Detection method and system of electric heating medical beauty equipment

Abstract

The invention belongs to the field of electromagnetic variable measurement, and discloses a detection method and system of electric heating medical beauty equipment. Comprising a bionic skin model and a main measurement center. The skin model is constructed based on a multilayer composite material and a dynamic regulation and control technology, a high-resolution camera array is deployed outside the bionic skin model, a high-precision electric-magnetic-thermal composite sensor array is integrated inside the bionic skin model, and multi-dimensional data is acquired and transmitted to an upper computer and a main measurement center; the pre-examination safety table is used for performing stable pre-examination on the output of the beauty equipment; the fusion detection module generates a multi-modal data report by using a multi-branch distribution method, deeply learns the multi-modal data report by using a machine learning model, outputs a quantitative index, calculates and judges whether the beauty equipment is qualified or not according to the quantitative index, and sends the qualified beauty equipment to the upper computer, and the upper computer is used for visual interaction and data storage. The detection of the electric heating medical beauty equipment is realized.

Description

Technical Field

[0001] The present invention relates to the field of electromagnetic variable measurement, and more specifically, to a detection method and system for an electrothermal medical beauty device. Background Art

[0002] When detecting an electrothermal medical beauty device, experiments cannot be directly carried out on real people, which poses certain risks. However, before mass production of the electrothermal medical beauty device, pre-experiments are required. Therefore, a bionic skin dynamic simulation platform is designed for a high-precision bionic test system for detecting electrothermal medical beauty devices, which simulates the biophysical properties of human skin at different ages (such as thickness, thermal conductivity, resistivity, water content, elastic modulus, etc.). However, generally only electrothermal is detected during testing, and magnetic variation is rarely considered.

[0003] Secondly, due to the complex texture effect of human skin, electricity, heat, and magnetism do not act on the skin alone. Electricity and magnetism will couple, and electricity and heat will also couple. Therefore, there are certain limitations in measuring electricity and magnetism alone, and it cannot comprehensively reflect the effects of beauty devices. Moreover, the measuring devices also have a large difference from the physical and chemical reactions of real skin.

[0004] In view of this, the present invention proposes a detection system for an electrothermal medical beauty device to solve the above problems. Summary of the Invention

[0005] To overcome the above defects of the prior art and to achieve the above object, the present invention provides the following technical solution: A detection system for an electrothermal medical beauty device, comprising a bionic skin model and a main measurement center; The bionic skin model is constructed based on multi-layer composite materials and dynamic regulation technology, and uses a three-layer gradient structure to simulate bionic skin, and the layers are bonded by gradient bonding technology; and the skin state under different environmental conditions is simulated through an external control system; an integrated bionic microcirculation physiology is deployed, including embedding photosensitive materials, microfluidic chips and an integrated graphene heating layer; A high-resolution camera array is deployed outside the bionic skin model, and a high-precision electric-magnetic-thermal composite sensor array is integrated inside to form a three-dimensional sensing network, and multi-dimensional data is collected and transmitted to the upper computer and the main measurement center; The main test center includes a pre-inspection safety station and a fusion detection module. The pre-inspection safety station is used to stably pre-inspect the output of beauty equipment; the fusion detection module uses a multi-branch shunt method to perform multi-interaction fusion processing on the collected multi-dimensional data, generate a multi-modal data report, pre-define and quantify the detection indicators of beauty equipment, use a machine learning model to deeply learn the multi-modal data report and then output quantitative indicators, and introduce a causal chain into the machine learning model. The causal chain adopts a hierarchical dynamic selection architecture, which is divided into a bottom causal layer, a middle causal layer, and a top causal layer; Calculate and determine whether the beauty equipment is qualified according to the quantitative indicators and send it to the host computer, which is used for visual interaction and data storage.

[0006] Preferably, the method for constructing the bionic skin model based on multi-layer composite materials and dynamic regulation technology includes: Use a three-layer gradient structure for bionic skin simulation, namely the epidermis layer, the dermis layer, and the subcutaneous tissue layer; the thickness of each layer is dynamically adjusted according to age, and the layers are bonded through gradient bonding technology; The epidermis layer uses a polymer to simulate the stratum corneum, and a hydrophobic coating is applied on the surface to simulate the skin barrier function; the dermis layer uses a hydrogel substrate to simulate the collagen fiber density and elastic modulus; the subcutaneous tissue layer uses fat to simulate the gel to regulate the water content and thermal conductivity; set an external control interface and connect it to an external control system to simulate the skin state under different environmental conditions; embed photosensitive materials in the epidermis layer or the dermis layer, and pre-define the area and concentration of the photosensitive materials; Deploy an integrated bionic microcirculation physiological evolution, including embedding a microfluidic chip and an integrated graphene heating layer between the dermis layer and the subcutaneous tissue layer. The microfluidic chip is used to simulate the capillary network, the network density is adjusted according to age, set a fluid medium to simulate blood, and the flow rate is adjusted by programming; the integrated graphene heating layer is used to simulate the electrical and thermal responses of skin tissue; The microfluidic chip is driven by an external peristaltic pump. By adjusting the flow rate and fluid temperature, dynamically simulate the heat conduction characteristics of the skin under different environmental temperatures or exercise states; use a Peltier effect heating / cooling device to control the initial temperature of the bionic skin; adjust the water content of the bionic skin through a micro-humidifier and a desiccant; Conduct electrical property simulation on the bionic skin, including setting different resistivity and conductivity according to the characteristics of the skin layer; Preset the microfluidic flow rate and skin temperature according to requirements, simulate the skin state under different environmental conditions, and dynamically adjust the parameters of the bionic skin according to the output of the beauty equipment.

[0007] Preferably, the dynamically adjusting the parameters of the bionic skin according to the output of the beauty equipment includes: The predetermined working outputs of the beauty device include radio frequency, ultrasonic waves, lasers, electrical current stimulation, and cold and heat therapies; For the heat, electrical current, electromagnetic, or mechanical vibrations output by the beauty device, the high molecular polymers and hydrophobic coatings in the epidermis layer undergo deformation or surface property changes; the hydrogel substrate in the dermis layer undergoes changes in elastic modulus and water content; the fat-mimicking gel in the subcutaneous tissue layer undergoes changes in thermal conductivity and water content; the microcirculation dynamically adjusts the flow rate and fluid temperature through an external control system; the capillary network density is dynamically adjusted through the external control system to simulate the dynamic process of skin tissue repair or aging; Simulate the resistivity and conductivity of each layer of the bionic skin; the sensor array in the electromagnetic-thermal composite sensor array real-time collects multi-dimensional data of the beauty device output acting on the bionic skin model; Preset a biochemical protection mechanism, including when the temperature of the dermis layer is too high, reducing the initial temperature or increasing the water content through a Peltier effect device or a micro-humidifier to simulate the self-protection of the skin; if the mechanical vibration of the subcutaneous tissue layer is too strong, reducing the microfluidic flow rate through an external control system to simulate the stress relaxation of the skin tissue.

[0008] Preferably, the integrated deployment method of the high-precision electro-magnetic-thermal composite sensor array includes: Deploy a high-resolution camera array outside the bionic skin model to comprehensively monitor the surface changes of the bionic skin model; embed micro strain sensors in the dermis layer and the subcutaneous tissue layer to measure the local stress changes caused by heat; For resistance, current, and voltage, an electrical variable sensor combines a microelectrode array and an impedance analyzer for measurement; introduce a wide-band frequency sweep technology; Use a TMR magnetoresistive sensor array to measure the magnetic field strength and magnetic flux changes; for the temperature of each layer, a thermal field sensor uses a thermocouple array for measurement; use FLIR Lepton3.5 infrared thermal imaging to real-time collect the surface temperature distribution map; The sensor array is evenly distributed in the epidermis layer, dermis layer, and subcutaneous tissue layer of the bionic skin at a spacing of 5 mm to form a three-dimensional sensing network, and the number of sensors is determined according to the size of the test area; it is connected to a data acquisition card through a multiplexer, and the data acquisition card uses a high-precision analog-to-digital converter and is real-time transmitted to the upper computer and the main measurement center through USB or Wi-Fi.

[0009] Preferably, the pre-inspection safety platform is used to stably pre-inspect the output of the beauty device, and the method includes; The pre-inspection safety platform uses a high-precision analog-to-digital converter ADC, combines the fast Fourier transform algorithm, and calculates the total harmonic distortion rate THD of the PWM signal through spectrum analysis; If the detected duty cycle fluctuation exceeds 0.5% or the THD exceeds 3%, the host computer issues a warning and switches the mode or shuts down the beauty device.

[0010] Preferably, the module uses a multi-branch shunt method to perform multi-interaction fusion processing on the collected multi-dimensional data and generate a multi-modal data report. The method includes: The report content includes a three-dimensional electric field distribution map, a three-dimensional magnetic field distribution map, a three-dimensional thermal field distribution map, and key demonstration characteristics; The method for obtaining the three-dimensional electric field distribution map is to collect the spatial distribution data of electrical parameters through a microelectrode array, and use the Kriging interpolation method or the finite element analysis algorithm to construct the three-dimensional electric field distribution map of the bionic skin model. The electrical parameters include resistance, current, and voltage; The method for obtaining the three-dimensional magnetic field distribution map is to perform denoising processing on the magnetic field intensity and magnetic flux changes collected by the TMR sensor array, and use the magnetic tomography algorithm. Based on Maxwell's equations and finite element analysis, the spatial distribution of the magnetic field in the bionic skin model is inverted. The algorithm input is the magnetic field intensity and magnetic flux changes, and the output is the three-dimensional magnetic field distribution map. The abnormal areas in the magnetic field distribution map are identified through gradient analysis, and their spatial positions are marked; The method for obtaining the three-dimensional thermal field distribution map is to fuse the point measurement data of the thermocouple array with the surface temperature distribution map of the FLIR Lepton 3.5 infrared thermal imaging through the Bayesian estimation algorithm to generate the three-dimensional thermal field distribution map; The methods for obtaining key demonstration characteristics include: Use a microelectrode array and an impedance analyzer to measure the complex impedance of the graphene heating layer, analyze the impedance spectrum through the Nyquist diagram and Bode diagram, extract the equivalent circuit parameters of the graphene heating layer, and establish a time-series dynamic change curve according to the sampling frequency; Measure the temperature values of each layer at a fixed frequency through the thermocouple array, and combine the Fourier heat conduction equation to calculate the thermal conductivity and thermal diffusivity of the epidermis layer, dermis layer, and subcutaneous tissue layer; Use the finite element analysis algorithm to establish the heat conduction model of the bionic skin model. The input parameters include the thermal conductivity, water content, microfluidic flow rate, and fluid temperature of each layer, and the output is the heat conduction path and heat flux density distribution; For the data collected by the TMR magnetoresistive sensor array and the microelectrode array, use the Kalman filter algorithm to eliminate noise and interference; establish an electromagnetic field coupling model based on Maxwell's equations, extract key parameters through the coupling model, and establish a time-series dynamic change curve according to the sampling frequency; Use the finite element analysis algorithm to establish the thermal-mechanical coupling model of the bionic skin model. The input parameters include the thermal expansion coefficient, elastic modulus, and thermal stress distribution of each layer of material, and the output is the deformation distribution map and the stress concentration area.

