Method for protecting against electromagnetic exposure of the human body based on optimization of medical implant structures

Abstract

The present application belongs to the technical field of human electromagnetic exposure safety protection, and especially is a human electromagnetic exposure safety protection method based on medical implant structure optimization. The method comprises the following steps: S1: wireless power transmission (WPT) system and human body modeling containing medical implants; S2: human electromagnetic exposure uncertainty evaluation framework modeling; S3: medical implant structure optimization. The Bayesian neural network proposed in the present application can greatly reduce the uncertainty quantification calculation cost. In addition, the medical implant occluder structure optimization design method proposed in the present application can effectively alleviate the safety risk of human electromagnetic exposure. Under the influence of uncertain factors, the probability of exceeding the limit of human electromagnetic exposure indicators in each scene is significantly reduced. The sparrow search algorithm is used to optimize the size parameters of the occluder in the present application, so as to reduce the electromagnetic exposure dose in the human body, and finally effectively reduce the probability of exceeding the electromagnetic exposure indicators of the patient containing the occluder implant under the influence of uncertain factors.

Description

Technical Field

[0001] This invention relates to the field of electromagnetic exposure safety protection technology, specifically a method for electromagnetic exposure safety protection of the human body based on the structural optimization of medical implants. Background Technology

[0002] With the increasing popularity of electric vehicles (EVs), the frequency of use of electric vehicle wireless power transfer (EV-WPT) systems is gradually increasing. EV-WPT relies on alternating magnetic fields in space to achieve magnetic coupling and resonant energy transfer. Because the magnetic field generated by the transmitting coil cannot be completely coupled to the receiving coil, leakage magnetic fields are inevitable, posing an electromagnetic exposure safety threat to nearby personnel such as drivers and passengers. Cardiovascular disease is a common disease in human health, and the proportion of people receiving cardiovascular medical implants has been rising in recent years. Among them, cardiac occluders, as a mainstream minimally invasive interventional medical device, have been widely used for the treatment of heart defects. Occluders are usually made of nickel-titanium alloy, whose conductivity is much higher than that of human tissue. When patients with occluders are exposed to the electromagnetic field environment of an EV-WPT system, induced currents and field strength accumulation are easily formed in their bodies, significantly increasing the electromagnetic exposure safety risk.

[0003] Currently, most EV-WPT system electromagnetic exposure safety assessment technologies only consider deterministic operating conditions, neglecting actual uncertainties such as WPT system manufacturing errors, installation offsets, and fluctuations in human tissue parameters. This results in assessment results that cannot effectively reflect reality. Regarding uncertainty quantification methods, traditional Monte Carlo methods rely on large-scale samples, leading to enormous computational time; multinomial chaotic expansion surrogate model methods suffer from the curse of dimensionality. To address these shortcomings, this invention proposes for the first time the application of Bayesian neural networks (BNNs) to human electromagnetic exposure safety uncertainty assessment, enabling accurate prediction of the probability of exceeding limits for electromagnetic exposure indicators within the human body under small sample conditions.

[0004] Furthermore, existing electromagnetic exposure safety technologies for people with medical implants are limited to electromagnetic shielding structure design and system topology optimization, with very little parameter optimization for the geometric structure of the medical implant itself. This invention innovatively proposes a occluder implant structure optimization method based on the sparrow search algorithm. By optimizing the key dimensional parameters of the occluder (diameter of the upper and lower plates, waist diameter, and height), it significantly reduces the electromagnetic exposure dose and the risk of exceeding the standard, providing a scientific guide for electromagnetic exposure protection of people with medical implants in the electromagnetic field environment of the EV-WPT system. Summary of the Invention

[0005] The purpose of this section is to outline some aspects of the embodiments of the present invention and to briefly describe some preferred embodiments. Simplifications or omissions may be made in this section, as well as in the abstract and title of this application, to avoid obscuring the purpose of these documents; however, such simplifications or omissions should not be construed as limiting the scope of the invention.

[0006] To address the aforementioned technical problems, according to one aspect of the present invention, the present invention provides the following technical solution:

[0007] A method for safe protection against electromagnetic exposure of the human body based on the structural optimization of medical implants includes the following steps:

[0008] S1: WPT system and human modeling with implants: Establish an electromagnetic numerical model of the wireless power transmission system of electric vehicles, as well as an adult male human model containing a metal medical implant occluder, and construct three typical daily high-risk exposure scenarios: the person is located outside the rear of the vehicle, the person is located outside the vehicle, and the person is located inside the rear seat of the vehicle, to provide a basis for subsequent human electromagnetic exposure assessment and analysis.

