Composite reactive resonance shielding coil structure and robust optimization method thereof

Abstract

The invention is suitable for the technical field of wireless charging of electric vehicles, and provides a composite reactive resonance shielding coil structure and a robust optimization method thereof. The composite reactive resonance shielding coil structure is formed by combining the nanocrystalline, the aluminum plate and the reactive resonance shielding coil, the leakage electromagnetic field of an electric vehicle wireless charging (EV-WPT) system can be weakened, the high-energy transmission efficiency and the light weight of the system are guaranteed, and the offset tolerance and the shielding effect are enhanced. According to the robust optimization method, an Attention-Unet uncertainty agent model and a multi-target whale optimization algorithm (MOWOA) are combined, the maximum value of the energy transmission efficiency of the system and the maximum value of the intensity of the induced electric field in the human body serve as optimization targets, uncertainty factors are considered, and rapid multi-target optimization design is achieved. The method provides practical application significance and theoretical guidance for electromagnetic shielding design of the EV-WPT system, and also provides a brand new research thought for human body electromagnetic exposure safety and protection.

Description

Technical Field

[0001] This invention belongs to the field of wireless charging technology for electric vehicles, and particularly relates to a composite reactive resonant shielded coil structure and its robust optimization method. Background Technology

[0002] As the automotive industry continues to develop towards intelligence and green technology, electric vehicles (EVs), as an important green transportation tool, have gradually risen to the forefront of the global industry. Among the charging methods for EVs, traditional wired charging suffers from limitations such as inconvenient cable plugging and unplugging, easily damaged charging interfaces, and large footprint of charging facilities. Furthermore, the limited range and inconvenience of charging during long-distance travel restrict its application scenarios. Against this backdrop, wireless power transfer (WPT) technology has gained widespread attention due to its advantage of contactless power transmission. This technology eliminates the need for manual plugging and unplugging of charging cables, improving the user experience and avoiding potential safety hazards such as leakage and short circuits during cable insertion and removal. It also operates normally in inclement weather, reducing limitations on charging scenarios. However, as a loosely coupled structure, the EV-WPT system works by using electromagnetic induction between the transmitting and receiving coils to achieve power transmission through magnetic field coupling. With the increasing demand for charging power, the leakage electromagnetic field dose generated during system operation also increases, potentially posing an electromagnetic safety threat to surrounding drivers and passengers. Related animal studies have shown that exposure to electromagnetic fields may have a tumor-promoting effect. For example, mice treated with carcinogens showed a significantly higher number of lung and liver tumors after exposure to electromagnetic fields compared to the control group. Mice exposed to low-frequency magnetic fields for extended periods also showed an increased incidence of mammary tumors. Furthermore, the position of the human body relative to the WPT device is not fixed (e.g., the position of the driver, passenger, or passerby is not fixed), and during automatic parking, insufficient algorithm accuracy may lead to parking position deviations. These factors can cause changes in the electromagnetic exposure dose within the human body, further exacerbating health risks. Therefore, ensuring the safety of human electromagnetic exposure is a pressing issue to be addressed in the development of wireless charging technology for electric vehicles.

[0003] To ensure the safe operation of WPT equipment, safety standards for electromagnetic exposure doses to the human body are indispensable. Several international organizations have conducted relevant research. For example, ICNIRP 2010 and IEEE C95.1 are standards related to electromagnetic exposure safety around EV-WPT systems. Various shielding measures have been proposed to address the leakage magnetic field (LMF) of EV-WPT systems. Currently, shielding technologies for EV-WPT systems are mainly divided into passive shielding and active shielding. Active shielding technology is less commonly used due to its high cost, complex structure, and large space requirements; while passive shielding technology, with its unique characteristics, has become the most widely used method. It generally relies on the structure and properties of the materials themselves for electromagnetic shielding. Materials are typically categorized as ferromagnetic materials and metallic conductors. The most common form is adding ferrite plates and aluminum plates to the outside of the transmitting and receiving coils, utilizing the magnetic permeability and eddy current effect of high-permeability materials to prevent the escape of LMF. However, ferromagnetic materials such as ferrites suffer from problems such as large weight, high brittleness, and significant temperature-dependent magnetic properties, posing a challenge to the compact design of EV-WPT systems when used for large-area coverage. Metallic conductors generate counteracting magnetic fields through internal eddy currents, but the system efficiency decreases more significantly the closer they are to the system. Some researchers have also attempted to construct shielding structures for EV-WPT systems using metamaterials; however, metamaterials can only achieve shielding effects at specific frequencies, and their shielding performance rapidly declines with changes in the system's operating frequency, limiting their application in the electromagnetic shielding field of EV-WPT systems. Furthermore, several studies have proposed ferrite composite structures that can effectively control magnetic flux and improve efficiency, but these are still limited by the inherent defects of ferrites. While nanocrystals are considered a substitute for ferrites, possessing superior magnetic and mechanical properties, related research has used nanocrystalline cores to design magnetically permeable structures, demonstrating that they can achieve higher efficiency and lower leakage electromagnetic fields. However, their high cost limits their widespread use; therefore, relying solely on the laying up of these materials is insufficient to achieve ideal electromagnetic shielding.