[0011] Preferably, the method for calculating and determining whether a beauty device is qualified according to quantization indexes includes: It is set that the detection of the beauty device includes functional evaluation and safety evaluation, and quantization indexes for functional evaluation and safety evaluation are defined according to the mode of the beauty device; For the multi-modal data report, a hierarchical dynamic selection architecture is adopted to extract description primitives for representing data features, a predefined machine learning model is used to learn the description primitives, and quantization indexes are output. A causal chain is introduced into the machine learning model; Calculate the deviation value between the quantization index and the predefined effect index; Multi-objective comprehensive evaluation is deployed. The quantization indexes are weighted and averaged to obtain a comprehensive score. Hard constraints are predefined for the comprehensive score and the deviation value. If the beauty device meets all the hard constraints, it is determined to be qualified and can be put into mass production; otherwise, it is determined to be unqualified and further design improvement or re-detection experiment is required; Among them, the causal chain adopts a hierarchical dynamic selection architecture and is divided into a bottom causal layer, a middle causal layer, and a top causal layer; In the bottom causal layer, the heat output by the beauty device affects the surface temperature distribution map collected by FLIR Lepton3.5 infrared thermal imaging through the deformation of the polymer and hydrophobic coating in the epidermis layer, and then affects the three-dimensional thermal field distribution map. The current stimulation affects the complex impedance measured by the microelectrode array through the change of the elastic modulus and water content of the hydrogel substrate in the dermis layer, and then affects the three-dimensional electric field distribution map. The electromagnetic affects the change of the magnetic field strength and magnetic flux measured by the TMR magnetoresistive sensor array through the change of the thermal conductivity of the fat analog gel in the subcutaneous tissue layer, and then affects the three-dimensional magnetic field distribution map. The mechanical vibration affects the temperature values of each layer measured by the thermocouple array by reducing the microfluidic flow rate through the external control system, and then affects the three-dimensional thermal field distribution map; In the middle causal layer, the deformation of the epidermis layer affects the deformation distribution map and stress concentration area of the thermo-mechanical coupling model by measuring the local stress change through the micro strain sensor. The change of the elastic modulus and water content of the dermis layer affects the three-dimensional thermal field distribution map by measuring the temperature values of each layer through the thermocouple array. The change of the microfluidic chip flow rate and fluid temperature affects the heat conduction characteristics by adjusting the capillary network density through the external control system, and then affects the three-dimensional thermal field distribution map by fusing the thermocouple array data and the infrared thermal imaging data through the Bayesian estimation algorithm. When the temperature of the dermis layer is too high, the initial temperature or water content is adjusted through the Peltier effect device or the micro humidifier to affect the heat conduction characteristics, and then affects the three-dimensional thermal field distribution map; In the high-level causal layer, the electrical parameters of the three-dimensional electric field distribution map affect the three-dimensional magnetic field distribution map through Maxwell's equations. The abnormal regions of the three-dimensional magnetic field distribution map are identified through gradient analysis and feedback through the electromagnetic field coupling model to affect the three-dimensional electric field distribution map. The three-dimensional thermal field distribution map calculates the thermal conductivity and thermal diffusivity through the Fourier heat conduction equation, which affects the heat conduction path and heat flux density distribution of the heat conduction model, and further affects the deformation distribution map and stress concentration region of the thermal-mechanical coupling model.

[0012] Preferably, the method for extracting description primitives using a hierarchical dynamic selection architecture to represent data features includes: For the numerical data in the multimodal data report, after normalization, it is processed using a multi-layer perceptron; For the curve data in the multimodal data report, first perform dynamic time warping to synchronously align with the time slices of the three-dimensional map data; adopt a temporal Transformer structure to output a sequence of feature vectors in the time dimension; For the three-dimensional map data in the multimodal data report, use the numerical data as the physical benchmark for cross-modal alignment, expand the numerical data into the query vector of the attention mechanism, splice the temporal features and spatial features as the key and value to construct a spatio-temporal joint representation; develop a dual-channel parallel architecture based on sparsity detection. If the sparsity > 70%, preferentially enable CNN local feature extraction. If the sparsity < 30%, enable PCA global dimensionality reduction, and use a gradient penalty term to constrain the feature redundancy; Among them, for CNN, add channel importance scoring after each convolutional layer, close the channels with importance scores lower than the expected performance, and use octree adaptive sampling; Record the features extracted from the numerical data, curve data, and three-dimensional map data as description primitives.

[0013] Preferably, the method for the host computer to be used for visual interaction and data storage includes: The host computer is developed based on an open technology stack and is used to provide a user interface, including setting detection parameters, displaying real-time data, and generating detection reports, supporting data visualization and data storage.

[0014] Preferably, a detection method for an electrothermal medical beauty device includes: Step S1: Construct a bionic skin model. The thickness of each layer and the density of the capillary network are dynamically adjusted according to the age group. Set an external control interface to connect to an external control system to simulate the skin state under different environmental conditions. Predefine the area and concentration of photosensitive materials, blood flow rate, and temperature. Use a Peltier effect device to control the initial temperature, and adjust the water content through a micro-humidifier and a desiccant; Step S2: Collect multi-dimensional data through a high-resolution camera array and an electro-magnetic-thermal composite sensor array, and transmit it to the host computer and the main measurement center; Step S3: Conduct a stable pre-inspection on the output of the beauty device through the pre-inspection safety platform; Step S4: The main measurement center uses a multi-branch shunt method to perform multi-interaction fusion processing on the multi-dimensional data to generate a multi-modal data report; Step S5: Adopt a hierarchical dynamic selection architecture to extract description primitives, use a machine learning model to learn and output quantization metrics, introduce causal chain analysis, calculate the deviation value between the quantization metrics and the predefined effect metrics, conduct multi-objective comprehensive evaluation, and determine whether the device meets the hard constraints. If it is qualified, it can be mass-produced; otherwise, the design needs to be improved or re-tested; Step S6: The host computer conducts visual interaction.

[0015] Technical effects and advantages of the detection system for an electro-thermal medical beauty device of the present invention: 1. A bionic skin model constructed based on multi-layer composite materials and dynamic regulation technology simulates the biophysical properties of human skin through a three-layer gradient structure (epidermal layer, dermal layer, and subcutaneous tissue layer), including thickness, thermal conductivity, resistivity, water content, elastic modulus, etc. The thickness of each layer and the capillary network density can be dynamically adjusted according to age. Combining a microfluidic chip to simulate blood circulation and a graphene heating layer to simulate electrical and thermal responses, it truly restores the dynamic responses of the skin under different environmental conditions. Compared with traditional detection systems that only use single materials or static models, the bionic skin model of this design is closer to the physical and chemical properties of real human skin, avoids the danger of direct testing on the human body, and at the same time significantly improves the authenticity and reliability of the detection results, providing a more scientific basis for the research and mass production of electro-thermal medical beauty devices.

[0016] 2. A high-precision electro-magnetic-thermal composite sensor array is integrated in the bionic skin model to form a three-dimensional sensing network, which can collect multi-dimensional data of the output of the beauty device acting on the bionic skin in real time. Traditional detection systems often only focus on a single parameter (such as electricity or heat) and ignore multi-field coupling effects such as electro-magnetism and thermo-force. Through multi-dimensional data collection, this design comprehensively reflects the comprehensive impact of the beauty device on the skin, overcomes the limitations of traditional detection methods, and provides more comprehensive data support for device performance evaluation.

[0017] 3. The multi - interaction fusion processing of the collected multi - dimensional data is carried out using the multi - branch shunt method. By combining the hierarchical dynamic selection architecture to extract descriptive primitives and introducing causal chain analysis into the machine learning model, the accurate evaluation of the functionality and safety of the beauty device is realized. Compared with the traditional detection system that relies on manual analysis or simple statistics, this design can automatically identify abnormal areas in the data (such as magnetic field anomalies, heat stress concentration, etc.) through data fusion and intelligent analysis, and reveal the internal correlation between the device output and skin reactions through the causal chain (bottom layer, middle layer, upper layer). This method not only improves the evaluation accuracy but also significantly reduces human error, providing scientific guidance for the optimized design of the device.