[0009] S2: Uncertainty Assessment Framework Modeling for Human Electromagnetic Exposure: An uncertainty quantification method based on Bayesian neural networks is employed, using random variables such as the parameters of the electric vehicle's wireless power transmission system and the electromagnetic properties of human tissue as inputs, with the maximum induced electric field E in the human body as the starting point. max As output, by efficiently solving the posterior probability distribution of the network model parameters, the human body E under various scenarios is calculated. max Probability of exceeding the standard;

[0010] S3: Medical Implant Structure Optimization: Incorporating Human E max As the optimization objective, the geometric structural parameters of the occluder are used as optimization variables. These parameters include the diameters of the upper and lower plates, the waist diameter, and the height. A sparrow search algorithm is used for global optimization to obtain the optimal design solution for the occluder. Uncertainty assessment is then used to verify the optimized occluder structure and its effect on the human body (E). max Probability of exceeding the standard.

[0011] As a preferred embodiment of the human electromagnetic exposure safety protection method based on medical implant structure optimization described in this invention, in step S1, the electromagnetic numerical model of the electric vehicle wireless power transmission system includes an LCC-LCC compensation topology, coil and ferrite structure, and the working frequency is set to 85kHz; three typical daily high-risk exposure scenarios include the human body being located outside the rear of the vehicle, the human body being located outside the vehicle, and the human body being located inside the rear seat of the vehicle.

[0012] As a preferred embodiment of the human electromagnetic exposure safety protection method based on medical implant structure optimization described in this invention, in the electromagnetic numerical model of the electric vehicle wireless power transmission system, the WPT device is installed in the rear chassis of the vehicle and adopts a dual-coil magnetic coupling resonant working mechanism. Both the transmitting and receiving coils are made of copper, and both ends are equipped with ferrite cores. The LCC-LCC compensation topology is in resonance condition, and the relevant resonance formula is shown below:

[0013]

[0014]

[0015] Among them, L f1 L f2 These represent the primary and secondary compensation inductances, respectively, C f1 C1 represents the parallel and series compensation capacitors on the primary side. f2 C2 represents the parallel and series compensation capacitors on the secondary side, L t L r These represent the inductance of the primary and secondary coupled coils, respectively.

[0016] As a preferred embodiment of the human body electromagnetic exposure safety protection method based on medical implant structure optimization described in this invention, the specific method of S2 is as follows:

[0017] S2.1: Based on the probability distribution of input variables in the framework for quantifying the safety uncertainty of electromagnetic exposure to the human body in electric vehicle wireless power transmission systems, sample points X={x} of N input variables are obtained through sampling. 1 x 2 ,...,x N};

[0018] S2.2: Based on the sample set X, use finite element simulation to solve for the electromagnetic dose value of the human body corresponding to each sample point, and obtain the training sample set. ,in The number of sample points for the output variable

[0019] S2.3: Use the training sample set D to train the Bayesian neural network and build a proxy model for human electromagnetic exposure safety in electric vehicle wireless power transmission systems;

[0020] S2.4: Call the trained Bayesian neural network proxy model to obtain the probabilistic statistical characteristics of human electromagnetic dose.

[0021] As a preferred embodiment of the human electromagnetic exposure safety protection method based on medical implant structure optimization described in this invention, the mathematical model expression of the Bayesian neural network is as follows:

[0022]

[0023] in, This represents the output of the Bayesian neural network model. Represents the parameters of a Bayesian neural network. This represents the connection weight parameters of the i-th neuron in the L-th hidden layer. This represents the connection weight parameter from the i-th neuron in the m-th hidden layer to the j-th neuron in the same layer. This represents the bias parameters of each neuron in the output layer. This represents the bias parameters of the neurons in the remaining hidden layers. For the activation functions of the remaining hidden layers, This indicates the number of neurons in the remaining hidden layers.

[0024] As a preferred embodiment of the human electromagnetic exposure safety protection method based on medical implant structure optimization described in this invention, the Hamiltonian Monte Carlo method is used to solve the posterior distribution of Bayesian neural network parameters, assuming... Obey Gaussian noise The probability model output by the Bayesian neural network is expressed as:

[0025]

[0026] in In the training sample set middle, The number of sample points for the output variable; the likelihood function expression for the Bayesian neural network model parameters is:

[0027]

[0028] Obtaining the posterior distribution of Bayesian neural network model parameters Then, sampling is performed. If the number of samplings is The predicted distribution of the Bayesian neural network output is represented as:

[0029]

[0030] Where M' represents the sampling execution. The specific number of samplings during the process; therefore, the mean and variance of the predicted output response for any input x are expressed as:

[0031]

[0032]

[0033] After training the Bayesian neural network model, a proxy model for assessing human electromagnetic exposure safety in electric vehicle wireless power transmission systems can be obtained. By calling this proxy model, N can be obtained. MCS sample And the corresponding output response, based on the acquired samples, directly solve and analyze the statistical characteristic parameters of human electromagnetic exposure dose, including the maximum induced electric field E of the human body. max Probability of exceeding the standard.

[0034] As a preferred embodiment of the human electromagnetic exposure safety protection method based on medical implant structure optimization described in this invention, the specific method of global optimization using the sparrow search algorithm in step S3 is as follows:

[0035] S3.1: Population initialization and role division: The sparrow population is divided into two categories: explorers and followers, and a certain number of guards are randomly assigned. The position of an individual in the population represents the combination of optimization variables, and the fitness value reflects the quality of the target solution.

