses zero-threshold voltage GaN diodes and integrates self-biased synchronous rectification technology. A dynamic impedance matching network uses a digitally tunable LC resonant circuit and tracks the change of the TEG internal resistance in real time through an impedance analyzer.

[0025] Energy conversion core unit: A DC-DC boost chip based on magnetic coupling resonance wireless energy transfer realizes adaptive multi-topology switching. The supercapacitor buffer uses graphene / CNT composite electrode materials to increase the pulse charge and discharge tolerance.

[0026] Back-end energy storage and distribution unit: The hybrid energy storage device includes short-term energy storage using 3D silicon anode lithium batteries, instantaneous buffering using solid-state micro-supercapacitors, and an intelligent power distribution network that manages the priorities of the CPU, USB, and battery charging weights and has a wireless backcharging function.

[0027] Preferably, the intelligent control module specifically includes:

[0028] Multi-modal sensing network unit: The distributed temperature sensing array uses a hybrid network of MEMS infrared thermopiles and fiber Bragg grating sensors to realize the three-dimensional thermal field reconstruction of the CPU / GPU surface. An integrated zero-drift operational amplifier is used to collect the output characteristics of each TEG unit in real time.

[0029] Intelligent decision-making core unit: Adopts edge computing, runs a thermal-electric coupling model, customizes energy efficiency optimization instructions based on the RISC-V instruction set, an adaptive PMIC integrates a digital decoupling algorithm to eliminate crosstalk between channels, and uses federated learning to realize cross-device fault prediction.

[0030] Actuator network unit: Through the piezoelectric micro-motion platform, the contact pressure of the TEG unit is dynamically adjusted, and the bionic muscle structure is adopted; based on GaN power devices, the dynamic reorganization of TEG units in series and parallel is realized, and lossless current commutation is supported;

[0031] Key control strategy unit: Use the temperature difference priority power generation algorithm combined with multi-objective optimization control. Its constraints include reliability constraints, efficiency constraints, and energy consumption constraints. The NSGA-Ⅲ algorithm is used for Pareto frontier optimization.

[0032] Furthermore, a method for recovering waste heat from a laptop computer based on a thermoelectric conversion effect is provided to realize a system for recovering waste heat from a laptop computer based on a thermoelectric conversion effect as described above, comprising:

[0033] The infrared thermal imaging sensor array scans the three-dimensional temperature distribution of the CPU, GPU and heat pipe surface in real time, identifies the efficient recovery area of the temperature difference based on the gradient descent algorithm, and preferentially activates the micro TEG unit array at the corresponding position;

[0034] Liquid metal thermal interface material is used to dynamically adjust the heat flow path, and the flow rate of the nanofluid is controlled by an electromagnetic pump to stabilize the temperature of the TEG hot end within a predetermined temperature range; the inclination angle of the cold end micro-heat sink fins and the fan PWM duty cycle are adjusted synchronously to maintain the cold end temperature below the predetermined temperature;

[0035] The millivolt voltage output by the TEG is processed in three stages, including charge pump pre-boost, resonant DC-DC conversion, and frequency-band charging and discharging of the hybrid energy storage system;

[0036] The LSTM neural network is used to predict the CPU power consumption trend in the short term and dynamically adjust the TEG unit series-parallel topology.

[0037] When it is detected that the efficiency attenuation of a single TEG unit exceeds a set ratio, it is automatically switched to a pure heat conduction channel and the power supply path is redistributed.

[0038] Preferably, the three-dimensional temperature distribution of the CPU, GPU and heat pipe surface is scanned in real time by an infrared thermal imaging sensor array, and the efficient recovery area of the temperature difference is identified based on a gradient descent algorithm, and the micro TEG unit array at the corresponding position is preferentially activated, specifically including:

[0039] A hybrid network of quantum dot infrared detectors and MEMS thermopiles is used to form a hexagonal monitoring grid with a predetermined spacing on the CPU top cover and the bend of the heat pipe;

[0040] Calculate nanoscale heat flux density on silicon chip surface hot spots, evaluate TEG installation quality on metal interface contact thermal resistance, and detect interface material aging on voids inside thermal paste layers;

[0041] Filter the vibration noise of the fan based on wavelet transform, reconstruct the three-dimensional temperature field by using moving least squares interpolation, and calculate the heat flux vector by combining Fourier's law and the Navier-Stokes equation;

[0042] Based on the selective area algorithm of adaptive gradient descent, calculate the temperature gradient, locate the gradient extreme area, filter through the temperature difference threshold, and output the optimal activation sequence; calculate the dynamic load balancing strategy through the multi-objective optimization function.

[0043] Optionally, the dynamic adjustment of the heat flow path by using the liquid metal thermal interface material, controlling the flow rate of the nanofluid through the electromagnetic pump, and stabilizing the hot end temperature of the TEG in a predetermined temperature range; synchronously adjusting the inclination angle of the cold end micro heat sink fins and the PWM duty cycle of the fan to maintain the cold end temperature below the predetermined temperature specifically includes:

[0044] The liquid metal thermal interface material includes a base alloy doped with diamond nanoparticles, a magnetic additive modified with a surface silane coupling agent, and a packaging layer with a microgroove capillary structure;

[0045] The drive of the electromagnetic pump adopts a combination of a Halbach array permanent magnet and a micro solenoid;

[0046] The inclination angle of the micro heat sink fins is driven by a shape memory alloy wire, and the intelligent speed regulation of the fan is completed through a multi-parameter coupling control model.