[0004] Resonant reactive power shielding technology utilizes the system's own energy to generate a counteracting magnetic field, eliminating the need for an external power supply. It offers excellent shielding performance, minimal impact on system efficiency, and also contributes to the lightweight design of WPT (Electromagnetic Transmission Platform) devices. Related research includes placing passive shielding coils externally to address insulation issues arising from direct connections between the power transfer coil and the shielding coil, or using capacitor compensation to achieve resonant reactive power shielding and alter the equivalent inductance of passive shielding coils. However, these studies fail to consider uncertainties present in practical applications, affecting their shielding effectiveness in complex scenarios. Robust Design Optimization (RDO) is a design method that incorporates unavoidable uncertainties into the optimization design. In real-world applications, uncertainties such as parking position shifts and changes in the distance between the human body and the EV-WPT can affect electromagnetic exposure doses. Existing research has not fully integrated RDO into the shielding structure design, making it difficult to address complex situations in real-world scenarios.

[0005] In research on EV-WPT system design, to simultaneously consider transmission efficiency and shielding effectiveness, most researchers choose to establish the shielding structure using the finite element method (FEM) and combine it with multi-objective optimization algorithms. However, the high computational cost of FEM simulation software leads to increased time costs for each optimization iteration. Although some studies have proposed using the sparse chaotic polynomial expansion (PCE) method to quantify the impact of uncertainties to optimize transmission efficiency, PCE suffers from the "curse of dimensionality" problem, resulting in similarly high computational costs. Furthermore, the optimization design of existing electromagnetic shielding structures is mostly based on deterministic states and does not incorporate comprehensive analysis with Reactive Power Discharge (RDO), making it difficult to guarantee electromagnetic safety in practical applications. Therefore, this invention proposes a composite reactive resonant shielding coil structure and its robust optimization method. Summary of the Invention

[0006] The purpose of this invention is to provide a composite reactive resonant shielded coil structure and its robust optimization method, aiming to solve the problems mentioned in the background art.

[0007] The objective of this invention is achieved through the following technical solution: A composite reactive resonant shielded coil structure is disclosed for use in a wireless charging system for electric vehicles. The structure includes a reactive resonant shielded coil, a nanocrystalline layer, and an aluminum plate. The nanocrystalline layer is adhered to the periphery of the reactive resonant shielded coil, and the aluminum plate is adhered to the nanocrystalline layer and located outside the reactive resonant shielded coil. The reactive resonant shielded coil generates a reverse canceling magnetic field through a leakage magnetic field to weaken the electromagnetic field strength in the non-operating area of ​​the wireless charging system. The nanocrystalline layer and aluminum plate are used to attract the leaked electromagnetic field to the vicinity of the reactive resonant shielded coil through their high permeability, thereby increasing the reverse magnetic field strength and reducing the external magnetic field strength, thus reducing the electromagnetic exposure dose to the human body.

[0008] Furthermore, the relative permeability of the nanocrystalline layer is 22000 under operating conditions at room temperature and 85 kHz.

[0009] Furthermore, the reactive resonant shielding coil is provided with four coils, including two transverse reactive resonant shielding coils and two longitudinal reactive resonant shielding coils, which are symmetrically arranged on the transverse and longitudinal sides of the receiving coil, respectively; both the transverse and longitudinal reactive resonant shielding coils are connected to a compensation capacitor.

[0010] Furthermore, the distance between the transverse reactive resonant shielding coil and the edge of the receiving coil is 0-0.05m; the distance between the longitudinal reactive resonant shielding coil and the edge of the receiving coil is 0-0.05m; the compensation capacitor of the transverse reactive resonant shielding coil is 500-1500nF; and the compensation capacitor of the longitudinal reactive resonant shielding coil is 500-1500nF.

[0011] A robust optimization method for the composite reactive resonant shielded coil structure described above includes: An Attention-Unet uncertainty surrogate model is constructed, with design variables and uncertainty factors as inputs, and the system energy transmission efficiency, the maximum induced electric field strength of the human body, the maximum induced electric field strength of the brain, and the maximum induced electric field strength of the lungs as outputs. The design variables include the distance between the lateral reactive resonant shielding coil and the edge of the receiving coil, the distance between the longitudinal reactive resonant shielding coil and the edge of the receiving coil, and the compensation capacitance of the lateral and longitudinal reactive resonant shielding coils. The uncertainty factors include the lateral offset of the human body, the longitudinal offset of the human body, the lateral offset of the vehicle, the longitudinal offset of the vehicle, the distance between the transmitting coil and the receiving coil, and the human body orientation deflection angle. Based on the output of the Attention-Unet uncertainty surrogate model, a multi-objective whale optimization algorithm is used to perform multi-objective optimization, solve the Pareto front, and obtain robust optimal parameters (i.e., the optimal distance between the horizontal reactive resonant shielding coil and the edge of the receiving coil, the distance between the vertical reactive resonant shielding coil and the edge of the receiving coil, and the compensation capacitance of the horizontal and vertical reactive resonant shielding coils).