[0036] S3.2: Position update of explorers, followers and guards: explorers are responsible for guiding the global actions of the sparrow population, followers continuously track the movement trajectory of explorers and conduct local exploration in their vicinity. The two work together to find the global optimum, while guards are used to guide the population to avoid local optima during the iteration process.

[0037] S3.3: Boundary handling and elite retention: Boundary handling: Check whether the updated position exceeds the search space boundary. If it does, reset it to the boundary value. Elite retention: Compare the fitness values ​​of individuals before and after the position update. If the fitness of the new position is higher, accept the position; otherwise, retain the original position.

[0038] Compared with existing technologies, the beneficial effects of this invention are as follows: The Bayesian Neural Network (BNN) proposed in this invention can significantly reduce computational costs, facilitating practical engineering applications. Furthermore, the uncertainty assessment results of BNN are essentially consistent with the accuracy of the traditional Monte Carlo (MC) method, ensuring the reliability of the assessment. In addition, the occluder structure optimization design method proposed in this invention can effectively mitigate the safety risks of electromagnetic exposure to the human body. Under the influence of uncertain factors, the probability of exceeding limits for electromagnetic exposure indicators in various scenarios is significantly reduced. This invention uses a sparrow search algorithm to optimize the occluder size parameters to reduce the electromagnetic exposure dose in the human body, ultimately effectively reducing the probability of exceeding limits for electromagnetic exposure indicators in patients with occluder implants under the influence of uncertain factors. Attached Figure Description

[0039] To more clearly illustrate the technical solutions of the embodiments of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and detailed embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0040] Figure 1 This is a flowchart of the method for protecting human electromagnetic exposure based on the structure optimization of medical implants according to the present invention.

[0041] Figure 2 Numerical modeling diagrams of electric vehicle wireless power transfer (EV-WPT) involved in the research on the human electromagnetic exposure safety protection method based on medical implant structure optimization of the present invention are shown in (a) and (b).

[0042] Figure 3 The basic structure diagram of the LCC-LCC compensation network involved in the research on the human electromagnetic exposure safety protection method based on medical implant structure optimization of this invention is shown below.

[0043] Figure 4 This research project involves a detailed human body model and diagrams of major organs related to the proposed method for safe protection against electromagnetic exposure of the human body based on the structure optimization of medical implants.

[0044] Figure 5 Geometric model diagram of the medical implant occluder involved in the research on the human electromagnetic exposure safety protection method based on medical implant structure optimization of this invention;

[0045] Figure 6 The research on the human electromagnetic exposure safety protection method based on medical implant structure optimization of the present invention includes the following daily electromagnetic exposure scenarios: (a) is a scene of a person standing at the rear of a car, (b) is a scene of a person standing on the side of a car, and (c) is a scene of a person sitting in the back seat of a car.

[0046] Figure 7 This diagram illustrates the optimization process of medical implants involved in the research conducted in accordance with the present invention, which is a method for protecting the safety of human electromagnetic exposure based on the structure optimization of medical implants.

[0047] Figure 8 The research on the human electromagnetic exposure safety protection method based on the optimization of medical implant structure of the present invention includes the distribution of human induced electric field in the rear scene of the vehicle exterior, where (a) is the distribution of human induced electric field with initial implant structure and (b) is the distribution of human induced electric field with optimized implant structure.

[0048] Figure 9 The following diagrams illustrate the distribution of the human body's induced electric field in a vehicle-side scene, which is part of the research conducted on the human body electromagnetic exposure safety protection method based on the optimization of medical implant structure according to the present invention. (a) is the distribution of the human body's induced electric field with the initial implant structure, and (b) is the distribution of the human body's induced electric field with the optimized implant structure.

[0049] Figure 10 The research on the human electromagnetic exposure safety protection method based on the optimization of medical implant structure of the present invention includes the distribution of human induced electric field in the rear seat of a vehicle, wherein (a) is the distribution of human induced electric field with initial implant structure, and (b) is the distribution of human induced electric field with optimized implant structure.

[0050] Figure 11 To investigate the maximum induced electric field E of the human body in the rear-end scenario of a vehicle, in relation to the human body electromagnetic exposure safety protection method based on medical implant structure optimization of this invention, research was conducted on this invention. max The probability distribution results are shown in the figure, where (a) is the probability distribution of the maximum induced electric field in the human body with the initial implant structure, and (b) is the probability distribution of the maximum induced electric field in the human body with the optimized implant structure.

[0051] Figure 12 To investigate the maximum induced electric field E of the human body in the external vehicle scene, focusing on the human body electromagnetic exposure safety protection method based on medical implant structure optimization as described in this invention. max The probability distribution results are shown in the figure, where (a) is the probability distribution of the maximum induced electric field in the human body with the initial implant structure, and (b) is the probability distribution of the maximum induced electric field in the human body with the optimized implant structure.