[0047] Optionally, the millivolt-level voltage output by the TEG is processed in three stages, including charge pump pre-boosting, resonant DC-DC conversion, and segmented charging and discharging of the hybrid energy storage system; predicting the change trend of the CPU power consumption in the short term in the future through the LSTM neural network, and dynamically adjusting the series-parallel topology of the TEG unit specifically includes:

[0048] The charge pump pre-boosting stage uses zero-threshold voltage MOSFETs to construct a cross-coupled charge pump and integrates self-oscillation control;

[0049] For the resonant DC-DC conversion, an ultra-thin planar transformer and GaN switching tubes are used, and dual-mode control of variable frequency - variable duty cycle is adopted;

[0050] The LSTM neural network prediction is through a power consumption prediction model, and the time domain features of this model include CPU occupancy rate, core temperature, and instruction throughput, and the frequency domain features include the energy distribution of 0.1 - 100 Hz decomposed by FFT;

[0051] The dynamic reconstruction of the TEG topology is based on a GaN matrix switch, and the topology modes include: high temperature difference steady state with full parallel connection; transient load impact with 3 series 2 parallel connection; low temperature environment with full series connection.

[0052] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0053] The present invention proposes to directly generate electricity by using the waste heat inside a laptop through a distributed thermoelectric array module, significantly improving the energy utilization efficiency and reducing energy waste; by effectively recovering and utilizing waste heat, it helps to reduce thermal pollution and carbon emissions, enhancing the environmental protection performance of the product; the adaptive thermal management module optimizes the temperature difference between the hot end and the cold end, improving the thermoelectric conversion efficiency, thus reducing the burden on the traditional cooling system and enhancing the overall performance and operation stability of the laptop; by reducing the thermal resistance and optimizing the heat dissipation design, it helps to reduce the damage of electronic components caused by overheating, thereby extending the service life of the product; the ingenious integration design of the thermoelectric conversion module and the radiator not only improves the thermoelectric conversion efficiency but also enhances the stability and durability of the system; the ultra-thin design of the thermoelectric module makes it suitable for integration into ultra-thin laptops, meeting the requirements of modern electronic products for thin and light; the intelligent control module dynamically adjusts the working area according to real-time temperature data, preferentially selects the position with the largest temperature difference for power generation, and matches the power supply demand through the PMIC chip, achieving the optimal utilization of energy; the integrated and compact design of the present invention has a wide application market, not only applicable to laptops but also extendable to other electronic devices that require efficient heat dissipation and energy recovery. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 It is an internal framework diagram of a laptop waste heat recovery and utilization system based on the thermoelectric conversion effect;

[0055] Figure 2 It is an internal framework diagram of the distributed thermoelectric array module;

[0056] Figure 3 It is an internal framework diagram of the adaptive thermal management module;

[0057] Figure 4 It is an internal framework diagram of the intelligent control module;

[0058] Figure 5 It is a flowchart of a method for recovering and utilizing laptop waste heat based on the thermoelectric conversion effect;

[0059] Figure 6 It is a laptop system structure diagram;

[0060] Figure 7 It is a heat dissipation module structure diagram;

[0061] Figure 8 It is a schematic diagram of the implementation principle;

[0062] Figure 9 It is a schematic diagram of the working principle;

[0063] Figure 10 It is a circuit diagram of the LED boost driver for camera fill light power supply;

[0064] Figure 11 A brightness detection circuit diagram for powering the camera fill light. Specific implementation manner

[0065] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments in the following description are only examples, and those skilled in the art can think of other obvious variations.

[0066] Refer to Figure 1 As shown, a waste heat recovery and utilization system for a laptop computer based on the thermoelectric conversion effect includes:

[0067] Distributed thermoelectric array module: Composed of multiple micro-TEG units, distributed in the high-heat generation areas of the laptop, including the CPU heat dissipation module substrate, the middle bending area of the heat pipe, and the air outlet guide fins, directly generating electricity through the temperature difference and supplying electrical energy to the device function module;

[0068] Adaptive thermal management module: Includes a thermal conductive silicone layer and micro heat dissipation fins to optimize the temperature difference between the hot end and the cold end; the hot end forms a directional heat channel with the CPU top cover through laser micro-welding, and the fin spacing at the cold end is designed with a gradient;

[0069] Multi-stage power management module: Includes a rectifier circuit, a DC-DC boost chip, and a recycled energy storage battery device, converting the unstable low-voltage direct current output by the TEG into stable electrical energy, storing it and supplying it to the target device;

[0070] Intelligent control module: Dynamically adjusts the working area of the micro-thermoelectric power generation unit according to the temperature sensor data, preferentially selects the position with the largest temperature difference for power generation, and matches the power supply demand through the PMIC chip.

[0071] It should be noted that for the thermo-electro-mechanical co-optimization, through the space-temperature mapping relationship between the liquid metal thermal interface and the micro-TEG array, the maximum hot-end temperature gradient is achieved, and the mechanical stress distribution is made uniform; the electromagnetic-thermal joint simulation is optimized using the COMSOL multi-physics model, where the angle θ between the thermoelectric unit spacing and the heat flow direction is controlled within 15 ± 5°;

[0072] The improvement of the energy conversion chain efficiency is achieved through the control of the full-link loss, including the use of laser micro-welding + nano-silver seed layer optimized thermal interface conduction; the use of gradient ZT materials optimized thermoelectric conversion; the use of resonant DC-DC + GaN matrix switch optimized power management.