[0012] Furthermore, the core framework of the Attention-Unet uncertain agent model is the Attention-Unet network, which is formed by introducing the Attention Gate mechanism on the basis of the U-Net network. The U-Net network is a fully convolutional deep network model based on encoder-decoder. The input data is normalized before entering the Attention-Unet network; First, the data passes through an encoder, which consists of convolutional layers and downsampling layers. The convolutional operation automatically extracts local spatial or sequence features from the input data, and a ReLU activation function is applied after convolution. After passing through the convolutional layers, the downsampling layer... MaxPool The algorithm slides through windows, retaining the maximum value within each window, progressively compressing information to preserve the most critical features. After processing through a downsampling layer, the information enters the Attention Gate, which controls the transmission of encoder features to the decoder. The decoding path first passes through an upsampling layer, then concatenates with the output of the Attention Gate, and finally performs convolution and ReLU activation. Finally, it is mapped to the output dimension through another convolution to obtain the final output. The set of is : ; In the formula, This is the set of predicted outputs of the model; W final These are the final convolution kernel parameters; The feature input before entering the convolutional mapping; ρ For bias.

[0013] Furthermore, the multi-objective whale optimization algorithm updates the population position through three stages: prey search, prey encirclement, and bubble web attack. The position update formula includes: Hunting phase: ; ; In the formula, X ( t +1) indicates the updated whale location; X r ( t The location of the whale is randomly selected by the population. A and C For coefficient vectors; D The distance between the individual whale and the optimal solution; t This represents the current iteration number; X ( t () indicates the current location of the whale; The stage of surrounding the prey: ; ; In the formula, z It is a constant that defines the shape of the spiral; X z ( t () represents the optimal whale location obtained so far; Bubble web attack phase: ; ; In the formula, l It is a random floating-point number in the interval (-1, 1).

[0014] Compared with the prior art, the beneficial effects of the present invention are: The composite reactive resonant shielding coil structure proposed in this invention is composed of nanocrystals, aluminum plates, and a reactive resonant shielding coil. While ensuring high energy transfer efficiency, it achieves system lightweighting, effectively weakens the leakage electromagnetic field of the EV-WPT system, enhances system offset tolerance and shielding effect, and provides a safer and more reliable electromagnetic environment for surrounding users. Its robust optimization method combines the Attention-Unet uncertainty surrogate model with the Multi-Objective Whale Optimization Algorithm (MOWOA). Using system energy transfer efficiency and the maximum induced electric field strength inside the human body as optimization objectives, and considering uncertainty factors, it achieves rapid multi-objective optimization design, improving the robustness of the composite reactive resonant shielding coil structure and making the system energy transfer efficiency and human electromagnetic exposure dose more stable and safe under the influence of uncertainty. Simulation experiments show that, considering uncertainty factors, the maximum induced electric field strength of the human body after shielding is [not specified]. E max The mean value decreased by 48.38%, and the maximum value of the induced electric field intensity in the lungs... E max The average value of the electromagnetic shielding coil decreased by 41.73%, the average energy transfer efficiency of the WPT system increased by 6.46%, and the probability of the maximum value of the induced electric field intensity inside the human body exceeding the standard limit decreased to 0%, proving the effectiveness of the composite reactive resonant shielding coil structure and its design method. This invention provides practical application significance and theoretical guidance for the electromagnetic shielding design of EV-WPT systems, and also provides a new research approach for the safety and protection of human electromagnetic exposure. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the EV-WPT system.

[0016] Figure 2 This is a schematic diagram of a composite reactive resonant shielded coil structure.

[0017] Figure 3 This is the equivalent circuit diagram of the EV-WPT system.

[0018] Figure 4 This represents the positional relationship between the receiving coil and the reactive resonant shielding coil.

[0019] Figure 5 This is the Attention-Unet network structure.

[0020] Figure 6 The system's overall structure is shown in Figure 1; where (a) represents the exposure scenario and (b) represents the offset between the human body and the electric vehicle.

[0021] Figure 7 The model is a human body simulation model; where (a) is the human torso and (b) is an organ.

[0022] Figure 8 The results of each method (Attention-Unet uncertain surrogate model, FEM simulation, traditional BP neural network and GRU network) are compared; where (a) is the system energy transfer efficiency. η (b) is the lungs E max (c) represents the brain. E max (d) refers to the human body E max .

[0023] Figure 9 This is a schematic diagram of the Pareto front.

[0024] Figure 10 The magnetic field distribution of the EV-WPT system is shown in Figure 1; where (a) is the magnetic field distribution of the EV-WPT system before shielding and (b) is the magnetic field distribution of the EV-WPT system after shielding.

[0025] Figure 11 The electric field distribution of the human body is shown in Figure 1; where (a) is the electric field distribution of the human body before shielding and (b) is the electric field distribution of the human body after shielding.