[0052] Figure 13 To support the research conducted on the human body electromagnetic exposure safety protection method based on medical implant structure optimization of this invention, the maximum induced electric field E of the human body in the rear seat scenario of an in-vehicle environment is investigated. max The probability distribution results are shown in the figure, where (a) is the probability distribution of the maximum induced electric field in the human body with the initial implant structure, and (b) is the probability distribution of the maximum induced electric field in the human body with the optimized implant structure. Detailed Implementation

[0053] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0054] Secondly, the present invention is described in detail with reference to the schematic diagrams. When detailing the embodiments of the present invention, for ease of explanation, the cross-sectional views illustrating the device structure may be partially enlarged, not according to the usual scale. Furthermore, the schematic diagrams are merely examples and should not limit the scope of protection of the present invention. In addition, actual fabrication should include three-dimensional spatial dimensions of length, width, and depth.

[0055] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0056] Please see Figure 1 This invention proposes a method for safe protection against electromagnetic exposure of the human body based on the structural optimization of medical implants, comprising the following steps:

[0057] S1: Modeling of WPT System and Human Body with Implants: An electromagnetic numerical model of the electric vehicle wireless power transfer (EV-WPT) system was established, including LCC-LCC compensation topology, coil and ferrite structure, and the operating frequency was set to 85kHz; an adult male human body model including a metal medical implant occluder was also established, and three typical daily high-risk exposure scenarios were constructed (human body located outside the rear of the vehicle, human body located outside the vehicle, and human body located inside the rear seat of the vehicle). Samples of uncertain variables related to the EV-WPT system and the human body were collected to provide a basis for subsequent human electromagnetic exposure assessment and analysis.

[0058] S2: Uncertainty Assessment Framework Modeling for Human Electromagnetic Exposure: An uncertainty quantification method based on Bayesian neural networks (BNN) is employed, using random variables such as EV-WPT system parameters (coil offset, compensation element error) and human tissue electromagnetic property parameters (conductivity) as inputs, with the maximum induced electric field E in the human body as the starting point. max As output, by efficiently solving the posterior probability distribution of the network model parameters, the human body E under various scenarios is calculated. max Probability of exceeding the standard.

[0059] S3: Medical Implant Structure Optimization: Incorporating Human E max As the optimization objective, the geometric structural parameters of the occluder (diameters of the upper and lower plates, waist diameter, and height) are used as optimization variables. A sparrow search algorithm is employed for global optimization to obtain the optimal design solution for the occluder, effectively reducing the electromagnetic exposure dose to the human body. Uncertainty assessment is then used to verify the improved electromagnetic radiation dose (END) to the human body after the occluder structure optimization. max The probability of exceeding the standard has been significantly reduced.

[0060] EV-WPT system and human modeling with implants:

[0061] The numerical simulation model of the EV-WPT system in this invention is as follows: Figure 2As shown. The vehicle's dimensions are as follows: length 4.5m, width 2m, height 1.5m, consistent with mainstream family car models. The WPT device is installed in the rear chassis of the vehicle and adopts a dual-coil magnetic coupling resonant working mechanism. Both the transmitting and receiving coils are made of copper windings, with 12 turns and 15 turns respectively, and the coil wire diameter is 2mm. In addition, ferrite cores are installed at both ends of the transmitter and receiver. The ferrite plates at both ends are 600mm×600mm×10mm, mainly used to confine the magnetic lines of force to the coupling area of ​​the system.

[0062] The compensation network is a core component of the WPT system, directly determining the overall system performance. Low-order compensation topologies mainly include SS, SP, PP, and PS types; high-order compensation topologies are mainly LCC-S, LCC-LCC, and LCCL-LCCL. Among them, the LCC-LCC compensation topology is widely used in WPT due to its advantages such as good transmission performance, strong anti-offset capability, and flexible parameter design. This invention selects this topology as the compensation network structure. Figure 3 The diagram shows the basic structure of an LCC-LCC compensation circuit.

[0063] Among them, L f1 L f2 These represent the primary and secondary compensation inductances, respectively, C f1 C1 represents the parallel and series compensation capacitors on the primary side. f2 C2 represents the parallel and series compensation capacitors on the secondary side, L t L r These represent the inductances of the primary and secondary coupled coils, respectively; M represents the mutual inductance between the two coils; and R... L Let U represent the load impedance and Us represent the input voltage of the compensation network. In order to achieve the best working efficiency of the wireless power transmission system, it is usually necessary to ensure that the LCC-LCC compensation network is in resonance condition. The relevant resonance formula is shown below.

[0064] (1)

[0065] (2)

[0066] The human numerical simulation model constructed in this invention uses an adult male as the baseline, with a height of 1.8m and a weight of 75kg. It also establishes detailed models of major organs and tissues such as the brain, heart, lungs, liver, kidneys, and stomach, as detailed below. Figure 4 As shown.