[0018] In summary, the problems of insufficient detection accuracy, single data, ignoring the coupling effect, and potential safety hazards in the prior art are solved. Its technical effects and advantages not only enhance the scientific nature and reliability of the detection of electrothermal medical beauty devices but also provide strong technical support for the research and development optimization and mass production of the devices, with significant innovation and practical value. Brief Description of the Drawings

[0019] Figure 1 It is a schematic structural diagram of a detection system for an electrothermal medical beauty device of the present invention; Figure 2 It is a schematic flow diagram of a detection system for an electrothermal medical beauty device of the present invention; Figure 3 It is a schematic diagram of a detection method for an electrothermal medical beauty device of the present invention. Detailed Embodiments

[0020] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0021] Embodiment 1 Please refer to Figure 1 、 Figure 2 and Figure 3 As shown, the main design content of a detection system for an electrothermal medical beauty device in this embodiment is as follows: The method for constructing a bionic skin model based on multi - layer composite materials and dynamic regulation technology includes: Using a three - layer gradient structure for bionic skin simulation, namely the epidermis layer, dermis layer, and subcutaneous tissue layer; the thickness of each layer is dynamically adjusted according to the age group, and the layers are bonded through gradient bonding technology; For example: Ages 20 - 30: Epidermis layer is 0.1 mm, dermis layer is 1.5 mm, and subcutaneous tissue layer is 3.0 mm.

[0022] Ages 40+: Epidermis layer is 0.15 mm, dermis layer is 1.2 mm, and subcutaneous tissue layer is 2.5 mm.

[0023] The seamless transition between layers is achieved through gradient bonding technology to ensure the continuity of heat conduction and electrical properties.

[0024] The epidermis layer uses a polymer (such as polydimethylsiloxane PDMS) to simulate the stratum corneum, and a hydrophobic coating is applied on the surface to simulate the skin barrier function; the dermis layer uses a hydrogel substrate (such as polyethylene glycol PEG) to simulate the collagen fiber density and elastic modulus; the subcutaneous tissue layer uses a fat - like gel (such as a silicone - based composite material) to regulate the water content and thermal conductivity; an external control interface is set up and connected to an external control system to simulate the skin state under different environmental conditions; photosensitive materials (such as particles simulating melanin) are embedded in the epidermis layer or dermis layer, and the area and concentration of the photosensitive materials are predefined; For example: Ages 20 - 30: Collagen fiber density is 120 mg / cm³, elastic modulus is 2.5 MPa; water content is 70%, thermal conductivity is 0.25 W / m·K, and low - concentration photosensitive materials are set at the corners of the eyes and the cheekbones.

[0025] Ages 40+: Collagen fiber density is 60 mg / cm³, elastic modulus is 0.5 MPa; water content is 50%, thermal conductivity is 0.18 W / m·K, and high - concentration photosensitive materials are set at the corners of the eyes and the cheekbones.

[0026] Through material formulations and processing techniques (such as 3D printing, microfluidic perfusion), the dynamic matching of parameters for each layer is achieved. Using an external control system (such as a temperature control module, humidity adjustment module), the water content, thermal conductivity, and elastic modulus of the bionic skin are adjusted in real - time to simulate the skin state under different environmental conditions.

[0027] Deploy an integrated bionic microcirculation physiological evolution, including embedding a microfluidic chip and an integrated graphene heating layer between the dermis layer and the subcutaneous tissue layer. The microfluidic chip is used to simulate the capillary network; the microfluidic channels are fabricated using lithography technology, with a channel diameter of 50 - 100 μm, and the network density is adjusted according to age (higher density for young skin, lower density for mature skin); a fluid medium (saline or a bionic liquid containing red blood cells) is set to simulate blood, and the flow rate is adjusted by programming (range 0 - 15 mL / min); the integrated graphene heating layer is used to simulate the electrical and thermal responses of skin tissue; The microfluidic chip is driven by an external peristaltic pump to simulate the absorption and diffusion of heat by blood circulation. By adjusting the flow rate and fluid temperature (flow rate range 0-15mL / min, temperature range 20℃-40℃), the thermal conductivity characteristics of the skin under different ambient temperatures or motion states are dynamically simulated; the Peltier effect heating / cooling device is used to control the initial temperature of the bionic skin (range 20℃-40℃); the accuracy is ±0.1℃, which is used to simulate different ambient temperatures or initial skin states. The water content of the bionic skin is adjusted by a micro humidifier and desiccant (range 30%-80%); it is used to simulate the changes in the characteristics of the skin under different humidity environments.

[0028] For example: The heat transfer properties are dynamically matched via the thermal diffusivity: Young skin: Thermal diffusion coefficient 0.25W / m·K, simulating efficient heat diffusion.

[0029] Mature skin: Thermal diffusion coefficient is 0.18 W / m·K, simulating lower heat diffusion capacity.

[0030] Simulating electrical properties on the bionic skin, including setting different resistivity and conductivity according to the characteristics of the skin layer; For example: For resistivity: The resistivity of the epidermis is relatively high (simulating the insulating properties of the stratum corneum), about 10 5 Ω·m.

[0031] The dermis and subcutaneous tissue layers have a lower resistivity (simulating tissue with a higher water content), approximately 10²Ω·m.

[0032] For electrical conductivity: The electrical conductivity of each layer can be regulated by doping conductive particles (such as carbon nanotubes, graphene) into the hydrogel matrix, fat-mimicking gel and polymer.

[0033] Using external electric fields or ion implantation technology, the resistivity and conductivity of the bionic skin can be dynamically adjusted to simulate the changes in the electrical properties of the skin under different humidity or electrical stimulation.

[0034] The microfluidic flow rate and skin temperature are preset according to needs to simulate the state of the skin under different environmental conditions (exercise or rest), and the parameters of the bionic skin are dynamically adjusted according to the output of the beauty device to simulate the operation of the beauty device in different usage scenarios.

[0035] Dynamically adjust the parameters of the bionic skin based on the output of the beauty device, including: Work outputs scheduled for aesthetic devices include radiofrequency, ultrasound, laser, electrical current stimulation, and hot and cold therapies; Beauty devices usually have multiple working modes, including radio frequency (RF), ultrasound, laser, electrical stimulation, cryo- and thermo-therapy, etc. The outputs of these modes act on the bionic skin model in different forms of energy (such as electromagnetic energy, thermal energy, mechanical energy, etc.).

[0036] Specifically, the action mechanisms of the outputs of various modes are as follows: 1. Radio frequency mode: The radio frequency device outputs high-frequency alternating current (usually in the range of 0.3 MHz to 10 MHz), which acts on the skin through electrodes, generating an electromagnetic field and Joule heat.

[0037] Acting on the bionic skin model: Electromagnetic field action: The electromagnetic field output by radio frequency acts on the epidermis, dermis, and subcutaneous tissue layers of the bionic skin. Due to different resistivity and conductivity of each layer (preset through material design), the electromagnetic field causes changes in the charge distribution, simulating the electrical response of skin tissue.

[0038] Thermal effect: Joule heat generates heat in the dermis and subcutaneous tissue layers, and the heat distribution is affected by the thermal conductivity and water content of each layer. The integrated high-precision electromagnetic-thermal composite sensor array measures the thermal field distribution in real time (through a thermocouple array and infrared thermal imaging) and the changes in electrical parameters (through a microelectrode array).

[0039] Microcirculation simulation: The radio frequency thermal effect drives the fluid (simulating blood) in the microfluidic chip through an external peristaltic pump, adjusting the flow rate and fluid temperature, and dynamically simulating the accelerated microcirculation phenomenon of the skin after heating.

[0040] 2. Ultrasound mode: The ultrasound device outputs high-frequency mechanical vibrations (usually in the range of 1 MHz to 10 MHz), which act on the skin through a transducer, generating mechanical waves and local thermal effects.

[0041] Acting on the bionic skin model: Mechanical wave action: The mechanical vibration of ultrasound acts on each layer of the bionic skin, especially the hydrogel substrate (simulating collagen fibers) in the dermis and the fat-simulating gel in the subcutaneous tissue layer. The vibration causes tiny deformations inside the material, simulating the mechanical response of skin tissue.

[0042] Thermal effect: Ultrasound generates heat due to tissue impedance during propagation, especially at the junction of the dermis and subcutaneous tissue layers. The thermocouple array and infrared thermal imaging measure the heat distribution in real time.

[0043] Microcirculation simulation: The mechanical vibration of ultrasound simulates the vibration effect of the capillary network through the microfluidic chip, and may adjust the flow rate through an external control system to simulate the enhancement of blood circulation.

[0044] 3. Laser Mode: The laser device outputs light energy of a specific wavelength (such as 532nm, 1064nm, etc.), which acts on the skin through photothermal effect or photochemical effect.

[0045] Acting on the bionic skin model: Photothermal effect: The laser energy is absorbed by the polymer (simulating the stratum corneum) in the epidermis and the hydrogel substrate in the dermis and converted into heat. The heat distribution is monitored in real time through a thermocouple array and infrared thermal imaging.

[0046] Photochemical effect: By embedding photosensitive materials (such as particles simulating melanin) in the epidermis or dermis to simulate skin aging or pigmentation, the laser action triggers a photochemical reaction to simulate pigment decomposition or tissue repair of skin tissue.

[0047] Microcirculation simulation: The laser thermal effect adjusts the flow rate and fluid temperature of the microfluidic chip through an external control system to simulate the enhancement of local blood circulation.

[0048] 4. Current Stimulation Mode: The current stimulation device outputs low-frequency or medium-frequency current (such as microcurrent, iontophoresis, etc.), which acts on the skin through electrodes to promote tissue repair or substance penetration.

[0049] Acting on the bionic skin model: Electrical effect: The current passes through the epidermis, dermis, and subcutaneous tissue layers. Due to the different resistivity and conductivity of each layer, there are differences in the current distribution. The microelectrode array measures the changes in resistance, current, and voltage in real time.

[0050] Substance penetration simulation: If simulating iontophoresis, a solution simulating drugs or cosmetic ingredients can be coated on the surface of the epidermis, and the current action simulates substance penetration through the gradient adhesive layer.

[0051] Microcirculation simulation: The current stimulation adjusts the flow rate of the microfluidic chip through an external control system to simulate the enhancement of blood circulation.

[0052] 5. Cold and Heat Therapy Mode: The surface temperature of the skin is changed through a heating or cooling module (such as a Peltier effect device).

[0053] Acting on the bionic skin model: Thermal effect: The heating mode increases the surface temperature of the bionic skin, and the heat is transferred to the dermis and subcutaneous tissue layers through heat conduction. The thermocouple array and infrared thermal imaging monitor the heat distribution in real time.