[0026] Figure 12 The electric field distribution in the lungs is shown in Figure 1; where (a) is the electric field distribution in the lungs before shielding and (b) is the electric field distribution in the lungs after shielding.

[0027] Figure 13 Let be the probability distribution function (PDF); where (a) is the system energy transfer efficiency. η PDF of (b) is of the lungs. E max The PDF, (c) is of the human body E max PDF, (d) for brain E max The PDF.

[0028] Figure 14 The results are from MOAT; where (a) represents the effect of input variables on the system energy transfer efficiency. η The degree of influence, (b) is the effect of input variables on the lungs E max The degree of influence, (c) represents the impact of the input variable on the human body. E max The degree of influence, (d) is the effect of the input variable on the brain. Emax The extent of the impact. Detailed Implementation

[0029] In order to provide a clearer understanding of the technical features, objectives and beneficial effects of the present invention, the technical solution of the present invention will now be described in detail below, but it should not be construed as limiting the scope of implementation of the present invention.

[0030] This invention provides a composite reactive resonant shielded coil structure and its robust optimization method, including: 1. Composite reactive resonant shielded coil structure and its equivalent circuit; The EV-WPT system is based on the principle of magnetic coupling resonance, using an electromagnetic field as a medium, and transmitting through a transmitting coil. Tx With receiving coil Rx Wireless power transmission is achieved through resonance between the transmitting coil and the electromagnetic field. First, AC power is input from the mains, converted to DC by the transmitter's rectifier circuit, and then converted back to AC by a high-frequency inverter circuit. The output frequency of the inverter circuit is the resonant frequency of the transmitting resonant circuit. At this point, the transmitting circuit operates in a resonant state under the action of the transmitter's compensation circuit, exhibiting purely resistive behavior. According to Faraday's law of electromagnetic induction, the transmitting coil... Tx A high-frequency electromagnetic field with a resonant frequency is excited into space, and is received by a receiving coil. Rx During reception, under the action of the receiving end compensation circuit, the receiving end circuit also operates in a resonant state, exhibiting pure resistivity. The receiving end rectifier circuit and chopper circuit convert high-frequency electrical energy into DC power required by the load. To ensure maximum efficiency of the EV-WPT system during normal operation, this invention employs a bilateral SS compensation circuit. An EV-WPT system with a reactive resonant circuit is as follows... Figure 1 As shown.

[0031] in, U S As a source of motivation, C 1 is the transmitting coil Tx Series compensation capacitor at the end, L 1 is the transmitting coil Tx The coil self-inductance at the end, L 2 is the receiving coil Rx The coil self-inductance at the end, L 3 represents the self-inductance of the reactive resonant shielded coil. M 12 For transmitting coil Tx With receiving coil Rx Mutual inductance between the coils M 13 For transmitting coil Tx Mutual inductance between the coil and the reactive resonant shielded coil M 23 For receiving coil Rx Mutual inductance between the coil and the reactive resonant shielded coil C 2 is the receiving coil Rx Series compensation capacitor at the end, R L For load resistance, U 1 represents the voltage of the transmitting coil. U 2 represents the voltage of the receiving coil. C 3 represents the reactive power resonant shielded coil compensation capacitor. Generally, the EV-WPT system only considers the equivalent internal resistance of the coil. When the system is in resonance, the components satisfy the following relationship: Formula 1: ; In the formula, ω ω is the angular frequency.

[0032] In EV-WPT systems, the high-frequency coupled electromagnetic field around the coupling coil is crucial for energy transfer. The current generated by the leakage magnetic field in the reactive resonant shielded coil produces a counteracting magnetic field, thus weakening the electromagnetic field in non-working areas. For EV-WPT systems, the transmission of electromagnetic waves within the shielding structure directly determines its shielding effectiveness. Traditional reactive resonant shielded coils provide good shielding at the coil center and within its internal region, but their shielding effect around the coil boundaries and external regions is poor. Because magnetic field lines are closed and cannot be severed, the only way to achieve magnetic shielding is to guide the magnetic field lines to reduce the magnetic field escaping to the outside of the transmission mechanism. The following relationship exists in the magnetic field: Formula 2: ; In the formula, Indicates magnetic flux. F m It represents magnetomotive force. R m This represents magnetic reluctance. To overcome the above problems, this invention proposes a composite reactive resonant shielded coil structure based on the traditional reactive resonant shielded coil structure (see...). Figure 2The device consists of a reactive resonant shielding coil, a nanocrystalline layer, and an aluminum plate. The nanocrystalline layer is adhered to the periphery of the reactive resonant shielding coil, and the aluminum plate is adhered to the nanocrystalline layer and located outside the reactive resonant shielding coil. The reactive resonant shielding coil is made of wound copper wire, a material with good conductivity and low cost. The high-permeability ultrathin nanocrystalline layer and aluminum plate weaken the leakage electromagnetic field around the reactive resonant shielding coil. First, the nanocrystalline layer gathers the magnetic field, diverting a portion of the magnetic field from the non-working area into the working area. Then, the aluminum plate electrically shields the remaining leakage electromagnetic field, reducing the electromagnetic field intensity in the non-working area through the eddy current effect. At room temperature and 85kHz operating conditions, the relative permeability of the nanocrystalline layer can reach 22000. Under the same conditions, the relative permeability of traditional magnetic material ferrite can only reach 3300. By attracting the previously unused leakage electromagnetic field around the reactive resonant shielding coil to the coil's periphery, the nanocrystalline layer further increases the reverse magnetic field strength while reducing the external magnetic field strength, thus achieving a better shielding effect and reducing the electromagnetic exposure dose to the human body.