[0067] Due to the differences in cellular composition among various organs and tissues of the human body, electrical properties such as conductivity and dielectric constant vary significantly between different organs, exhibiting strong frequency-dependent characteristics. In this invention, the WPT system operates at 85 kHz. At this frequency, the electrical and density parameters of human tissues are shown in Table 1, where σ represents conductivity and ε...r ρ is the relative permittivity, and ρ is the tissue density.

[0068] Table 1: Electromagnetic properties and density parameters of human tissue at 85kHz

[0069]

[0070] According to the safety guidelines for low-frequency electromagnetic fields issued by the International Commission on Non-Ionizing Radiation Protection (ICNIRP), when the electromagnetic frequency is below 100kHz, the induced electric field strength is the core evaluation indicator for human electromagnetic safety. Therefore, this invention selects the human induced electric field as the basis for subsequent electromagnetic exposure safety assessment. The induced electric field strength of the human body will be uniformly represented by the symbol E below.

[0071] Furthermore, this invention employs numerical simulation modeling of implantable cardiac occluders, and the established occluder model is as follows: Figure 5 As shown, the upper and lower mesh discs are woven from nickel-titanium alloy wire mesh into a disc-like structure, while the middle section is used to seal the heart defect. In the geometric parameters of the occluder, D... d1 D d2 W represents the diameters of the upper and lower disks, respectively. h W d These represent the height and diameter of the waist, respectively.

[0072] The established occluder model is primarily made of nickel-titanium alloy. Since the electrical conductivity of metals is much higher than that of human biological tissue, this leads to the formation of low-impedance pathways. Simultaneously, the metal implant easily forms an equivalent capacitance with human tissue. These coupling effects cause electromagnetic energy accumulation, resulting in an increase in the induced electric field intensity within the body. Therefore, in the electromagnetic environment of the EV-WPT system, patients with implanted medical devices face significantly higher electromagnetic exposure safety risks than healthy individuals.

[0073] Based on the actual usage scenarios of EV-WPT devices, three typical relative positions between the human body and electric vehicles are established as follows:

[0074] 1) The human body is located outside the vehicle, in the rear area of ​​the vehicle, simulating a scenario where a person stands at the rear of the vehicle waiting for WPT charging.

[0075] 2) The human body is located outside the vehicle, on the side of the vehicle, simulating a scenario where a person stands on the side of the vehicle waiting for WPT charging.

[0076] 3) The human body is located in the back seat of the vehicle, simulating the scenario of a person sitting in the back seat of the car waiting for WPT to charge.

[0077] The three scenarios mentioned above correspond to the following in sequence. Figure 6(a), (b), and (c) This invention quantitatively evaluates the intensity E of the human body's induced electric field based on the constructed scene model, and analyzes it from two dimensions: determinism and uncertainty.

[0078] Modeling a framework for assessing the uncertainty of human electromagnetic exposure:

[0079] Human electromagnetic exposure safety assessment using EV-WPT systems typically involves high-dimensional input variables. Existing uncertainty quantification methods require a large number of training samples and often employ data-driven approaches, which can easily lead to significant errors in the calculation results. Therefore, this invention employs a Bayesian neural network (BNN) to construct a high-dimensional human electromagnetic exposure safety assessment surrogate model to calculate the probabilistic statistical characteristic parameters of electromagnetic dose within the human body. The specific implementation steps are as follows.

[0080] 1) Based on the probability distribution of input variables in the quantification framework of the human electromagnetic exposure safety uncertainty of the EV-WPT system, sample points X={x} of N input variables are obtained through sampling. 1 x 2 ,...,x N};

[0081] 2) Based on the sample set X, use finite element simulation to solve for the electromagnetic dose value of the human body corresponding to each sample point, and obtain the training sample set. ,in The number of sample points for the output variable;

[0082] 3) Use the training sample set D to train the Bayesian neural network (BNN) and construct a proxy model for human electromagnetic exposure safety in the EV-WPT system;

[0083] 4) Call the trained BNN proxy model to obtain the probabilistic statistical characteristics of human electromagnetic dose.

[0084] Bayesian Neural Networks (BNNs) rely on the Bayesian algorithm to solve for the posterior distribution of network parameters. They can build a stable model with only a small number of training samples. The mathematical model expression is as follows:

[0085] (3)

[0086] in, This represents the output of the BNN model. Represents the parameters of the BNN network. This represents the connection weight parameters of the i-th neuron in the L-th hidden layer. This represents the connection weight parameter from the i-th neuron in the m-th hidden layer to the j-th neuron in the same layer. This represents the bias parameters of each neuron in the output layer. This represents the bias parameters of the neurons in the remaining hidden layers. For the activation functions of the remaining hidden layers, This indicates the number of neurons in the remaining hidden layers.