[0073] Refer to Figure 2 As shown, the distributed thermoelectric array module specifically includes:

[0074] Dynamic Thermal Impedance Matching Unit: A micro thermal flux sensor is integrated at the bottom of each TEG unit to monitor the thermal flux distribution in real time; the inclination angle of the unit is dynamically adjusted by a piezoelectric ceramic actuator to align the heat flow direction with the optimal orientation of the thermoelectric lattice;

[0075] Three-dimensional Interconnected Power Supply Unit: Laser-induced graphene wiring is used to directly form a circuit on the surface of the heat dissipation module. The TEG units in each area are adaptively reorganized in series-parallel through a topology optimization algorithm to maintain the stability of the output voltage within a given temperature difference range;

[0076] Failure Tolerance Mechanism Unit: A health assessment model of the TEG unit is built-in, and the remaining life is predicted through the decay rate of the Seebeck coefficient; when the efficiency of a certain unit drops to a given threshold, the system automatically switches it to a pure heat conduction channel, and the total power output is maintained through the current redistribution of adjacent units.

[0077] CPU Heat Dissipation Module Substrate Area: Multiple micro TEG units are deployed in an asymmetric star-shaped arrangement, and the center point is offset from the geometric center of the CPU top cover to match the actual hot spot distribution; a thick boron nitride thermal conductive insulating film is filled between the TEG units to achieve electrical isolation while maintaining the lateral thermal conductivity;

[0078] Bent Area in the Middle of the Heat Pipe: Multiple flexible TEG units are deployed, and the substrate uses a copper-graphene composite foil to fit the curved shape of the heat pipe; a micro shape memory alloy spring is integrated to automatically increase the contact pressure when the temperature of the heat pipe exceeds a given temperature;

[0079] Air Outlet Flow Guide Fin Area: Multiple air flow-driven TEG units are deployed, and a micron-level turbulent structure is processed on the fin surface to reduce the cold end temperature by using the exhaust air flow convection.

[0080] It should be noted that the dynamic thermal impedance matching unit includes:

[0081] The micro thermal flux sensor uses a hybrid sensing of MEMS thin film thermopile and fiber Bragg grating, and its thermal flux measurement range is 1 - 500 W / cm 2 ;

[0082] The piezoelectric ceramic actuator uses a hinge-type amplification mechanism (5 times stroke) to improve the displacement accuracy to ±50 nm, uses a multi-layer co-fired ceramic structure to make the maximum thrust reach 8 N, and reduces the power consumption to 5 mW / axis through a self-powered energy recovery design;

[0083] The lattice orientation optimization algorithm, based on the pre-computed database of density functional theory, calculates the optimal inclination angle according to the heat flow vector and drives the piezoelectric ceramic to rotate.

[0084] The three-dimensional interconnected power supply unit includes:

[0085] The process parameters of the laser-induced graphene wiring, its laser wavelength is 355 nm (ultraviolet), and the power density is 105 W / cm 2 The line width / spacing is 30μm / 50μm; the sheet resistance of the electrical conductivity is <0.1Ω / sq, and the bending tolerance is >1000 times (radius of curvature 3mm);

[0086] The topology optimization algorithm includes: maximizing the output current (>500mA) through full parallel connection, ΔT > 25°C; boosting the voltage to 3.3V ± 5% through 3 series and 2 parallel connections, 5°C < ΔT < 15°C; suppressing the output voltage ripple and transient load through dynamic switching (<100ns).

[0087] The health management of the TEG unit in the failure tolerance mechanism unit, online detection of the Seebeck coefficient decay rate (ΔS / S0), prediction of the remaining life (error < 5%);

[0088] Health assessment model:

[0089] In the formula, Remaining-Life is the remaining life; S(t) is the health state at time t; S0 is the initial health state, usually at time t = 0; α is a constant representing the influence coefficient of temperature on the health state; T(τ) is the temperature at time τ; T ref is the reference temperature, usually a standard or reference temperature; α = 0.025 (aging coefficient of Bi2Te3), T_ref = 80°C.

[0090] The current redistribution strategy includes: for failed TEG units, through current compensation by adjacent units, then increasing the drive voltage of parallel units and adjusting the DC-DC duty cycle to maintain the total power output;

[0091] For extreme cases, use phase change material fusing protection + liquid metal shunting to cope with instantaneous 100°C thermal shock, and use PTC film preheating + series topology boost to cope with -20°C low-temperature startup.

[0092] In the CPU heat dissipation module substrate area, the non-symmetric star-shaped arrangement is offset 3mm from the CPU hot spot towards the center point, and the radial branch angle is 22.5° intervals; the thickness of the boron nitride insulating film is 0.05mm, and the thermal conductivity is 15W / mK (in-plane), 5W / mK (normal).

[0093] In the middle bending area of the heat pipe, the flexible TEG unit uses copper-graphene composite foil, and its bending fatigue life is >10 5 times (R = 25mm), in-plane thermal conductivity: 650W / mK; the shape memory alloy spring has a trigger temperature of 60°C (NiTiNOL-Cu variant), and the pressure gradient is 0.5MPa / °C (in the 60 - 90°C range).

[0094] The air outlet flow guiding fin area, and the turbulent flow structure design includes: the groove depth is 50 μm at the near-TEG end and 80 μm at the far-TEG end; the surface roughness is Ra = 1.6 μm at the near-TEG end and Ra = 3.2 μm at the far-TEG end; the air flow acceleration ratio is 1.8x at the near-TEG end and 1.2x at the far-TEG end.