[0033] For ease of calculation, the power supply and conversion circuit are respectively represented as two equivalent voltage sources. The equivalent internal resistance of each coil is too small to be ignored. Therefore, Figure 1 It can be further simplified to Figure 3 The equivalent circuit is shown. The inductor is shown in the diagram. L 3 and compensation capacitor C 3 is simplified to an equivalent inductance L eq The formula is as follows: Formula 3: ; The loop voltage equation is: Formula 4: ; In the formula, This serves as the voltage source for the transmitting coil circuit. This serves as the voltage source for the receiving coil circuit. j The imaginary unit; This refers to the current in the transmitting coil circuit. To receive the coil circuit current; This is the current in the reactive resonant shielded coil circuit.

[0034] When the reactive resonant shielded coil is working, let M 13 = αM 12 , M 23 = βM 12 ,in α and β These are all proportional parameters, the purpose of which is to... M 13 andM 23 All utilize M 12 This is to facilitate calculation and simplification. According to Equation 4, we can calculate: Formula 5: ; From Equation 5, we can see that if α and β A very small value indicates that the coupling effect between the reactive resonant shielding coil and the transmitting and receiving coils is very small. Therefore, the original system is not affected after the reactive resonant shielding coil operates. and It can be considered a constant. The reactive resonant circuit generates current solely due to the leakage magnetic field. Therefore, equation 4 can be simplified to equation 6: Formula 6: ; According to the Biot-Saffar law, the magnetic field strength produced by a current in a coil is directly proportional to the current. Therefore, the shielding effect is determined by... It was decided. The magnitude is mainly determined by the mutual inductance coefficient. M 13 , M 23 With equivalent inductance L eq Decision, and M 13 , M 23 and L eq It is determined by the structure of the coil itself, the distance between the reactive resonant shielded coil and the EV-WPT system. d and compensation capacitor C 3. The shielding effect of the reactive resonant shielded coil is determined solely by its distance from the EV-WPT system. Once the coil structure is determined, the shielding effect of the reactive resonant shielded coil is determined solely by its distance from the EV-WPT system. d and compensation capacitor C 3. Decision.

[0035] However, in practice, only one pre-placed equivalent inductance is allowed. L eqYes, it exists. When all coupling mechanisms except the reactive resonant shielding coil are determined, the aforementioned parameters are closely related to the number of coil turns, coil size, and compensation capacitor parameters. Therefore, finding the optimal parameter combination through numerous trials is unrealistic. Furthermore, due to medical ethics concerns, current research on human electromagnetic exposure mainly relies on FEM simulation software. However, due to the fine meshing of simulation software, the time cost of each calculation is enormous. Therefore, this invention proposes a surrogate model based on Attention-Unet, which can effectively shorten the computation time cost, generalize the human electromagnetic exposure scenario with relatively less data, and obtain accurate results. In practical applications, various uncertainties are unavoidable. This requires the composite reactive resonant shielding coil structure designed in this invention to enhance its tolerance to various uncertainties, reduce the electromagnetic exposure dose to the human body around the EV-WPT system, and thus have greater practical application value.

[0036] The design variables and their distribution characteristics, as well as the uncertainty factors and their distribution characteristics of the reactive resonant shielded coil proposed in this invention, are shown in Tables 1 and 2, respectively. This is because the EV-WPT system receiving coil... Rx The structure is symmetrical, and the positions of the four reactive resonant shielding coils are also symmetrical. Therefore, the parameters of the two horizontal reactive resonant shielding coils and the two vertical reactive resonant shielding coils should be the same. The positional relationship between the receiving coil and the reactive resonant shielding coils is as follows: Figure 4 As shown in the figure. Here, U represents a uniform distribution. Since the offset between the coils in the EV-WPT system is equally likely as the offset of the human body in all directions, it is usually considered a uniform distribution in engineering applications.