[0087] This invention employs the Hamiltonian Monte Carlo (HMC) method to solve for the posterior distribution of Bayesian neural network (BNN) parameters, assuming... Obey Gaussian noise The probability model output by BNN can be expressed as:

[0088] (4)

[0089] in In the training sample set middle, Given the number of sample points for the output variable, the likelihood function expression for the BNN model parameters is:

[0090] (5)

[0091] Obtaining the posterior distribution of BNN model parameters Then, sampling can be performed. If the number of samplings is The predicted distribution of the BNN output can be represented as:

[0092] (6)

[0093] Where M' represents the sampling execution. The specific number of samplings during the process means that the mean and variance of the predicted output response for any input x can be expressed as:

[0094] (7)

[0095] (8)

[0096] After completing the BNN model training, a proxy model for human electromagnetic exposure safety assessment of the EV-WPT system can be obtained. By calling this proxy model, N can be obtained. MCS sample And the corresponding output response; based on the acquired samples, the statistical characteristic parameters of human electromagnetic exposure dose can be directly solved and analyzed, including the maximum induced electric field E of the human body. max The probability of exceeding the standard can be compared with the calculation results of traditional Monte Carlo (MC) to verify the calculation accuracy of the proposed BNN method in this invention.

[0097] Medical implant structure optimization:

[0098] In the electromagnetic exposure environment of EV-WPT systems, implantable medical devices are the main sensitive source leading to increased electromagnetic exposure levels in the human body. Therefore, optimizing their physical structure is of great significance for reducing electromagnetic dose to the human body. This invention explores the influence of the geometric parameters of the occluder on the induced electric field intensity E in the human body, selecting... Figure 4 The geometric part D shown d1 D d2 W d W h It was analyzed as an optimization parameter.

[0099] This invention employs a sparrow search algorithm to optimize the aforementioned size parameters. The sparrow algorithm is a novel intelligent algorithm based on the foraging and anti-predation behaviors of sparrows. Its optimization solution steps are mainly as follows:

[0100] 1) Population initialization and role assignment. The sparrow population is mainly divided into two categories: explorers and followers, with a certain number of guards randomly assigned. The position of each individual in the population represents the combination of optimization variables, and the fitness value reflects the quality of the objective solution.

[0101] 2) Position updates of explorers, followers, and guards. Explorers are responsible for guiding the global actions of the sparrow population, followers continuously track the movement trajectory of explorers and conduct local exploration in their vicinity, and the two work together to find the global optimum. Guards are used to guide the population to avoid local optima during the iteration process.

[0102] 3) Boundary handling and elite retention. Boundary handling: Check if the updated position exceeds the search space boundary; if so, reset it to the boundary value. Elite retention: Compare the fitness values ​​of the individual's position before and after the update; if the new position has a higher fitness, accept the position; otherwise, retain the original position.

[0103] This invention takes a scenario where a person is standing at the rear of a vehicle as an example, and sets the optimization interval as follows: D d1 ∈[16mm,32mm],D d2 ∈[16mm,32mm],W d ∈[2mm,3mm],W h ∈[4mm,5mm], the objective function is the maximum value of the induced electric field intensity E in the human body. max The algorithm population size and number of iterations are set to 50 and 30 respectively, and the optimization process is as follows: Figure 7 As shown.

[0104] By performing optimization operations in the solution space, the optimized design results of the occluder were obtained, as shown in Table 2. In subsequent analysis and verification, this invention will further compare the human body induced electric field intensity E values ​​under other scenarios before and after the occluder structure optimization.

[0105] Table 2: Comparison of Occluder Structures Before and After Optimization

[0106]

[0107] Example:

[0108] Based on the aforementioned model, this invention employs the COMSOL finite element numerical method to solve for the induced electric field intensity E of the human body containing the plug. According to the limits for low-frequency electromagnetic field exposure set forth by the International Commission on Non-Ionizing Radiation Protection (ICNIRP 2010), the induced electric field intensity of the human body must be below 11.5 V / m. This invention uses 11.5 V / m as the threshold criterion in subsequent numerical evaluation and analysis. First, an analysis and evaluation under deterministic operating conditions is conducted, and the results are as follows... Figures 8 to 10 As shown, (a) and (b) represent the distribution of the human body induced electric field before and after the optimization of the occluder, respectively.

[0109] In all scenarios, the maximum induced electric field intensity E of the human body max The distribution locations of these [items] are almost all near the cardiac occluder. Under deterministic conditions, the human body E [item name] in each scenario [is described]. max All values ​​are within the safe threshold (<11.5V / m). By scientifically adjusting the geometric parameters of the plug, such as the disc diameter, waist height, and diameter, E can be effectively reduced. max Indicators. When the implanted occluder is located at the rear of a vehicle, the optimized occluder structure (E) max The value can be reduced by 37.1% compared to before optimization; when the person with the implanted occluder is located on the side of the vehicle, the optimized occluder structure E max The value can be reduced by 38.9% compared to before optimization; and when the implanted occluder is located in the back seat of the vehicle, the optimized occluder structure reduces the value by 38.9%. max The value can be reduced by 26% compared to before optimization. The evaluation results show that the proposed optimized design of the medical implant occluder is applicable to all scenarios.