[0054] Cold effect: The cooling mode reduces the surface temperature of the bionic skin. The water content may be adjusted through a micro humidifier and desiccant to simulate the contraction or moisturizing reaction of the skin in a low-temperature environment.

[0055] Microcirculation simulation: Heat and cold stimuli can regulate the flow rate and fluid temperature of the microfluidic chip through an external control system to simulate the dynamic changes in blood circulation.

[0056] The bionic skin model, through its multi-layer composite material design, dynamic regulation technology, and integrated high-precision sensor array, can respond autonomously according to the output of beauty devices. The following is the specific reaction mechanism: For the heat, current, electromagnetic, or mechanical vibration output by beauty devices, the polymer and hydrophobic coating on the epidermis layer generate deformation or changes in surface properties; for example, heat may cause a change in the surface tension of the hydrophobic coating, simulating the dynamic adjustment of the skin barrier function; current may change the resistivity of the epidermis layer, simulating the electrical barrier function of the skin.

[0057] The hydrogel substrate in the dermis layer undergoes changes in elastic modulus and water content; for example, the thermal effect of radio frequency or ultrasonic waves may cause water migration inside the hydrogel, simulating the thermoplastic change of collagen fibers; current stimulation may change the ion distribution in the hydrogel, simulating tissue repair.

[0058] The fat-mimicking gel in the subcutaneous tissue layer undergoes changes in thermal conductivity and water content; for example, the thermal effect of laser or radio frequency may cause local "melting" or water evaporation of the fat-mimicking gel, simulating the thermal response of subcutaneous fat.

[0059] Microcirculation dynamically regulates the flow rate and fluid temperature through an external control system; for example, the thermal effect of radio frequency or laser may increase the flow rate through program control, simulating the accelerated blood circulation after the skin is heated; the cold therapy mode may reduce the flow rate, simulating the slowdown of blood circulation in the skin at low temperatures.

[0060] The capillary network density of the microfluidic chip can be preset according to age (for example, the network density of young skin is higher and that of old skin is lower). The output of beauty devices (such as ultrasonic or current stimulation) may dynamically adjust the capillary network density through an external control system to simulate the dynamic process of skin tissue repair or aging; Simulate the resistivity and conductivity of each layer of the bionic skin; the microelectrode array will measure these changes in real time. For example, current stimulation may cause a decrease in the resistivity of the epidermis layer, simulating a temporary weakening of the skin barrier function; radio frequency electromagnetic fields may cause an increase in the conductivity of the dermis layer, simulating the electrical activation of skin tissue. The external control system will dynamically adjust the microfluidic flow rate or the electrical properties of the material layer according to these changes in electrical parameters to simulate the dynamic response of the skin under electrical stimulation.

[0061] In the electromagnetic-thermal composite sensor array, the sensor array (including microelectrode array, thermocouple array, fluxgate sensor, infrared thermal imaging, etc.) will collect multi-dimensional data (resistance, current, voltage, magnetic field, temperature, etc.) output by the beauty device acting on the bionic skin model in real time. These data are transmitted to the host computer through a multiplexer (MUX) and a high-precision analog-to-digital converter (ADC).

[0062] A preset biochemical protection mechanism is included. When the output of the beauty device (radio frequency) causes the temperature of the dermis layer to be too high, the initial temperature is reduced or the water content is increased through a Peltier effect device or a micro-humidifier to simulate the self-protection of the skin. If the output of the ultrasonic device causes the mechanical vibration of the subcutaneous tissue layer to be too strong, the microfluidic flow rate is reduced through an external control system to simulate the stress relaxation of the skin tissue.

[0063] The method of integrating a high-precision electro-magnetic-thermal composite sensor array in the bionic skin model includes: To achieve real-time collection of multi-dimensional data, a high-precision electromagnetic-thermal composite sensor array is integrated in the bionic skin model.

[0064] A high-resolution camera array is deployed outside the bionic skin model to comprehensively monitor the surface changes of the bionic skin model (e.g., mechanical deformation); the deformation measurement accuracy is 0.01 mm.

[0065] Micro strain sensors (such as fiber Bragg grating sensors, FBG) are embedded in the dermis layer and the subcutaneous tissue layer to measure local stress changes caused by heat. The strain measurement range is ±5000 με, and the resolution is 1 με; For resistance, current, and voltage, an electrical variable sensor combines a microelectrode array and an impedance analyzer for measurement; for example, the measurement range: resistance 10 Ω - 10 6 Ω, current 0.1 mA - 100 mA, voltage 0.1 V - 50 V; a wide-band swept-frequency technology is introduced; the impedance change of the graphene heating layer is measured through the wide-band swept-frequency technology to evaluate its dynamic response of electrical and thermal properties. A wide-band sweep from 10 Hz to 1 MHz is used, covering the common low-frequency (current stimulation) and high-frequency (radio frequency) output ranges of beauty devices. The sweep resolution is 0.1 Hz to ensure high-precision capture of impedance changes.

[0066] A TMR magnetoresistive sensor array is used to measure the magnetic field strength and magnetic flux changes; for example, the measurement range: magnetic field strength 0.01 μT - 1 mT, accuracy ±0.1 μT.

[0067] For the temperature of each layer, a thermal field sensor measures it using a thermocouple array; for example, the measurement range is 0°C - 100°C, with an accuracy of ±0.1°C. A FLIR Lepton 3.5 infrared thermal imager is used to collect the surface temperature distribution map in real time; the resolution is 160×120 pixels, the temperature measurement range is -10°C - 140°C, and the accuracy is ±0.5°C. The thermal imaging layer covers the surface of the bionic skin and is used to collect the surface temperature distribution map in real time.

[0068] The sensor array is evenly distributed in the epidermis layer, dermis layer, and subcutaneous tissue layer of the bionic skin at a spacing of 5 mm, forming a three-dimensional sensing network. The number of sensors is determined according to the size of the test area; for example, 400 sensor nodes are arranged in a 10 cm×10 cm test area. It is connected to a data acquisition card through a multiplexer (MUX). The sampling frequency is 100 Hz. The data acquisition card uses a high-precision analog-to-digital converter (ADC) with a resolution of 16 bits and is transmitted to the host computer and the main measurement center in real time through USB or Wi-Fi for subsequent analysis and visualization.

[0069] Traditional detection methods only focus on electrical variables and thermal field data, but the magnetic changes caused by electrothermal effects are ignored. Therefore, this solution innovatively introduces magnetic variable measurement to evaluate the electromagnetic compatibility during the operation of beauty devices and the electromagnetic characteristics of graphene materials. This not only improves the comprehensiveness of detection but also provides data support for the electromagnetic optimization design of new-generation beauty devices.

[0070] Case: A radio frequency beauty device acts on a bionic skin model Output of the beauty device: The radio frequency device outputs at a frequency of 1 MHz with a power of 50 W and acts on the surface of the bionic skin model.

[0071] Acting on the bionic skin model: Electromagnetic field action: The radio frequency electromagnetic field penetrates the epidermis layer and acts on the dermis layer and subcutaneous tissue layer. Due to the high conductivity of the hydrogel substrate in the dermis layer, the electromagnetic field will cause a redistribution of charges inside the dermis layer, and the microelectrode array measures the resistance and current changes in real time.

[0072] Thermal effect: The radio frequency energy is converted into heat in the dermis layer and subcutaneous tissue layer. The thermocouple array measures that the temperature in the dermis layer rises to 42°C, and the temperature in the subcutaneous tissue layer rises to 40°C. The infrared thermal imaging generates a surface temperature distribution map, showing that the heat is concentrated in the dermis layer.

[0073] Microcirculation response: According to the heat distribution data, the external control system automatically increases the flow rate of the microfluidic chip (from 5 μL / min to 10 μL / min) to simulate the acceleration of blood circulation for heat dissipation.

[0074] Autonomous response of the bionic skin model: Material-level reaction: The hydrogel substrate in the dermis layer expands due to heat, and its elastic modulus decreases, simulating the thermoplastic change of collagen fibers. The fat-simulating gel in the subcutaneous tissue layer evaporates water due to heat, and its thermal conductivity decreases, simulating the thermal barrier function of subcutaneous fat.

[0075] Dynamic regulation: After the host computer analyzes the sensor data, if it detects that the temperature of the dermis layer exceeds the safety threshold (e.g., 45 °C), it will reduce the initial temperature of the bionic skin through the Peltier effect device or increase the water content of the dermis layer through a micro-humidifier, simulating the self-protection mechanism of the skin.

[0076] Adjustment of electrical properties: The radio frequency electromagnetic field causes the electrical conductivity of the dermis layer to increase, and the microelectrode array measures that the resistance drops by about 20%. The host computer adjusts the microfluidic flow rate according to this change, simulating the enhanced blood circulation of skin tissue under electrical stimulation.

[0077] The main detection center includes a pre-inspection safety station, which is used for stable pre-inspection of the output of beauty equipment; Beauty equipment (such as radio frequency mode, current stimulation mode) usually outputs high-frequency pulse width modulation (PWM) signals to control the energy output intensity and mode. The pre-inspection safety station evaluates the stability and accuracy of the beauty equipment output by capturing and analyzing the duty cycle, frequency, and waveform characteristics of the PWM signal in real time.

[0078] The pre-inspection safety station uses a high-precision analog-to-digital converter ADC (resolution 16 bits, sampling frequency above 100 Hz), combined with the fast Fourier transform algorithm, to analyze the duty cycle fluctuation of the PWM signal in real time, with an accuracy of 0.1%. For example, for a 1 MHz PWM signal output by a radio frequency device, the system can detect a small change in the duty cycle from 50% to 50.1%.

[0079] Calculate the total harmonic distortion rate THD of the PWM signal through spectrum analysis to evaluate the uniformity of the device output energy and ensure the signal quality of the beauty equipment output. THD ≤ 3% indicates that the harmonic interference of the device output signal is low, avoiding unexpected electrical or thermal effects on the bionic skin model.