[0095] Referring to Figure 3 as shown, the adaptive thermal management module specifically includes:

[0096] Hot end conduction unit: The thermal conductive silicone layer uses graphene / h-BN hybrid fillers and adds phase change microcapsules, and the critical temperature of the phase change microcapsules is defined;

[0097] Cold end heat dissipation unit: The micro heat dissipation fins are designed with gradient arrangement according to the near heat source area, far heat source area, and transition area, and the transition area adopts a bionic spiral gradient design;

[0098] Heat flux sensor unit: Through an embedded micro heat flux sensor, the heat flux distribution is monitored in real time, combined with a shape memory alloy actuator, the fin inclination angle is dynamically adjusted, and computational fluid dynamics simulation is used to guide the optimization of the fin inclination angle, so that the heat dissipation efficiency is automatically optimized with the change of the load; phonon engineering is introduced to optimize the silicone filler network and optimize the intermediate frequency phonon transmission efficiency.

[0099] It should be noted that the material ratio and properties of the graphene / h-BN hybrid filler in the hot end conduction unit include: the content of graphene nanosheets is 25 wt%, which improves the in-plane thermal conductivity (≥180 W / mK); the content of hexagonal boron nitride (h-BN) is 15 wt%, which enhances the electrical insulation (breakdown voltage > 5 kV / mm); the content of the silicone matrix (PDMS) is 60 wt%, which provides flexibility and interface adhesion;

[0100] The core material of the phase change microcapsule is paraffin (C22H46) and a silica shell layer (thickness 100 ± 10 nm), and its critical temperature is 80 ± 2 °C (matching the CPU TJmax threshold), which can absorb 15 J / g of transient thermal shock, and the measured peak hot spot temperature is reduced by 8 - 12 °C.

[0101] The structural parameters of the gradient fins in the cold end heat dissipation unit are: the fin density in the near heat source area is 120 pieces, the height is 8 mm, and the surface is treated by laser micro-texturing (Ra = 1.6 μm); the fin density in the transition area is 90 pieces, the height is 6.5 mm, and the surface is treated by bionic spiral gradient (pitch 2 mm); the fin density in the far heat source area is 60 pieces, the height is 5 mm, and the surface is treated by a hydrophobic nano-coating (θ = 152°);

[0102] The bionic spiral gradient design is inspired by the aerodynamic structure of eagle feathers. Its spiral angle varies continuously from 18° to 25°, the air flow acceleration ratio is 1.8x at the proximal end → 1.2x at the distal end, and the wind resistance optimization effect is 35% lower than that of traditional straight fins.

[0103] In the heat flux sensor unit, the micro heat flux sensor uses a MEMS thin film type (Pt1000 temperature measurement element), with a measurement range of 1 - 500 W / cm 2 , a response time < 1 ms, and 4 sensing nodes are deployed per square centimeter;

[0104] In the shape memory alloy actuator, the phase change temperature of the driving material is adjustable (55 - 90 °C), containing NiTiNOL - Cu (Cu added 3 wt%); the strain recovery rate of displacement output > 99.5%, with an inclination adjustment of 0 - 15°; the response speed is driven by pulsed current (5 A / pulse), 50 ms (heating) / 80 ms (cooling);

[0105] Control equation:

[0106] Where PWM% is the percentage of pulse width modulation, representing the magnitude of the control output; K p is the proportional gain, representing the intensity of proportional control; ΔT is the temperature deviation, that is, the difference between the set value and the actual value; K i is the integral gain, representing the intensity of integral control; ∫ΔTdt is the integral term, representing the accumulation of temperature deviation over time; K d is the derivative gain, representing the intensity of derivative control; is the derivative term, representing the rate of change of temperature deviation; the coefficients are tuned by the Ziegler - Nichols method: K p = 0.8, K i = 0.05, K d = 0.3; when the SMA temperature > 120 °C is detected, power is cut off forcibly.

[0107] Computational fluid dynamics simulation optimization, multi - scale modeling, the macroscopic model solves the RANS equation (k - ωSST turbulence model), and the boundary condition is a wind speed of 2 ± 0.5 m / s (corresponding to the notebook fan working condition); the microscopic model uses LBM (lattice Boltzmann method) to analyze the boundary layer on the fin surface, and the grid size is the first - layer grid near the wall y+ ≈ 1.

[0108] Optimization of phonon transmission paths, constructing an h - BN "thermal bridge" network, arranging by electric - field induction (field strength 1 kV / mm), and using an amino - silane coupling agent to enhance the interfacial phonon coupling, with the transmission efficiency of intermediate - frequency phonons (1 - 10 THz) increased by 60%.

[0109] Refer to Figure 4As shown, the intelligent control module specifically includes:

[0110] Multimodal sensing network unit: The distributed temperature sensing array uses a hybrid network of MEMS infrared thermopiles and fiber Bragg grating sensors to realize the three-dimensional thermal field reconstruction of the CPU / GPU surface; an integrated zero-drift operational amplifier is used to collect the output characteristics of each TEG unit in real time;

[0111] Intelligent decision-making core unit: Edge computing is adopted to run the thermal-electric coupling model, and energy efficiency optimization instructions are customized based on the RISC-V instruction set; an adaptive PMIC integrates a digital decoupling algorithm to eliminate crosstalk between channels; federated learning is used to realize cross-device fault prediction;

[0112] Actuator network unit: Through a piezoelectric micro-motion platform, the contact pressure of the TEG unit is dynamically adjusted, and a bionic muscle structure is adopted; based on GaN power devices, dynamic series-parallel recombination of the TEG unit is realized, and lossless current commutation is supported;

[0113] Key control strategy unit: The temperature difference priority power generation algorithm is used, combined with multi-objective optimization control. Its constraint conditions include reliability constraint, efficiency constraint, and energy consumption constraint, and the NSGA-Ⅲ algorithm is used for Pareto front optimization.