[0110] Next, we will conduct an evaluation and analysis under uncertain operating conditions. Based on the actual scenario, the present invention selects relevant random variables as shown in Table 3. For the EV-WPT system, the uncertain variables mainly include: the WPT coil offset caused by improper driver operation, and the manufacturing error of the system compensation network components. For the human body, the uncertain variable is mainly the conductivity parameter of human tissue. Since the frequency of the WPT system studied in this invention is below 100kHz, the electromagnetic field environment meets the quasi-static characteristics. Under this condition, the displacement current can be ignored. Compared with the dielectric constant, the conductivity has a significant impact on the electric field strength (E value) of the human body. Therefore, this invention selects conductivity instead of dielectric constant as the influencing factor.

[0111] In Table 3, L f1 L f2Indicates the compensating inductance, C f1 C1, C2, C f2 This represents the compensation capacitor; x0 and y0 represent the horizontal and vertical offsets of the WPT coil group, respectively; d0 represents the vertical offset between the transmitter and receiver; σ M σ B σ H σ L σ l σ K σ S These represent the electrical conductivity of the human muscles, brain, heart, lungs, liver, kidneys, and stomach, respectively. The symbols U and N represent uniform distribution and normal probability distribution, respectively.

[0112] Table 3: Uncertain variables involved in EV-WPT exposure scenarios

[0113]

[0114] This invention extracts samples from the distribution range of the aforementioned uncertain variables, and then simultaneously uses BNN and MC to conduct uncertainty quantification analysis and comparison. The sample size for the MC method and the BNN method are set to 5000 and 750, respectively. The maximum induced electric field E of the human body calculated by the two methods is... max The probability distribution results are as follows Figures 11 to 13 As shown, (a) and (b) correspond to the situation before and after the optimization of the plug structure, respectively. The part exceeding the safety threshold of 11.5V / m is marked in light gray. Table 4 lists the E values ​​under various scenarios. max Probability of exceeding the limit.

[0115] Table 4: E under various exposure scenarios max Over-limit probability

[0116]

[0117] It can be seen that the solution results of the BNN surrogate model are basically consistent with those of the Monte Carlo (MC) method. Since BNN requires fewer samples than the MC method, this reflects that BNN can significantly reduce computational costs while maintaining the accuracy of uncertainty assessment. Table 4 shows that when a human body is inside the cabin of an electric vehicle (EV), its electromagnetic exposure risk is significantly lower than outside the vehicle, which is attributed to the electromagnetic shielding effect of the vehicle itself. Before the optimization of the occluder structure, the maximum induced electric field intensity E of the human body... max The probability of exceeding the limit varies under different exposure scenarios, and the occluder structure optimization method can significantly reduce this risk of exceeding the limit. When a human body is located at the rear of the vehicle, the E value after occluder structure optimization is [not specified]. max The probability of exceeding the standard is less than 5%, a reduction of more than 20% compared to before optimization. When the human body is located on the outside of the vehicle, the E value after the blocker structure optimization is also lower. maxThe probability of exceeding the standard is also less than 5%, a reduction of more than 20% compared to before optimization. When a person is inside the vehicle, the optimized occluder structure... max There is no probability of exceeding the standard. The above results verify the feasibility of the proposed medical implant design method in terms of human electromagnetic exposure safety protection.

[0118] Although the present invention has been described above with reference to embodiments, various modifications can be made and components can be replaced with equivalents without departing from the scope of the invention. In particular, as long as there is no structural conflict, the features in the disclosed embodiments can be combined with each other in any manner. The lack of an exhaustive description of these combinations in this specification is merely for the sake of brevity and resource conservation. Therefore, the present invention is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A method for safe protection against electromagnetic exposure of the human body based on the structural optimization of medical implants, characterized in that, Includes the following steps: S1: WPT system and human modeling with implants: Establish an electromagnetic numerical model of the wireless power transmission system of electric vehicles, as well as an adult male human model containing a metal medical implant occluder, and construct three typical daily high-risk exposure scenarios: the person is located outside the rear of the vehicle, the person is located outside the vehicle, and the person is located inside the rear seat of the vehicle, to provide a basis for subsequent human electromagnetic exposure assessment and analysis. S2: Uncertainty Assessment Framework Modeling for Human Electromagnetic Exposure: An uncertainty quantification method based on Bayesian neural networks is employed, using random variables of electric vehicle wireless power transmission system parameters and human tissue parameters as inputs, with the maximum induced electric field E in the human body as the starting point. max As output, by efficiently solving the posterior probability distribution of the network model parameters, the human body E under various scenarios is calculated. max The probability of exceeding the standard limit; S3: Medical Implant Structure Optimization: Incorporating Human E max As the optimization objective, the geometric structural parameters of the occluder are used as optimization variables. These parameters include the diameters of the upper and lower plates, the waist diameter, and the height. A sparrow search algorithm is used for global optimization to obtain the optimal design solution for the occluder. Uncertainty assessment is then used to verify the optimized occluder structure and its effect on the human body (E). max Probability of exceeding the standard.