[0080] If it detects that the duty cycle fluctuation exceeds 0.5% or THD exceeds 3%, the host computer issues a warning, indicating that the device may have unstable output problems, and switches the mode or shuts down the beauty equipment to protect the skin tissue.

[0081] The electro-magnetic-thermal composite sensor array is the core component of the bionic skin dynamic simulation platform, which is used to collect, analyze, and process multi-dimensional data (electrical, magnetic, thermal parameters) of the beauty equipment output acting on the bionic skin model in real time. Through the high-precision sensor array for data collection, the comprehensive monitoring and feedback of the dynamic response of the bionic skin model are realized.

[0082] Perform multi-interaction fusion processing on the collected multi-dimensional data, that is, integrate electrical parameters (resistance, current, voltage), magnetic parameters (magnetic field strength, magnetic flux change), and thermal parameters (temperature, thermal conductivity) to generate a comprehensive multi-modal data report. The methods include: The report content includes three-dimensional electric field distribution map, three-dimensional magnetic field distribution map, three-dimensional thermal field distribution map, and key demonstration characteristics; The method for obtaining the three-dimensional electric field distribution map is to collect the spatial distribution data of electrical parameters through a microelectrode array (distributed in the epidermis layer, dermis layer, and subcutaneous tissue layer), and use the Kriging interpolation method or finite element analysis algorithm to construct the three-dimensional electric field distribution map of the bionic skin model. The electrical parameters include resistance, current, and voltage; the spatial distribution of electric field strength, current density, and resistivity can be intuitively displayed in the figure.

[0083] In the current stimulation beauty device test, the system evaluates the penetration depth and uniformity of the current in the bionic skin model through the electric field distribution map. The beauty device is tested for its effectiveness.

[0084] The magnetic field spatial distribution is used to monitor and analyze the magnetic field output by the beauty device (such as the alternating magnetic field generated by a radio frequency device) and the magnetic response of the bionic skin model (such as magnetic flux change, magnetic leakage distribution). The high-frequency electromagnetic field output by the beauty device (such as a radio frequency device) will generate an alternating magnetic field, which may affect the electrical and thermal characteristics of the bionic skin model. The TMR (tunnel magnetoresistance) sensor array is used to measure the magnetic field strength and magnetic flux change in real time, with a sensitivity of 1 nT, capable of capturing weak magnetic leakage signals.

[0085] For the magnetic field parameters collected by the (32-channel) TMR magnetoresistance sensor array, the magnetic field measurement range is ±100 μT, the resolution is 1 nT, suitable for detecting the alternating magnetic field generated by radio frequency devices (usually in the range of 10 nT to 10 μT); the sampling frequency is 100 Hz, and the resolution is 16 bits. The magnetic field parameters include magnetic field strength (B) and magnetic flux change (dB / dt).

[0086] In the radio frequency beauty device test, the alternating magnetic field distribution generated by the radio frequency electromagnetic field is monitored through the TMR sensor array. For example, if the magnetic field strength at the dermis junction is detected to exceed 50 μT (which may cause tissue overheating).

[0087] To visually analyze the spatial distribution of the magnetic field in the bionic skin model, use the magnetic tomography algorithm to construct a three-dimensional magnetic field distribution map of the measured magnetic field strength and magnetic flux change, and locate the abnormal magnetic leakage area.

[0088] The method for obtaining the three-dimensional magnetic field distribution map is as follows: denoise the magnetic field intensity and magnetic flux changes (magnetic field data) collected by the TMR sensor array. For example, use the Wavelet Transform algorithm to remove high-frequency noise to ensure data quality. Then, use the magnetic tomography algorithm to invert the spatial distribution of the magnetic field in the bionic skin model based on Maxwell's equations and finite element analysis. The input of the algorithm is the magnetic field intensity and magnetic flux changes, and the output is the three-dimensional magnetic field distribution map (resolution 1mm). Identify the abnormal areas (such as areas with sudden changes in magnetic field intensity or magnetic flux leakage) in the magnetic field distribution map through gradient analysis, and mark their spatial positions; achieve abnormal magnetic flux leakage positioning: the abnormal magnetic flux leakage area may correspond to uneven output of the beauty device or material defects inside the bionic skin model.

[0089] In the test of radio frequency beauty devices, evaluate the penetration depth and uniformity of radio frequency electromagnetic fields through the three-dimensional magnetic field distribution map. For example, if an abnormal magnetic flux leakage area (with a sudden change in magnetic field intensity exceeding 20%) is detected at the junction of the dermis layer, the device may have a problem of uneven output.

[0090] Thermal parameter analysis is used to monitor and analyze the heat generated by beauty devices in real time (such as the thermal effects generated by radio frequency, laser, and ultrasound) and the thermal responses of the bionic skin model (such as temperature distribution and heat conduction characteristics); the thermal responses of the bionic skin model are key indicators for evaluating the safety and effectiveness of beauty devices.

[0091] The method for obtaining the three-dimensional thermal field distribution map is as follows: through the Bayesian estimation algorithm, fuse the point measurement data of the thermocouple array with the surface temperature distribution map of the FLIR Lepton 3.5 infrared thermal imaging to generate a three-dimensional thermal field distribution map; the resolution of the fused thermal field map is 1mm, which can accurately reflect the process of heat conduction, diffusion, and dissipation in the bionic skin model.

[0092] In the test of laser beauty devices, evaluate the penetration depth and uniformity of laser heat in the bionic skin model through the three-dimensional thermal field distribution map. For example, if the temperature of the dermis layer exceeds 45°C (which may cause tissue damage), the beauty device has excessive energy and exceeds the standard.

[0093] The methods for obtaining key demonstration features include: Use a microelectrode array and an impedance analyzer to measure the complex impedance (including resistance and reactance) of the graphene heating layer. The impedance resolution reaches 0.1Ω, which can detect small impedance changes caused by temperature, humidity, or mechanical vibration. Analyze the impedance spectrum through Nyquist plots and Bode plots, extract the equivalent circuit parameters (such as resistance, capacitance, and inductance) of the graphene heating layer, and establish a time-series dynamic change curve based on the sampling frequency to evaluate the stability of its electrical performance; In the testing of radio frequency or current stimulation beauty devices, the electrical response of the graphene heating layer is monitored through impedance spectroscopy. For example, if the output of a radio frequency device causes the temperature of the graphene heating layer to increase, the impedance may decrease by approximately 5% (due to the increase in conductivity caused by the temperature increase). Simulate the heat loss effect of the skin.

[0094] The heat conduction characteristics (such as thermal conductivity, thermal diffusivity) of the bionic skin model will vary dynamically due to the heat output by the beauty device, the water content of the material layer, and the microfluidic flow rate. The dynamic analysis of heat conduction characteristics can evaluate the conduction behavior of heat in the bionic skin model.

[0095] Measure the temperature values of each layer at a fixed frequency through a thermocouple array, and combine with the Fourier heat conduction equation to calculate the thermal conductivity (unit: W / m·K) and thermal diffusivity (unit: m² / s) of the epidermis layer, dermis layer, and subcutaneous tissue layer; Use the finite element analysis algorithm to establish a heat conduction model of the bionic skin model. The input parameters include the thermal conductivity, water content, microfluidic flow rate, and fluid temperature of each layer (material), and the output is the heat conduction path (curve) and heat flux density distribution (distribution map); In the testing of cold and heat therapy beauty devices, evaluate the thermal effects of cold and heat stimulation on the bionic skin model through dynamic analysis of heat conduction characteristics. For example, if the thermal diffusivity of the subcutaneous tissue layer is detected to be too low in the cooling mode (which may cause tissue frostbite).

[0096] The electromagnetic field output by the beauty device usually contains both electric field and magnetic field components, and the two may produce a coupling effect (such as electromagnetic induction, eddy current effect) in the bionic skin model. To comprehensively evaluate the influence of the electromagnetic field, perform a coupling analysis of the magnetic field and the electric field.

[0097] For the data collected by the TMR magnetoresistive sensor array and the microelectrode array, use the Kalman filter algorithm to eliminate noise and interference and improve data accuracy; establish an electromagnetic field coupling model based on Maxwell's equations to calculate the interaction between the electric field and the magnetic field in the bionic skin model. For example, the alternating magnetic field output by a radio frequency device may induce eddy currents in the dermis layer, thereby generating a local thermal effect. Through the coupling model, extract key parameters such as eddy current density (unit: A / m²), electromagnetic induction intensity (unit: V / m), and local thermal power density (unit: W / m³), and establish a time series dynamic change curve according to the sampling frequency; In the testing of radio frequency beauty devices, evaluate the thermal effects of eddy current effects on the bionic skin model through electromagnetic field coupling analysis. For example, if the eddy current density in the dermis layer is detected to be too high, it may cause the skin tissue to overheat.

[0098] The heat output by the beauty device may cause thermal stress and mechanical deformation (such as expansion and contraction) in the material layer of the bionic skin model, simulating the thermo-mechanical response of skin tissue. The thermal stress and mechanical deformation are used to evaluate the impact of heat on the structural stability of the bionic skin model through a sensor array and mechanical modeling.

[0099] Use the finite element analysis algorithm to establish a thermo-mechanical coupling model of the bionic skin model. The input parameters include the thermal expansion coefficient, elastic modulus, and thermal stress distribution of each layer of material, and the output is the deformation distribution map and the stress concentration area. In the radiofrequency beauty device test, the impact of heat on the structure of the bionic skin model is evaluated through thermal stress and mechanical deformation analysis. For example, if it is detected that the dermis layer has excessive deformation due to heat (which may cause tissue damage), the host computer will reduce the microfluidic flow rate through an external control system to simulate the stress relaxation of skin tissue.