[0114] It should be noted that for the three-dimensional thermal field reconstruction in the multimodal sensing network unit, high-frequency dynamic temperature tracking is realized through the MEMS infrared thermopile; precise positioning of microscopic hot spots is realized through the fiber Bragg grating sensor; and lossless acquisition of millivolt-level TEG signals is realized through the zero-drift operational amplifier.

[0115] The thermal field modeling algorithm is based on spatial interpolation of the moving least squares method (MLS) and integrates multispectral data (3-5μm + 8-12μm bands);

[0116] TEG characteristic monitoring, output impedance analysis, the Seebeck coefficient is calculated in real time through the sweep frequency method (10Hz - 100kHz), and the health assessment is based on establishing an exponential decay model of ΔS / S0 and aging time (R 2 > 0.99).

[0117] The edge computing architecture of the intelligent decision-making core unit includes:

[0118] RISC-V custom instruction set, including THERMAL_FMA as the thermal-electric coupling matrix operation fusion instruction, NSGA3_OPT as the multi-objective optimization hardware acceleration, and FED_UPDATE as the federated learning parameter aggregation instruction;

[0119] Thermal-electric coupling model:

[0120] In the formula, To represent the gradient operator, which is used to calculate the rate of change in space; k is the thermal conductivity, representing the ability of the material to conduct heat; is the temperature gradient, representing the rate of change of temperature in space; ρ is the resistivity, representing the degree of obstruction of the material to the flow of electric current; J is the current density, representing the current passing through per unit area; α is the thermoelectric conversion coefficient, representing the conversion efficiency between electrical energy and thermal energy; T is the temperature; represents heat conduction; ρJ 2 represents Joule heat, the heat generated when an electric current passes through a resistor; represents the thermoelectric effect, that is, the current caused by the temperature gradient; the solution method is the finite volume method.

[0121] In the actuator network unit, the bionic muscle structure includes: the biological inspiration for the strain rate comes from the muscles of the octopus tentacle, the biological inspiration for the energy density comes from the insect flight muscle, and the biological inspiration for the self-sensing ability comes from the human proprioceptor; the dynamic range of contact pressure control is 0.1 - 5 N (resolution 0.01 N), and the overshoot is < 3% (step response test);

[0122] The lossless commutation technology of the GaN topology reconstruction switch, zero voltage switching (ZVS) timing control, dead time < 2 ns; full parallel → 3 series 2 parallel topology mode, suitable for the game load mutation scenario, switching time is 80 ns, full series → full parallel topology mode, suitable for low-temperature startup, switching time is 120 ns.

[0123] In the key control strategy unit, the differential temperature priority power generation algorithm, based on morphological gradient detection, selects the 3 units with the largest temperature difference;

[0124] Multi-objective optimization control, using the Pareto front to solve, objective function:

[0125] max(ηTEG), min(T hotspot ), min(P control )

[0126] In the formula, ηTEG is the thermoelectric power generation efficiency, representing the efficiency of the thermoelectric generator (TEG) to convert thermal energy into electrical energy; T hotspot is the hot spot temperature, referring to the point with the highest temperature in the system; P control is the control power;

[0127] The constraint conditions of reliability constraint, efficiency constraint, and energy consumption constraint are respectively chip junction temperature ≤ 95 °C, ΔT ≥ 10 °C, and system power consumption ≤ 50 mW;

[0128] NSGA-Ⅲ parameters: population size 50, number of reference points 12;

[0129] Refer to Figure 5As shown, a method for recycling waste heat from a laptop based on the thermoelectric conversion effect

[0130] The three-dimensional temperature distribution on the surfaces of the CPU, GPU, and heat pipes is scanned in real time by an infrared thermal imaging sensor array. Based on the gradient descent algorithm, the high-efficiency recovery areas of the temperature difference are identified, and the micro-TEG unit array at the corresponding positions is preferentially activated;

[0131] A liquid metal thermal interface material is used to dynamically adjust the heat flow path, and an electromagnetic pump is used to control the flow rate of the nanofluid to keep the hot end temperature of the TEG stable within a given temperature range; The inclination angle of the cold-end micro heat sink fins and the fan PWM duty cycle are synchronously adjusted to maintain the cold-end temperature below the given temperature;

[0132] The millivolt-level voltage output by the TEG is processed in three stages, including charge pump pre-boosting, resonant DC-DC conversion, and segmented charge and discharge of the hybrid energy storage system;

[0133] The change trend of the CPU power consumption in the short term in the future is predicted by an LSTM neural network, and the series-parallel topology of the TEG units is dynamically adjusted;

[0134] When it is detected that the efficiency decay of a single TEG unit exceeds a given ratio, it is automatically switched to a pure heat conduction channel and the power supply path is reallocated.

[0135] It should be noted that the magnetic fluid characteristics of liquid metal are used at the hot end to achieve directional guidance of heat flow, and the heat flux distribution is dynamically adjusted by an electromagnetic pump; at the cold end, the fin inclination angle and fan PWM are optimized based on the pneumatic-thermal coupling model to maintain the power generation threshold of ΔT>15°C;

[0136] Energy conversion chain efficiency: η = η TEG ×η DC-DC ×η storage = 7% × 94% × 98% ≈ 6.5%

[0137] In the formula, η is the total energy conversion efficiency, η TEG is the conversion efficiency of the thermoelectric generator, η DC-DC is the efficiency of the DC-DC converter, η storage is the efficiency of the energy storage system.