2. The method for safe protection against electromagnetic exposure of the human body based on the structural optimization of medical implants according to claim 1, characterized in that, In S1, the electromagnetic numerical model of the electric vehicle wireless power transmission system includes an LCC-LCC compensation topology, coil and ferrite structure, and the operating frequency is set to 85kHz; three typical daily high-risk exposure scenarios include people being located outside the rear of the vehicle, people being located outside the vehicle, and people being located inside the rear seat of the vehicle.

3. The method for safe protection against electromagnetic exposure of the human body based on the structural optimization of medical implants according to claim 2, characterized in that, In the electromagnetic numerical model of the wireless power transfer system for electric vehicles, the WPT device is installed in the rear chassis of the vehicle and adopts a dual-coil magnetic coupling resonant working mechanism. Both the transmitting and receiving coils are made of copper, and ferrite cores are installed at both ends. When the LCC-LCC compensation topology is in resonance mode, the relevant resonance formula is as follows: Among them, L f1 L f2 These represent the primary and secondary compensation inductances, respectively, C f1 C1 represents the parallel and series compensation capacitors on the primary side. f2 C2 represents the parallel and series compensation capacitors on the secondary side, L t L r These represent the inductance of the primary and secondary coupled coils, respectively.

4. The method for safe protection against electromagnetic exposure of the human body based on the structural optimization of medical implants according to claim 1, characterized in that, The specific method of S2 is as follows: S2.1: Based on the probability distribution of input variables in the framework for quantifying the safety uncertainty of electromagnetic exposure to the human body in electric vehicle wireless power transmission systems, sample points X={x} of N input variables are obtained through sampling. 1 x 2 ,...,x N }; S2.2: Based on the sample set X, use finite element simulation to solve for the electromagnetic dose value of the human body corresponding to each sample point, and obtain the training sample set. ,in The number of sample points for the output variable; S2.3: Use the training sample set D to train the Bayesian neural network and build a proxy model for human electromagnetic exposure safety in electric vehicle wireless power transmission systems; S2.4: Call the trained Bayesian neural network proxy model to obtain the probabilistic statistical characteristics of human electromagnetic dose.

5. The method for safe protection against electromagnetic exposure of the human body based on the structural optimization of medical implants according to claim 4, characterized in that, The mathematical model expression of the Bayesian neural network is as follows: in, This represents the output of the Bayesian neural network model. Represents the parameters of a Bayesian neural network. This represents the connection weight parameters of the i-th neuron in the L-th hidden layer. This represents the connection weight parameter from the i-th neuron in the m-th hidden layer to the j-th neuron in the same layer. This represents the bias parameters of each neuron in the output layer. This represents the bias parameters of the neurons in the remaining hidden layers. For the activation functions of the remaining hidden layers, This indicates the number of neurons in the remaining hidden layers.

6. The method for safe protection against electromagnetic exposure of the human body based on the structural optimization of medical implants according to claim 5, characterized in that, The Hamiltonian Monte Carlo method is used to solve for the posterior distribution of the parameters of the Bayesian neural network, assuming... Obey Gaussian noise The probability model output by the Bayesian neural network is expressed as: in In the training sample set middle, The number of sample points for the output variable; the likelihood function expression for the Bayesian neural network model parameters is: Obtaining the posterior distribution of Bayesian neural network model parameters Then, sampling is performed. If the number of samplings is The predicted distribution of the Bayesian neural network output is represented as: Where M' represents the sampling execution. The specific number of samplings during the process; therefore, the mean and variance of the predicted output response for any input x are expressed as: After training the Bayesian neural network model, a proxy model for assessing human electromagnetic exposure safety in electric vehicle wireless power transmission systems can be obtained. By calling this proxy model, N can be obtained. MCS sample And the corresponding output response, based on the acquired samples, directly solve and analyze the statistical characteristic parameters of human electromagnetic exposure dose, including the maximum induced electric field E of the human body. max Probability of exceeding the standard.

7. The method for safe protection against electromagnetic exposure of the human body based on the structural optimization of medical implants according to claim 1, characterized in that, In S3, the specific method for global optimization using the sparrow search algorithm is as follows: S3.1: Population initialization and role division: The sparrow population is divided into two categories: explorers and followers, and a certain number of guards are randomly assigned. The position of an individual in the population represents the combination of optimization variables, and the fitness value reflects the quality of the target solution. S3.2: Position update of explorers, followers and guards: explorers are responsible for guiding the global actions of the sparrow population, followers continuously track the movement trajectory of explorers and conduct local exploration in their vicinity. The two work together to find the global optimum, while guards are used to guide the population to avoid local optima during the iteration process. S3.3: Boundary handling and elite retention: Boundary handling: Check whether the updated position exceeds the search space boundary. If it does, reset it to the boundary value. Elite retention: Compare the fitness values ​​of individuals before and after the position update. If the fitness of the new position is higher, accept the position; otherwise, retain the original position.

Images