[0100] The detection of beauty devices is set to include functional evaluation and safety evaluation, and quantitative indicators for functional evaluation and safety evaluation are defined according to the mode of the beauty device. For example, for the quantitative indicators of functional evaluation, the temperature increase range in the dermis layer in the radiofrequency mode (target: 40 - 45 °C), the collagen fiber density increase rate of 10%, the fat decomposition rate of 5% in the subcutaneous tissue layer in the ultrasonic mode, the mechanical vibration intensity of 2 Pa, and the optical response intensity of 3 for the photosensitive material in the epidermis layer and the pigment decomposition rate of 7% in the laser mode. For the quantitative indicators of safety evaluation, the temperature of the epidermis layer does not exceed 45 °C to avoid burns, the stress of the dermis layer does not exceed 2 MPa to avoid collagen fiber breakage, the vibration intensity of the subcutaneous tissue layer does not exceed 0.5 m / s² to avoid fat tissue damage, and the electromagnetic field intensity complies with the standards of the International Commission on Non-Ionizing Radiation Protection (ICNIRP).

[0101] Functional evaluation is to comprehensively evaluate the functional performance of beauty devices in different working modes (such as the thermal, electrical, and mechanical responses of radiofrequency, ultrasonic, laser, etc. to the skin) through multi-modal data reports, and compare with the expected beauty effects (such as firming, repair, anti-aging). Safety evaluation is based on the dynamic response of the bionic skin model to detect whether the output of the beauty device is within the safe range to avoid damage to skin tissue (such as overheating, excessive mechanical vibration, electromagnetic interference, etc.).

[0102] The stability, repeatability, environmental adaptability of the device and the optimization of user experience can also be added.

[0103] For multi-modal data reports, a hierarchical dynamic selection architecture is used to extract descriptive primitives to represent data features, a predefined machine learning model is used to learn the descriptive primitives, and quantitative indicators are output. A causal chain is introduced into the machine learning model (analyzing the causal reasoning between data). Calculate the deviation value between the calculated quantitative index and the predefined effect index; Deploy multi-objective comprehensive evaluation, perform weighted averaging on the quantitative indicators to obtain a comprehensive score, and predefined hard constraints for the comprehensive score and the deviation value (such as safety score SS, SS ≥ 90 points, temperature deviation value CPS ≤ 3). If the beauty device meets all hard constraints, it is determined to be qualified and can be put into mass production; otherwise, it is determined to be unqualified and further design improvement or re-testing of the experiment is required.

[0104] Among them, the input data of the causal chain comes from a multi-modal data report, including three-dimensional electric field distribution maps, three-dimensional magnetic field distribution maps, three-dimensional thermal field distribution maps, and key demonstration characteristics; a hierarchical dynamic selection architecture is adopted, which is divided into a bottom causal layer, a middle causal layer, and a top causal layer; In the bottom causal layer, the heat output by the beauty device affects the surface temperature distribution map collected by FLIR Lepton3.5 infrared thermal imaging through the deformation of the polymer and hydrophobic coating on the epidermis, and then affects the three-dimensional thermal field distribution map. The current stimulation affects the complex impedance measured by the microelectrode array through the change of the elastic modulus and water content of the hydrogel substrate in the dermis, and then affects the three-dimensional electric field distribution map. The electromagnetic affects the change of the magnetic field strength and magnetic flux measured by the TMR magnetoresistive sensor array through the change of the thermal conductivity of the fat simulation gel in the subcutaneous tissue layer, and then affects the three-dimensional magnetic field distribution map. The mechanical vibration affects the temperature values of each layer measured by the thermocouple array by reducing the microfluidic flow rate through the external control system, and then affects the three-dimensional thermal field distribution map; In the middle causal layer, the deformation of the epidermis affects the deformation distribution map and stress concentration area of the thermo-mechanical coupling model by measuring the local stress change through the micro strain sensor. The change of the elastic modulus and water content of the dermis affects the three-dimensional thermal field distribution map by measuring the temperature values of each layer through the thermocouple array. The change of the microfluidic chip flow rate and fluid temperature affects the heat conduction characteristics by adjusting the capillary network density through the external control system, and then affects the three-dimensional thermal field distribution map by fusing the thermocouple array data and the infrared thermal imaging data through the Bayesian estimation algorithm. When the temperature of the dermis is too high, the initial temperature or water content is adjusted through the Peltier effect device or the micro humidifier to affect the heat conduction characteristics, and then affects the three-dimensional thermal field distribution map; In the top causal layer, the electrical parameters of the three-dimensional electric field distribution map affect the three-dimensional magnetic field distribution map through Maxwell's equations. The abnormal areas of the three-dimensional magnetic field distribution map are identified through gradient analysis and feedback through the electromagnetic field coupling model to affect the three-dimensional electric field distribution map. The three-dimensional thermal field distribution map calculates the thermal conductivity and thermal diffusivity through the Fourier heat conduction equation to affect the heat conduction path and heat flux density distribution of the heat conduction model, and then affects the deformation distribution map and stress concentration area of the thermo-mechanical coupling model.

[0105] Establish an electro-magnetic-thermal-mechanical coupling model through the finite element analysis algorithm. The input parameters include thermal conductivity, water content, microfluidic flow rate, fluid temperature, coefficient of thermal expansion, elastic modulus, thermal stress distribution, resistivity, conductivity, magnetic field strength, and magnetic flux change, and the comprehensive causal influence diagram is output. Extract the time-series dynamic change curve through the time-series Transformer structure, identify the key causal nodes through the attention mechanism, and eliminate noise and interference through the Kalman filter algorithm. The output of the causal chain is verified through the pre-inspection safety platform. If the duty cycle fluctuation exceeds 0.5% or the THD exceeds 3%, the host computer issues a warning and adjusts the parameters of the bionic skin model for the functional evaluation and safety evaluation of beauty devices.

[0106] Because in the multi-modal data report, there are too many forms of data presentation, including not only numerical values but also curves and charts. Therefore, hierarchical processing is required to improve the efficiency of feature extraction.

[0107] The methods for extracting descriptive primitives to represent data features using a hierarchical dynamic selection architecture include: For the numerical data (such as resistivity and thermal conductivity) in the multi-modal data report, after normalization, it is processed using a multi-layer perceptron (MLP). For the curve data (such as the time-series dynamic change curve) in the multi-modal data report, first perform dynamic time warping (DTW) to synchronously align with the time slices of the 3D graph data. Adopt the time-series Transformer structure, and capture the long-term dependencies in the curve data (such as temperature fluctuation period and electromagnetic field transient change) through the self-attention mechanism, and output a sequence of feature vectors in the time dimension. For the 3D graph data (such as the 3D thermal field distribution map) in the multi-modal data report, use the numerical data as the physical benchmark for cross-modal alignment. Because it has clear biomedical significance (such as "the average temperature of the epidermis layer is 42°C"), it can effectively anchor the scale reference system of spatio-temporal data. Expand the numerical data into the query vector (Query) of the attention mechanism, which represents the core index of the skin state. Concatenate the time-series features (time curve evolution pattern) and spatial features (3D field distribution gradient) into the key (Key) and value (Value) to construct a spatio-temporal joint representation. Develop a two-channel parallel architecture based on sparsity detection. If the sparsity > 70%, preferentially enable CNN local feature extraction (retaining spatial correlation). If the sparsity < 30%, enable PCA global dimensionality reduction (retaining the principal components with 95% variance). The boundary of sparsity judgment can be adjusted adaptively, where sparsity = total number of elements / number of non-zero elements, and use the gradient penalty term to constrain the feature redundancy. , θij is the angle between feature vectors i and j, representing the correlation between features (θij → 0 indicates high correlation), and λ is the regularization coefficient (usually taken as 0.01 ∼ 0.1), which controls the penalty intensity. By penalizing highly correlated feature pairs, the network is forced to learn orthogonalized feature representations, solving the problem of information redundancy in multimodal data.

[0108] Among them, for CNN, the channel importance score is added after each convolutional layer, and the channels with importance scores lower than the expected performance are closed. For example, the importance score , dynamically close < 0.1·max( ) channels, reducing the computational amount by 40%. Octree adaptive sampling is used, where represents the batch size, represents the feature map of the c-th channel in the n-th sample; For example, for a three-dimensional thermal field distribution map, the sampling rate , maintaining high resolution in areas with large temperature gradients and downsampling in flat areas, reducing the computational amount by 65%; is the magnitude of the temperature gradient (unit: °C / mm), 1 / 8 means keeping 1 out of every 8 voxels (downsampling rate of 87.5%), and 1 means full-resolution sampling (no downsampling); The features extracted from numerical data, curve data, and three-dimensional map data are recorded as description primitives.

[0109] Example: For different modal data, description primitives are extracted. For example: Electric field data: Extract the peak value, mean value, variance, etc. of resistivity, conductivity, and current density distribution.

[0110] Magnetic field data: Extract the magnetic field strength, gradient of magnetic flux change, spatial position of abnormal regions, etc.

[0111] Thermal field data: Extract thermal conductivity, thermal diffusivity, temperature gradient, heat flux density, etc.

[0112] Mechanical data: Extract stress concentration regions, maximum values of deformation distribution, changes in elastic modulus, etc.

[0113] The machine learning model can be based on RNN, deep neural learning network, etc. Use pre-designed data for pre-training to obtain a machine learning model that meets the training standards.

[0114] The host computer is developed based on an open technology stack and is used to provide a user interface (GUI), including setting detection parameters, displaying real-time data, and generating detection reports, supporting data visualization (such as thermal field distribution maps, electromagnetic time series curves) and data storage.

[0115] Among them, the open technology stack includes open-source systems (Linux kernel, Android framework), cross-platform development tools (Python, LabVIEW), and standardized interfaces (Python's library ecosystem, LabVIEW's hardware drivers); Furthermore, it includes an Android intelligent interaction layer (based on the Linux kernel), a Python-LabVIEW hybrid development framework, supports modular expansion and cross-platform deployment. When making the top-level design, an open heterogeneous system architecture is recommended. For the development layer, a cross-platform development suite (Python / LabVIEW) is recommended, and for the runtime layer, a Linux / Android dual-mode runtime environment is recommended.

[0116] For example, the three-dimensional magnetic field distribution map is presented in the form of a heat map or a vector map. The color depth represents the magnetic field strength, and the arrow represents the magnetic field direction. Users can select a specific area through the interaction interface to view detailed magnetic field parameters (such as magnetic field strength, magnetic flux change rate, or expression of the time-series curve).