[0138] Refer to Figure 6 As shown in the laptop system structure diagram, the front view of the laptop is shown in the upper left corner of the figure, and the bottom view of the laptop is shown in the lower left corner; the figure on the right shows the internal structure of the laptop, and the positions of the key components are shown in the figure, including the TEG module, motherboard, DC-DC boost chip, battery, DCover, heat dissipation module, CPU chip, GPU chip, DCover, recycled energy storage battery device.

[0139] Refer toFigure 7 As shown in the figure, it is a structural diagram of a heat dissipation module, which shows the internal structure of the heat dissipation module. The key components include the TEG hot end: in contact with the heat pipe, used to absorb heat; the TEG cold end: in contact with the fins, used to dissipate heat; the heat pipe: used to conduct heat from the heat source to the TEG hot end; the heat dissipation fins: used to increase the heat dissipation area and improve the heat dissipation efficiency; the fan: controlled by PWM (pulse width modulation), used for forced convection heat dissipation. The TEG unit is embedded inside the heat dissipation module and integrated with the heat pipe and the vapor chamber to avoid occupying additional space.

[0140] Refer to Figure 8 As shown in the figure, it is a schematic diagram of the implementation principle, which needs to be combined with Figure 6 , 7 to view. The figure shows the layout and connection method of the TEG module inside the notebook computer. The series-connected TEG module, DC-DC boost chip, and recycled energy storage battery device. Its working process is that the heat pipe conducts the heat generated by heat-generating components such as the CPU to the hot end of the TEG module. The TEG module uses the thermoelectric effect to convert this heat into electrical energy. The generated electrical energy is converted into a suitable voltage and current through the boost chip. The converted electrical energy is used to supply power to low-power modules such as the camera, fill light, and sensor inside the notebook computer.

[0141] Refer to Figure 9 As shown in the figure, it is a schematic diagram of the working principle and application scenarios

[0142] Scenario 1: Camera fill light power supply. The recycled electrical energy is directly used to drive the LED fill light in the camera area, solving the problem of additional power consumption in video calls in low-light environments.

[0143] Scenario 2: Sensor and peripheral power supply. Provide auxiliary power for the keyboard backlight, touchpad, and biometric module (fingerprint / face recognition).

[0144] Refer to Figure 10 As shown in the figure, it is a LED boost drive circuit diagram for camera fill light power supply. This circuit is used to increase the input voltage to drive the LED, and precise current control and brightness adjustment are achieved through the feedback resistor.

[0145] U3 (LED boost drive chip): This is the core component of the circuit, responsible for increasing the input voltage and controlling the brightness of the LED; L1 (inductor): used for energy storage and voltage conversion;

[0146] D1 (diode): used to prevent reverse current flow and protect the circuit;

[0147] C1 (input capacitor): used to smooth the input voltage and reduce voltage fluctuations;

[0148] C2 (output capacitor): used to smooth the output voltage and ensure the stable operation of the LED;

[0149] R1, R2, R3 (Resistors): Used for current detection and feedback control;

[0150] Working principle: The input voltage is initially smoothed and boosted through inductor L1 and capacitor C1; The LED boost driver chip U3 controls the charging and discharging process of inductor L1 to achieve voltage increase; Diode D1 ensures unidirectional current flow and prevents reverse current from damaging the circuit; The output voltage is further smoothed by capacitor C2 to drive the LED; Resistors R1, R2, R3 are used for current detection and feedback to ensure stable operation of the circuit.

[0151] Refer to Figure 11 As shown, the brightness detection circuit diagram for the camera fill light power supply, which can convert optical signals into electrical signals and output through an analog-to-digital converter (ADC).

[0152] Photoresistor (R1): Used to detect the ambient light intensity, and its resistance value changes with the light intensity;

[0153] Operational amplifier (U1): Used to amplify the voltage change generated by the photoresistor;

[0154] Resistors (R2, R3): Used to set the gain and bias of the operational amplifier;

[0155] Reference voltage (V_REF_2V5): Provides a stable 2.5V reference voltage;

[0156] Analog-to-digital converter (ADC): Converts the analog voltage signal into a digital signal for further processing.

[0157] Working principle: The resistance value of the photoresistor R1 changes with the ambient light intensity, resulting in a voltage change across it; The operational amplifier U1 is configured as a voltage follower or amplifier to amplify the voltage change of the photoresistor; The amplified voltage signal is further processed through resistors R2 and R3 and then input to the analog-to-digital converter ADC; The ADC converts the analog voltage signal into a digital signal and outputs it to a microcontroller or other digital devices for processing;

[0158] Using an operational amplifier and an analog-to-digital converter to achieve high-precision brightness detection, and by adjusting the values of resistors R2 and R3, the gain and sensitivity of the circuit can be changed.

[0159] In summary, the advantages of the present invention are: An innovative thermoelectric waste heat recovery system is proposed, which is integrated inside the laptop. Through a unique design, efficient conversion and utilization of waste heat are achieved, thus achieving the goals of energy conservation and environmental protection;

[0160] The ingenious integration design of the thermoelectric conversion module and the radiator not only effectively reduces the thermal resistance, improves the thermoelectric conversion efficiency, but also extends the service life of the product, enhancing the stability and durability of the system; the thermoelectric module can convert waste heat into electrical energy, which can not only relieve the original heat dissipation pressure of the laptop, but also improve the overall performance of the product. The integrated design of the present invention is compact and highly integrated, very suitable for application in ultra-thin products such as laptops, and has extremely broad market application prospects.

[0161] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments, and what is described in the above embodiments and the specification is only the principle of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection required by the present invention is defined by the appended claims and their equivalents.