[0037] Table 1 Design variables and their distribution

[0038] Table 2. Uncertainty Factors and Their Distribution

[0039] 2. A robust optimization method based on the Attention-Unet uncertain surrogate model; The U-Net network is a fully convolutional deep network model based on an encoder-decoder architecture. It enables efficient learning using a small number of labeled samples in processing one-dimensional high-order data. The encoder extracts features through convolution and downsampling, deeply extracting high-dimensional information; the decoder reconstructs local information and restores the original data structure through upsampling and skip connections. The combination of these two approaches achieves multi-scale feature extraction. Furthermore, this invention introduces the Attention Gate (AG) mechanism on top of the U-Net network, forming the Attention-Unet network. This mechanism can calculate weights for each channel or skip connection, strengthening important features, suppressing redundant features, and dynamically adjusting the model's focus. Moreover, addressing the overfitting problem common in traditional U-Net regression networks, the Attention Gate mechanism automatically sparsifies feature representations, enhancing the model's generalization ability, exhibiting stronger robustness to noise or irrelevant features, and indicating the variables the model focuses on.

[0040] This invention uses the Attention-Unet network as the core framework to construct the Attention-Unet uncertainty surrogate model, and takes the design variables of the reactive resonant shielded coil in Table 1 as input variables. X =[ x 1, x 2,..., x N ],in N This refers to the dimensions of the optimization variables. Since there are a total of 4 optimization variables in the table, therefore... N =4. Since the ICNIRP 2010 guidelines explicitly state that the induced electric field strength is a limit for the human body and brain, and considering the lungs' proximity to the EV-WPT system, their large size, and existing research indicating that exposure to extremely low-frequency magnetic fields may lead to symptoms such as decreased respiratory rate and prolonged respiratory cycle, this invention also considers the electromagnetic exposure dose to the human lungs. In summary, this invention uses simulation to measure the system's energy transfer efficiency. η The maximum induced electric field strength in the human body, brain, and lungs E max As the regression variable in the model, let it be denoted as Y =[ y 1, y 2, y 3, y 4], of which y 1, y 2, y 3, y 4 represents the system energy transfer efficiency. η The maximum value of the induced electric field intensity of the human bodyE max The maximum value of the induced electric field intensity in the brain E max Maximum value of induced electric field intensity in the lungs E max Input data is normalized to improve model accuracy and enhance the model's ability to generalize to new data.

[0041] Formula 7: ; Formula 8: ; In the formula, For the first n The first set of data i Data normalization results; For the first n The first set of data i Each belongs to X The original data for the category; For the first i indivual X A collection of class data; The value is the normalized value; For the first n The first group of data j One set of raw data; For the first j indivual Y A collection of class data; ε It is a very small constant used to prevent the denominator from being 0; max and min represent the maximum and minimum value functions, respectively.

[0042] The input data is normalized before entering the Attention-Unet network, whose specific structure is as follows: Figure 5 As shown. First, the data passes through an encoder, which consists of convolutional layers and downsampling layers. The convolutional operation automatically extracts local spatial or sequence features from the input data. For the one-dimensional data in this invention, it's equivalent to sliding a small window to extract pattern fragments, thereby enhancing nonlinear expressive power. Furthermore, a ReLU activation function is applied after convolution, as shown in the following equation: Formula 9: ; In the formula, For the first l The output features of the layer; Represents the ReLU activation function; For the first l Layer weights; * indicates a one-dimensional convolution operation; For the first l -1 layer output features; For the first l Layer bias. After passing through the convolutional layer, the downsampling layer passes through... MaxPool Slide the window and retain the maximum value within each window. Dimensionality reduction gradually compresses the information, but still retains the most critical features.

[0043] Formula 10: ; In the formula, For the first l Features after layer downsampling.

[0044] After processing by the downsampling layer, the data enters the Attention Gate, which controls which features are passed from the Encoder to the Decoder. Let... m Encoder features g If the gate signal is for the decoder, then the attention coefficient is... τ As shown in the following formula: Formula 11: ; In the formula, W x This is the encoder feature weight matrix; x For encoder output features; W g This is the gated signal weight matrix for the decoder; ρ For bias.

[0045] The Attention Gate then outputs the following: Formula 12: ; In the formula, x gated This is the output of the Attention Gate.

[0046] The decoding path first passes through the upsampling layer: Formula 13: ; In the formula, For the first l Layer upsampling output features; ConvTransposelD This is a one-dimensional transposed convolution; For the first l Output features of layer +1.

[0047] Then it is concatenated with the output of the Attention Gate: Formula 14: ; In the formula, Concat To facilitate the splicing operation, features from different sources are merged to enrich the feature information.

[0048] Finally, convolution and ReLU activation are performed: Formula 15: ; In the formula, BN Indicates batch normalization; ConvlD This is a one-dimensional convolution operation.

[0049] Then, by mapping to the output dimension through a convolution, we obtain... The set of is .

[0050] Formula 16: ; In the formula, This is the set of predicted outputs of the model; W final These are the final convolution kernel parameters; The feature input before entering the convolutional mapping; ρ For bias.

[0051] On the one hand, by using uncertainty factors as input variables, this invention can obtain uncertainty surrogate models suitable for various scenarios. On the other hand, integrating the Attention Gate mechanism into U-Net enables the realization of spatial features at different locations, improving the model's fitting ability. The pseudocode of the Attention-Unet uncertainty surrogate model method is shown in Algorithm 1. The Attention-Unet uncertainty surrogate model method proposed in this invention can reduce the number of training groups and quickly obtain uncertainty surrogate models to assist in the multi-objective robust optimization of reactive resonant shielding coils.