[0117] Embodiment 2 Please refer to Figure 1 and Figure 3 As shown, for the parts not described in detail in this embodiment, refer to the description in Embodiment 1. A detection method for an electrothermal medical beauty device is provided, including: Step S1: Construct a bionic skin model. The thickness of each layer and the density of the capillary network are dynamically adjusted according to the age group. Set an external control interface to connect to an external control system to simulate the skin state under different environmental conditions. Pre-define the area and concentration of the photosensitive material, blood flow rate, and temperature. Use a Peltier effect device to control the initial temperature, and adjust the water content through a micro-humidifier and a desiccant. Dynamically adjust the parameters of the bionic skin (such as microfluidic flow rate, skin temperature, capillary network density, etc.) through the external control system to simulate the skin state under different environmental conditions. At the same time, preset a biochemical protection mechanism. For example, when the temperature of the dermis layer is too high, lower the temperature or increase the water content through the Peltier effect device or the micro-humidifier. When the mechanical vibration of the subcutaneous tissue layer is too strong, reduce the microfluidic flow rate to simulate the self-protection and stress relaxation of the skin. This dynamic regulation ability enables the system to adapt to the working modes of different types of beauty devices (such as radio frequency, ultrasonic, cold and heat therapy, etc.), and truly simulate the skin's reaction under extreme conditions, significantly improving the adaptability and practicality of the system.

[0118] Step S2: Collect multi-dimensional data through a high-resolution camera array and an electro-magnetic-thermal composite sensor array, and transmit it to the host computer and the main measurement center; Step S3: Perform stable pre-inspection on the output of the beauty device through the pre-inspection safety platform; perform stable pre-inspection on the output of the beauty device through the pre-inspection safety platform, and use a high-precision analog-to-digital converter (ADC) combined with the fast Fourier transform algorithm to calculate the total harmonic distortion rate (THD) of the PWM signal. If it is detected that the duty cycle fluctuation exceeds 0.5% or the THD exceeds 3%, the upper computer will issue a warning and take measures such as switching modes or shutting down the device. This function effectively avoids detection deviations or safety hazards caused by unstable device output, ensuring the reliability and safety of the detection process, and is especially applicable to device verification before mass production.

[0119] Step S4: The main measurement center uses the multi-branch shunt method to perform multi-interaction fusion processing on multi-dimensional data to generate a multi-modal data report; Step S5: Extract description primitives using a hierarchical dynamic selection architecture, use a machine learning model to learn and output quantization metrics, introduce causal chain analysis, calculate the deviation value between the quantization metrics and predefined effect metrics, conduct multi-objective comprehensive evaluation, and determine whether the device meets the hard constraints. If it is qualified, mass production can be carried out; otherwise, the design needs to be improved or re-inspected; through multi-objective comprehensive evaluation, the quantization metrics are weighted and averaged to obtain a comprehensive score, and hard constraints are predefined for the score and deviation value to determine whether the device meets the functional and safety requirements. If it is qualified, it can be put into mass production; otherwise, the design needs to be improved or re-inspected. This standardized evaluation method avoids the arbitrariness of subjective judgment in traditional inspections, provides an objective and scientific decision-making basis for device mass production, and significantly improves production efficiency and product quality.

[0120] Step S6: The upper computer conducts visual interaction. The user-friendly interface design and powerful data management functions facilitate R & D personnel to intuitively analyze the detection results, optimize the device design, and at the same time facilitate long-term storage and traceability of data, meeting the requirements of industrial production and quality control.

[0121] Embodiment 3 This embodiment publicly provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, it implements the operation mode of the detection system of the above-provided electrothermal medical beauty device.

[0122] Since the electronic device introduced in this embodiment is the electronic device adopted by the detection system of an electrothermal medical beauty device in the embodiments of the present application, based on the detection system of an electrothermal medical beauty device introduced in the embodiments of the present application, those skilled in the art can understand the specific implementation manners and various variations of the electronic device in this embodiment. Therefore, the specific implementation of how this electronic device implements the method in the embodiments of the present application will not be described in detail here. As long as those skilled in the art implement the electronic device adopted by the detection system of an electrothermal medical beauty device in the embodiments of the present application, it falls within the scope of protection of the present application.

[0123] The above formulas are all dimensionless and take their numerical values for calculation. The formulas are obtained by collecting a large amount of data for software simulation to obtain a formula that is closest to the actual situation. The preset parameters and threshold selection in the formulas are set by those skilled in the art according to the actual situation.

[0124] The above are only the preferred implementation manners of the present invention. The protection scope of the present invention is not limited to the above embodiments. All technical solutions falling within the idea of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technical users in the technical field, several improvements and refinements made without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A detection system for an electrothermal medical beauty device, characterized in that It includes a bionic skin model and a main measurement center; The bionic skin model is constructed based on multi-layer composite materials and dynamic regulation technologies, and uses a three-layer gradient structure to simulate bionic skin, with the layers bonded by gradient bonding technology; And it simulates the skin state under different environmental conditions through an external control system; Deploy an integrated bionic microcirculation physiological evolution, including embedding photosensitive materials, microfluidic chips and an integrated graphene heating layer; Deploy a high-resolution camera array outside the bionic skin model, and integrate a high-precision electro-magnetic-thermal composite sensor array inside to form a three-dimensional sensing network, collect multi-dimensional data and transmit it to the host computer and the main measurement center; The main measurement center includes a pre-inspection safety station and a fusion detection module. The pre-inspection safety station is used for stable pre-inspection of the output of beauty equipment; The fusion detection module uses a multi-branch shunt method to perform multi-interaction fusion processing on the collected multi-dimensional data, generates a multi-modal data report, pre-defines and quantifies the detection indicators of beauty equipment, uses a machine learning model to deeply learn the multi-modal data report and then outputs quantified indicators, and introduces a causal chain in the machine learning model. The causal chain adopts a hierarchical dynamic selection architecture, which is divided into a bottom causal layer, a middle causal layer and a top causal layer; Calculate and determine whether the beauty equipment is qualified according to the quantified indicators, and send it to the host computer, which is used for visual interaction and data storage.

2. The detection system of the electrothermal medical beauty device according to claim 1, characterized in that, The method for constructing the bionic skin model based on multi-layer composite materials and dynamic regulation technologies includes: Use a three-layer gradient structure to simulate bionic skin, namely the epidermis layer, the dermis layer and the subcutaneous tissue layer; The thickness of each layer is dynamically adjusted according to the age group, and the layers are bonded by gradient bonding technology; The epidermis layer uses a polymer to simulate the stratum corneum, and a hydrophobic coating is applied on the surface to simulate the skin barrier function; The dermis layer uses a hydrogel substrate to simulate the collagen fiber density and elastic modulus; The subcutaneous tissue layer uses fat to simulate gel to regulate the water content and thermal conductivity; Set an external control interface and connect it to the external control system to simulate the skin state under different environmental conditions; Embed photosensitive materials in the epidermis layer or the dermis layer, and pre-define the area and concentration of the photosensitive materials; Deploy an integrated bionic microcirculation physiological evolution, including embedding a microfluidic chip and an integrated graphene heating layer between the dermis layer and the subcutaneous tissue layer. The microfluidic chip is used to simulate the capillary network, the network density is adjusted according to the age group, set a fluid medium to simulate blood, and the flow rate is adjusted by programming; The integrated graphene heating layer is used to simulate the electrical and thermal responses of skin tissue; The microfluidic chip is driven by an external peristaltic pump. By adjusting the flow rate and fluid temperature, it dynamically simulates the heat conduction characteristics of the skin under different environmental temperatures or exercise states; Adopt a Peltier effect heating / cooling device to control the initial temperature of the bionic skin; Adjust the water content of the bionic skin through a micro-humidifier and a desiccant; Conduct electrical property simulation on the bionic skin, including setting different resistivity and conductivity according to the characteristics of the skin layer; Preset the microfluidic flow rate and skin temperature according to requirements, simulate the skin state under different environmental conditions, and dynamically adjust the parameters of the bionic skin according to the output of the beauty equipment.

3. The detection system of the electrothermal medical beauty device according to claim 2, wherein Dynamically adjusting the parameters of the bionic skin according to the output of the beauty device, including: The predetermined working outputs of the beauty device include radio frequency, ultrasonic, laser, electrical stimulation, and cold and heat therapies; For the heat, current, electromagnetic, or mechanical vibration output by the beauty device, the polymer and hydrophobic coating of the epidermis layer generate deformation or surface property changes; the hydrogel substrate of the dermis layer generates changes in elastic modulus and water content; the fat-mimicking gel of the subcutaneous tissue layer generates changes in thermal conductivity and water content; the microcirculation dynamically adjusts the flow rate and fluid temperature through an external control system; the capillary network density is dynamically adjusted through an external control system to simulate the dynamic process of skin tissue repair or aging; Simulating the resistivity and conductivity of each layer of the bionic skin; the sensor array in the electromagnetic-thermal composite sensor array real-time collects multi-dimensional data of the output of the beauty device acting on the bionic skin model; Presetting a biochemical protection mechanism, including when the temperature of the dermis layer is too high, reducing the initial temperature or increasing the water content through a Peltier effect device or a micro-humidifier to simulate the self-protection of the skin; if the mechanical vibration of the subcutaneous tissue layer is too strong, reducing the microfluidic flow rate through an external control system to simulate the stress relaxation of the skin tissue.