Claims

1. A waste heat recovery and utilization system for laptop computers based on the thermoelectric conversion effect, characterized in that Including: Distributed thermoelectric array module: Composed of multiple micro-TEG units, distributed in the high-heat generation areas of the notebook, including the CPU heat dissipation module substrate, the middle bending area of the heat pipe, and the air outlet guide fins, generating electricity directly through the temperature difference and supplying electrical energy to the device functional modules; Adaptive thermal management module: Including a thermal conductive silicone layer and micro heat dissipation fins, optimizing the temperature difference between the hot end and the cold end; The hot end forms a directional heat channel with the CPU top cover through laser micro-welding, and the cold end fin spacing is designed with a gradient; Multi-stage power management module: Including a rectifier circuit, a DC-DC boost chip, and a recycled energy storage battery device, converting the unstable low-voltage direct current output by the TEG into stable electrical energy, storing it and supplying it to the target device; Intelligent control module: Dynamically adjusts the working area of the micro-thermoelectric power generation unit according to the temperature sensor data, preferentially selects the position with the largest temperature difference to generate electricity, and matches the power supply demand through the PMIC chip.

2. The waste heat recovery and utilization system and method of a laptop based on the thermoelectric conversion effect according to claim 1, characterized in that, The distributed thermoelectric array module specifically includes: Dynamic thermal impedance matching unit: Each TEG unit is integrated with a micro heat flux sensor at the bottom to monitor the heat flux distribution in real time; dynamically adjusts the unit inclination angle through a piezoelectric ceramic actuator to align the heat flow direction with the optimal orientation of the thermoelectric lattice; Three-dimensional interconnected power supply unit: Adopts laser-induced graphene wiring to directly form a circuit on the surface of the heat dissipation module. The TEG units in each area achieve series-parallel adaptive recombination through a topology optimization algorithm, maintaining the stability of the output voltage within a given temperature difference range; Failure tolerance mechanism unit: Built-in TEG unit health assessment model, predicting the remaining life through the decay rate of the Seebeck coefficient; when the efficiency of a certain unit drops to a given threshold, the system automatically switches it to a pure heat conduction channel and maintains the total power output through the current redistribution of adjacent units.

3. The waste heat recovery and utilization system and method for a laptop based on the thermoelectric conversion effect according to claim 2, characterized in that, The distributed thermoelectric array module specifically includes: CPU heat dissipation module substrate area: Deploy multiple micro-TEG units in an asymmetric star-shaped arrangement, with the center point offset from the geometric center of the CPU top cover to match the actual hot spot distribution; fill the space between the TEG units with a thick boron nitride thermal conductive insulating film to achieve electrical isolation while maintaining the lateral thermal conductivity; Middle bending area of the heat pipe: Deploy multiple flexible TEG units, with the substrate made of copper-graphene composite foil, conforming to the curved shape of the heat pipe; integrate a micro shape memory alloy spring to automatically increase the contact pressure when the temperature of the heat pipe exceeds a given temperature; Air outlet guide fin area: Deploy multiple air flow-driven TEG units, with micron-level turbulent structures processed on the fin surface, using the exhaust air flow convection to reduce the cold end temperature.

4. The waste heat recovery and utilization system for laptop based on the thermoelectric conversion effect according to claim 3, characterized in that, The adaptive thermal management module specifically includes: Hot end conduction unit: The thermal conductive silicone layer uses graphene / boron nitride hybrid fillers and adds phase change microcapsules, defining the critical temperature of the phase change microcapsules; Cold end heat dissipation unit: The micro heat dissipation fins are designed with a gradient arrangement according to the near heat source area, far heat source area, and transition area, and the transition area adopts a bionic spiral gradient design; Heat flux sensor unit: Real-time monitoring of the heat flux distribution through an embedded micro heat flux sensor, combined with a shape memory alloy actuator to dynamically adjust the fin angle. Computational fluid dynamics simulation is used to guide the optimization of the fin angle, enabling the automatic optimization of the heat dissipation efficiency according to the load change. Phonon engineering is introduced to optimize the silica filler network and improve the intermediate-frequency phonon transmission efficiency.

5. A waste heat recovery and utilization system for a laptop based on the thermoelectric conversion effect according to claim 4, characterized in that, The multi-level power management module specifically includes: Front-end signal processing unit: A wide-input-range rectifier circuit using zero-threshold-voltage GaN diodes and integrating self-biased synchronous rectification technology; a dynamic impedance matching network using a digitally tunable LC resonant circuit and real-time tracking of the TEG internal resistance change through an impedance analyzer. Energy conversion core unit: A DC-DC boost chip based on magnetic-coupled resonance wireless energy transfer to achieve adaptive multi-topology switching; a supercapacitor buffer using graphene / CNT composite electrode materials to increase the pulse charge and discharge tolerance. Backend energy storage and distribution unit: The hybrid energy storage device includes short-term energy storage using a 3D silicon anode lithium battery; instantaneous buffering using a solid-state micro-supercapacitor; an intelligent power distribution network for priority management of the CPU, USB, and battery charging weights, and having a wireless backcharging function.