[0052]

[0053] Furthermore, this invention employs MOWOA (Multi-Objective Whale Optimization Algorithm) for robust optimization design. As a metaheuristic algorithm, MOWOA searches for the optimal solution by simulating the hunting behavior of whales, enabling efficient searching across the entire solution space and achieving fast convergence. Simultaneously, it balances the weights of each objective through chaotic mapping and adaptive weighting mechanisms, and its simple structure and few parameters result in high optimization efficiency. Its optimization process mainly consists of three stages, as follows: (1) Searching for prey: In the early stages of the algorithm, individual whales search for potential prey locations based on their own position and the fitness of the surrounding environment, as shown below: Equation 17: ; Formula 18: ; In the formula, X ( t +1) indicates the updated whale location; X r ( t The location of the whale is randomly selected by the population. A andC For coefficient vectors; D The distance between the individual whale and the optimal solution; t This represents the current iteration number; X ( t () indicates the current location of the whale.

[0054] (2) Encircling the prey: In the middle stage of the algorithm, individual whales will swim towards the prey's location (the better solution) to encircle it. During this process, the whale will form a bubble net underwater, and then quickly swim to the surface to capture the prey, as shown below: Formula 19: ; Formula 20: ; In the formula, z It is a constant that defines the shape of the spiral. X z ( t () represents the optimal whale location obtained so far.

[0055] (3) Bubble Web Attack: In the later stages of the algorithm, the whale captures its prey through specific foraging behaviors (bubble web attack), as shown below: Equation 21: ; Equation 22: ; In the formula, l It is a random floating-point number in the interval (-1, 1).

[0056] MOWOA achieves efficient search for the optimal solution through the three stages described above, making it particularly suitable for multi-objective optimization scenarios. First, the population is initialized and evaluated. Then, the Attention-Unet uncertainty surrogate model is embedded as the objective function in the optimization process. Through MOWOA's prey search, prey encirclement, and bubble net attacks, better individuals in the population are selected, thus outputting the Pareto front and ultimately obtaining the robust optimal structure of the composite reactive resonant shielded coil. This invention combines the Attention-Unet uncertainty surrogate model with MOWOA to achieve efficient and accurate robust optimization design of the composite reactive resonant shielded coil structure.

[0057] To achieve robust optimization design of the reactive resonant shielded coil in a high-efficiency EV-WPT system and protect human electromagnetic exposure safety, this invention uses a surrogate model based on the Attention-Unet method as the optimization objective function of MOWOA. The MOWOA pseudocode is shown in Algorithm 2. p It is a random number.

[0058]

[0059] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0060] Example 1: Simulation and optimization verification of composite reactive resonant shielded coil in EV-WPT system; I. Construction of a refined simulation model; This invention establishes a simulation model of the EV-WPT system and a composite reactive resonant shielded coil. Specific parameters are as follows: vehicle dimensions are 4m × 2m × 1.5m (suitable for family vehicles); transmitting coil... Tx The dimensions are 0.32m × 0.32m, and the receiving coil is... Rx The dimensions are 0.65m × 0.5m, both with 18 turns, using Litz coils with a wire diameter of 0.007m; the system operates at a frequency of 85kHz and a power of 11kW. Due to the receiving coil... Rx Fixed to the chassis of the electric vehicle, the composite reactive resonant shielded coil is placed in the receiving coil. Rx The four directions are used to shield the LMF in each direction. Human exposure scenarios and uncertainty factors are labeled as follows: Figure 6 As shown in (a) and (b). Furthermore, this invention establishes a detailed human body simulation model of a human body with a height of 1.75m (average male height) and some of its organs (brain, lungs), as shown in... Figure 7 As shown in (a) and (b).

[0061] The simulation model established in this invention runs on an Intel 13900K processor, requiring 11 minutes and 45 seconds for each calculation to obtain the corresponding simulation results. Based on the original data from the FEM simulation, an Attention-Unet uncertainty surrogate model is established using the parameters in Tables 1 and 2. The model output is the system energy transfer efficiency. η The maximum value of the induced electric field intensity in three parts of the human body (lungs, torso, and brain). E max .

[0062] II. Validation of the Attention-Unet Uncertainty Proxy Model; To verify the prediction accuracy of the Attention-Unet uncertain surrogate model, 30 sets of test data were selected from the parameters in Tables 1 and 2. When using the Attention-Unet uncertain surrogate model, FEM simulation, a traditional backpropagation (BP) neural network, and a gated recurrent unit (GRU) network were used as control groups to calculate the corresponding test datasets. η and three parts of the human body E max The comparison results are as follows: Figure 8As shown in (a)-(d), the prediction performance is measured by the mean absolute error (MAE), as shown in Table 3.