4. The detection system of the electrothermal medical beauty device according to claim 3, characterized in that, The integrated deployment method of the high-precision electro-magnetic-thermal composite sensor array includes: Deploying a high-resolution camera array outside the bionic skin model to comprehensively monitor the surface changes of the bionic skin model; embedding micro strain sensors in the dermis layer and the subcutaneous tissue layer to measure the local stress changes caused by heat; For resistance, current, and voltage, an electrical variable sensor combines a microelectrode array and an impedance analyzer for measurement; introducing a wide-band frequency sweep technology; Using a TMR magnetoresistive sensor array to measure the magnetic field strength and magnetic flux changes; for the temperature of each layer, a thermal field sensor uses a thermocouple array for measurement; using a FLIR Lepton 3.5 infrared thermal imager to real-time collect the surface temperature distribution map; The sensor array is evenly distributed in the epidermis layer, dermis layer, and subcutaneous tissue layer of the bionic skin at a spacing of 5 mm to form a three-dimensional sensing network, and the number of sensors is determined according to the size of the test area; it is connected to a data acquisition card through a multiplexer, and the data acquisition card uses a high-precision analog-to-digital converter and is transmitted to the upper computer and the main measurement center in real-time through USB or Wi-Fi.

5. The detection system of the electrothermal medical beauty device according to claim 4, characterized in that, The pre-inspection safety station is used to stably pre-inspect the output of the beauty device, and the method includes; The pre-inspection safety station uses a high-precision analog-to-digital converter ADC, combines the fast Fourier transform algorithm, and calculates the total harmonic distortion rate THD of the PWM signal through spectrum analysis; If it is detected that the duty cycle fluctuation exceeds 0.5% or the THD exceeds 3%, the upper computer issues a warning and switches the mode or shuts down the beauty device.

6. The detection system of the electrothermal medical beauty device according to claim 5, characterized in that, The module uses a multi-branch method to perform multi-interaction fusion processing on the collected multi-dimensional data to generate a multi-modal data report, and the method includes: The report content includes three-dimensional electric field distribution map, three-dimensional magnetic field distribution map, three-dimensional thermal field distribution map, and key demonstration characteristics; The method for obtaining the three-dimensional electric field distribution map is as follows: collect the spatial distribution data of electrical parameters through a microelectrode array, and use the Kriging interpolation method or the finite element analysis algorithm to construct the three-dimensional electric field distribution map of the bionic skin model. The electrical parameters include resistance, current, and voltage; The method for obtaining the three-dimensional magnetic field distribution map is as follows: denoise the magnetic field intensity and magnetic flux change collected by the TMR sensor array, adopt the magnetic tomography algorithm, and based on Maxwell's equations and finite element analysis, invert the spatial distribution of the magnetic field in the bionic skin model. The input of the algorithm is the magnetic field intensity and magnetic flux change, and the output is the three-dimensional magnetic field distribution map. Identify the abnormal areas in the magnetic field distribution map through the gradient analysis method and mark their spatial positions; The method for obtaining the three-dimensional thermal field distribution map is as follows: through the Bayesian estimation algorithm, fuse the point measurement data of the thermocouple array with the surface temperature distribution map of the FLIR Lepton 3.5 infrared thermal imaging to generate the three-dimensional thermal field distribution map; The methods for obtaining the key demonstration features include: Use a microelectrode array and an impedance analyzer to measure the complex impedance of the graphene heating layer, analyze the impedance spectrum through the Nyquist diagram and Bode diagram, extract the equivalent circuit parameters of the graphene heating layer, and establish a time-series dynamic change curve according to the sampling frequency; Measure the temperature values of each layer at a fixed frequency through the thermocouple array, and combine with the Fourier heat conduction equation to calculate the thermal conductivity and thermal diffusivity of the epidermis layer, dermis layer, and subcutaneous tissue layer; Use the finite element analysis algorithm to establish the heat conduction model of the bionic skin model. The input parameters include the thermal conductivity, water content, microfluidic flow rate, and fluid temperature of each layer, and the output is the heat conduction path and heat flux density distribution; For the data collected by the TMR magnetoresistive sensor array and the microelectrode array, use the Kalman filtering algorithm to eliminate noise and interference; establish an electromagnetic field coupling model based on Maxwell's equations, extract the key parameters through the coupling model, and establish a time-series dynamic change curve according to the sampling frequency; Use the finite element analysis algorithm to establish the thermal-mechanical coupling model of the bionic skin model. The input parameters include the thermal expansion coefficient, elastic modulus, and thermal stress distribution of each layer of material, and the output is the deformation distribution map and stress concentration area.

7. The detection system of the electrothermal medical beauty device according to claim 6, characterized in that, The method for calculating and determining whether the beauty device is qualified according to the quantization index includes: Set the detection of the beauty device to include functional evaluation and safety evaluation, and define the quantization indexes of functional evaluation and safety evaluation according to the mode of the beauty device; For the multi-modal data report, adopt a hierarchical dynamic selection architecture to extract the description primitives for representing the data features, use a pre-defined machine learning model to learn the description primitives, and output the quantization index, and introduce a causal chain in the machine learning model; Calculate the deviation value between the quantization index and the pre-defined effect index; Deploy a multi-objective comprehensive evaluation, perform a weighted average of the quantization indexes to obtain a comprehensive score, pre-define hard constraints for the comprehensive score and the deviation value. If the beauty device meets all the hard constraints, it is determined to be qualified and can be put into mass production; otherwise, it is determined to be unqualified and needs to further improve the design or re-detect the experiment; Among them, the causal chain adopts a hierarchical dynamic selection architecture, which is divided into a bottom-layer causal layer, a middle-layer causal layer, and a top-layer causal layer; In the bottom-layer causal layer, the heat output by the beauty device mode affects the three-dimensional thermal field distribution map through the deformation of the polymer and hydrophobic coating in the epidermis layer. The current stimulation affects the three-dimensional electric field distribution map through the changes in the elastic modulus and water content of the hydrogel substrate in the dermis layer. The electromagnetic field affects the changes in the magnetic field strength and magnetic flux through the changes in the thermal conductivity of the fat simulation gel in the subcutaneous tissue layer, and then affects the three-dimensional magnetic field distribution map. The mechanical vibration reduces the microfluidic flow rate through the external control system to affect the temperature values of each layer, and then affects the three-dimensional thermal field distribution map; In the middle-layer causal layer, the deformation of the epidermis layer affects the deformation distribution map and stress concentration area of the thermo-mechanical coupling model through local stress changes. The changes in the elastic modulus and water content of the dermis layer affect the three-dimensional thermal field distribution map through the temperature values of each layer. The changes in the microfluidic chip flow rate and fluid temperature affect the heat conduction characteristics through the capillary network density, and then affect the three-dimensional thermal field distribution map by fusing the thermocouple array data and infrared thermal imaging data through the Bayesian estimation algorithm. When the temperature of the dermis layer is too high, the initial temperature or water content is adjusted through the Peltier effect device or humidifier to affect the heat conduction characteristics, and then affect the three-dimensional thermal field distribution map; In the top-layer causal layer, the electrical parameters of the three-dimensional electric field distribution map affect the three-dimensional magnetic field distribution map through Maxwell's equations. The abnormal areas of the three-dimensional magnetic field distribution map are identified through gradient analysis and feedback through the electromagnetic field coupling model to affect the three-dimensional electric field distribution map. The three-dimensional thermal field distribution map calculates the thermal conductivity and thermal diffusivity through Fourier's heat conduction equation to affect the heat conduction path and heat flux density distribution of the heat conduction model, and then affects the deformation distribution map and stress concentration area of the thermo-mechanical coupling model.

8. The detection system of the electrothermal medical beauty device according to claim 7, characterized in that, The method for extracting descriptive primitives using a hierarchical dynamic selection architecture to represent data features includes: For the numerical data in the multi-modal data report, after normalization, it is processed using a multi-layer perceptron; For the curve data in the multi-modal data report, first perform dynamic time warping to synchronously align with the time slices of the three-dimensional map data; adopt a temporal Transformer structure to output a sequence of feature vectors in the time dimension; For the three-dimensional map data in the multi-modal data report, use the numerical data as the physical benchmark for cross-modal alignment, expand the numerical data into the query vector of the attention mechanism, splice the temporal features and spatial features into keys and values to construct a spatio-temporal joint representation; develop a two-channel parallel architecture based on sparsity detection. If the sparsity > 70%, give priority to enabling CNN local feature extraction. If the sparsity < 30%, enable PCA global dimensionality reduction, and use a gradient penalty term to constrain the feature redundancy; Among them, for CNN, add a channel importance score after each convolutional layer, close the channels with importance scores lower than the expected performance, and use octree adaptive sampling; Record the features extracted from the numerical data, curve data, and three-dimensional map data as descriptive primitives.

9. The detection system of the electrothermal medical beauty device according to claim 8, characterized in that, The method for the host computer to perform visual interaction and data storage includes: The host computer is developed based on an open technology stack and is used to provide a user interface, including setting detection parameters, displaying real-time data, and generating detection reports, and supports data visualization and data storage.

10. A detection method for an electrothermal medical beauty device, applied to the detection system of the electrothermal medical beauty device according to any one of claims 1-9, characterized in that, The detection method of the electrothermal medical beauty device includes: Step S1: Construct a bionic skin model, with the thickness of each layer and the density of the capillary network dynamically adjusted according to age. Set an external control interface to connect to an external control system to simulate the skin state under different environmental conditions. Pre-define the area and concentration of the photosensitive material, blood flow rate, and temperature. Use a Peltier effect device to control the initial temperature, and adjust the water content through a micro humidifier and a desiccant; Step S2: Collect multi-dimensional data through a high-resolution camera array and an electro-magnetic-thermal composite sensor array, and transmit it to the host computer and the main measurement center; Step S3: Perform stable pre-inspection on the output of the beauty device through a pre-inspection safety platform; Step S4: The main measurement center uses a multi-branch shunt method to perform multi-interaction fusion processing on the multi-dimensional data and generate a multi-modal data report; Step S5: Adopt a hierarchical dynamic selection architecture to extract descriptive primitives, use a machine learning model to learn and output quantitative indicators, introduce causal chain analysis, calculate the deviation value between the quantitative indicators and the pre-defined effect indicators, conduct multi-objective comprehensive evaluation, and determine whether the device meets the hard constraints. If it is qualified, it can be mass-produced; otherwise, the design needs to be improved or re-inspected; Step S6: The host computer conducts visual interaction.

Images