6. A waste heat recovery and utilization system for a laptop computer based on the thermoelectric conversion effect according to claim 5, characterized in that, The intelligent control module specifically includes: Multi-modal sensing network unit: A distributed temperature sensing array uses a hybrid network of MEMS infrared thermopiles and fiber Bragg grating sensors to achieve three-dimensional thermal field reconstruction of the CPU / GPU surface; an integrated zero-drift operational amplifier to collect the output characteristics of each TEG unit in real time. Intelligent decision-making core unit: Using edge computing to run a thermal-electric coupling model, customizing energy efficiency optimization instructions based on the RISC-V instruction set; an adaptive PMIC integrating a digital decoupling algorithm to eliminate crosstalk between channels; using federated learning to achieve cross-device fault prediction. Actuator network unit: Dynamically adjusting the contact pressure of the TEG unit through a piezoelectric micro-motion platform, adopting a bionic muscle structure; based on GaN power devices, achieving dynamic series-parallel recombination of the TEG unit and supporting lossless current commutation. Key control strategy unit: Applying the temperature difference priority power generation algorithm, combined with multi-objective optimization control, whose constraint conditions include reliability constraint, efficiency constraint, and energy consumption constraint, and using the NSGA-Ⅲ algorithm for Pareto front optimization.

7. A method for recycling waste heat of a laptop based on the thermoelectric conversion effect, according to the system for recycling waste heat of a laptop based on the thermoelectric conversion effect described in claims 1-6, characterized in that, Including: Real-time scanning of the three-dimensional temperature distribution of the CPU, GPU, and heat pipe surfaces through an infrared thermal imaging sensor array, identifying the high-efficiency recovery area of the temperature difference based on the gradient descent algorithm, and preferentially activating the micro-TEG unit array at the corresponding position. Using a liquid metal thermal interface material to dynamically adjust the heat flow path, controlling the nanofluid flow rate through an electromagnetic pump to keep the TEG hot end temperature stable within a given temperature range; synchronously adjusting the fin angle of the cold-end micro heat sink and the fan PWM duty cycle to maintain the cold-end temperature below the given temperature. Performing three-level processing on the millivolt-level voltage output by the TEG, including charge pump pre-boosting, resonant DC-DC conversion, and segmented charge and discharge of the hybrid energy storage system. Predicting the future short-term CPU power consumption change trend through an LSTM neural network and dynamically adjusting the TEG unit series-parallel topology. When it is detected that the efficiency decay of a single TEG unit exceeds a predetermined ratio, it is automatically switched to a pure heat conduction channel and the power supply path is reallocated.

8. The waste heat recovery and utilization system of a laptop based on the thermoelectric conversion effect according to claim 7, characterized in that, The real-time scanning of the three-dimensional temperature distribution on the surfaces of the CPU, GPU, and heat pipe by the infrared thermal imaging sensor array, and the identification of the high-efficiency recovery regions of the temperature difference based on the gradient descent algorithm, and the priority activation of the micro-TEG unit array at the corresponding positions specifically include: A hybrid network of quantum dot infrared detectors and MEMS thermopiles is used to form a hexagonal monitoring grid with a predetermined spacing at the top cover of the CPU and the bend of the heat pipe; Perform nanoscale heat flux density calculation on the hot spots on the surface of the silicon chip, evaluate the installation quality of the TEG for the metal interface contact thermal resistance, and detect the aging of the interface material for the voids inside the thermal paste layer; Filter out the fan vibration noise based on wavelet transform, perform three-dimensional temperature field reconstruction using the moving least squares interpolation method, and calculate the heat flux vector by combining Fourier's law and the Navier-Stokes equation; Based on the selection area algorithm of adaptive gradient descent, calculate the temperature gradient, locate the gradient extreme region, filter through the temperature difference threshold, and output the optimal activation sequence; calculate the dynamic load balancing strategy through the multi-objective optimization function.

9. The waste heat recovery and utilization system of a laptop based on the thermoelectric conversion effect according to claim 8, characterized in that The dynamic adjustment of the heat flow path by using the liquid metal thermal interface material, controlling the flow rate of the nanofluid by the electromagnetic pump to keep the temperature of the hot end of the TEG stable within a predetermined temperature range; synchronously adjusting the inclination angle of the micro heat sink fins at the cold end and the PWM duty cycle of the fan to keep the temperature at the cold end lower than the predetermined temperature specifically include: The liquid metal thermal interface material includes a base alloy doped with diamond nanoparticles, a magnetic additive modified with a surface silane coupling agent, and a packaging layer with a micro-groove capillary structure; The drive of the electromagnetic pump uses a combination of a Halbach array permanent magnet and a micro solenoid; The inclination angle of the micro heat sink fins is driven by a shape memory alloy wire, and the intelligent speed regulation of the fan is completed through a multi-parameter coupling control model.

10. A waste heat recovery and utilization system for a laptop based on the thermoelectric conversion effect according to claim 9, characterized in that, The three-level processing of the millivolt-level voltage output by the TEG includes charge pump pre-boost, resonant DC-DC conversion, and segmented charge and discharge of the hybrid energy storage system; predicting the change trend of the CPU power consumption in the short term in the future through the LSTM neural network, and dynamically adjusting the series-parallel topology of the TEG unit specifically include: The charge pump pre-boost stage uses zero-threshold voltage MOSFETs to construct a cross-coupled charge pump and integrates self-oscillation control; For the resonant DC-DC conversion, an ultra-thin planar transformer and GaN switching tubes are used, and dual-mode control of variable frequency - variable duty cycle is adopted; The LSTM neural network prediction is through a power consumption prediction model, and the time domain characteristics of this model include CPU occupancy rate, core temperature, and instruction throughput, and the frequency domain characteristics include the energy distribution of 0.1 - 100 Hz decomposed by FFT; The dynamic reconstruction of the TEG topology is based on a GaN matrix switch, and the topology modes include: high-temperature difference steady state with all-parallel connection; transient load impact with 3 series and 2 parallel connection; low-temperature environment with all-series connection.

Images