[0063] Table 3 MAE of different methods

[0064] Figure 8 The data in Table 3 show that the Attention-Unet uncertainty surrogate model is consistent with the results of FEM simulation, while the BP neural network and GRU neural network have significantly poor fitting performance. In the Attention-Unet section of Table 3, the system energy transfer efficiency... η The MAE is 0.0031, proving that the Attention-Unet uncertain surrogate model has accurate predictive performance. On the other hand, three parts of the human body... E max The MAE values ​​were 0.0112, 0.0514, and 0.0113, respectively, slightly higher than... η The MAE is low because of the strong nonlinearity and complexity between the LMF value and the uncertainty factors, but this accuracy is still acceptable. Therefore, the Attention-Unet uncertainty surrogate model established in this invention can replace FEM simulation as the objective function of robust optimization design, achieving an equivalent substitution for simulation.

[0065] III. Robust optimization based on the Attention-Unet uncertain proxy model; Based on the optimization parameters set in Table 1 and the uncertainty factors in Table 2, the tested Attention-Unet uncertainty surrogate model is used as the objective function for optimization, and the Pareto front of the optimization result is obtained, as follows: Figure 9 As shown in Table 4, the robust optimal parameters of the composite reactive resonant shielded coil obtained by the two methods (Attention-Unet and FEM) and the computation time of the two methods are shown in Table 4.

[0066] Table 4 Robust optimal parameters and computation time

[0067] Table 4 shows that the robust optimization design method based on the Attention-Unet uncertain surrogate model can reduce computation time by 80.06% compared to 10,000 FEM simulations while maintaining computational accuracy. This demonstrates that the method of this invention can improve the computational efficiency of the optimization process and significantly reduce the time cost of optimization design.

[0068] IV. Verification of the effectiveness of the composite reactive resonant shielded coil; To verify the effectiveness of the robust optimal structure and design method of the composite reactive resonant shielded coil obtained in the aforementioned work, and considering actual human electromagnetic exposure scenarios, this invention focuses on three parts of the human body. E max As a standard for measuring the shielding effect of composite reactive resonant shielded coils. According to ICNIRP 2010, the limits for public induced electric field strength. E lim The calculation formula is: Equation 23: ; In the formula, f The frequency of the electromagnetic field. The operating frequency of the EV-WPT system of this invention is 85kHz, therefore... E lim =11.475V / m.

[0069] The magnetic flux density distribution around the EV-WPT system before and after placing the composite reactive resonant shielding coil is as follows: Figure 10 As shown. Figure 10 (a) shows the magnetic flux density distribution without shielding. The overall trend is that the magnetic flux density is high in the middle area (red area) and gradually decreases towards the periphery, while the magnetic flux density decays relatively slowly at the edges. Figure 10 Figure (b) shows the magnetic flux density distribution with shielding. The high magnetic flux density region (red area) is significantly reduced, and the decrease in magnetic flux density at the edges is more significant. The overall magnetic flux density distribution is more concentrated and leakage to the outside is reduced. Taking the induced electric field intensity in human tissue as the analysis object, the induced electric field intensity in human tissue before and after shielding is compared, such as... Figure 11 (a) and (b) Figure 12 As shown in (a) and (b), the induced electric field distribution is more pronounced in the legs (part of the human torso), which are closest to the EV-WPT system. The images show the distribution of the induced electric field in three parts of the human body (lungs, torso, and brain) before and after shielding. E max As shown in Table 5.

[0070] Table 5. Shielding effect and system energy transfer efficiency before and after shielding

[0071] Table 5 data shows that the human body was shielded before the experiment. E max Exceeding the standard, due to the distance between the brain and the brain. E max The standard was not exceeded; the human body was shielded. E max Reduced by 61.72%, brain E max A decrease of 36.19%, lungs E maxThe efficiency was reduced by 40.39%, fully meeting the standard limits. Furthermore, through compensation by the composite reactive resonant shielded coil, the system energy transfer efficiency was improved. η It increased by 11.79%.

[0072] However, in practical applications, errors in both vehicle and human positioning are unavoidable, leading to uncertainty in the electromagnetic exposure dose to the human body. Based on the robust optimal structure of the aforementioned composite reactive resonant shielded coil and the distribution of uncertainty factors, the Attention-Unet uncertainty surrogate model is used to calculate the human body's... E max 、 brain E max With lungs E max The probability distribution function (PDF), such as Figure 13 As shown in (a)-(d), the statistical characteristics are shown in Table 6.

[0073] Table 6 Statistical Characteristics

[0074] Figure 13 As shown in Table 6, under the influence of uncertainties, without a composite reactive resonant shielded coil, E max There is a 43.09% probability that the human body will exceed the standard limit. E max The average value is 12.67 V / m. After shielding, even under the influence of uncertainties, the human body... E max It will still not exceed the standard limit, with an average value of 6.54V / m, and the system energy transfer efficiency is also good. η The average increased by 6.46%. E max The probability of exceeding the standard limit drops to 0%. It can be seen that, considering uncertainties, the composite reactive resonant shielded coil can still protect the human body. E max The mean decreased by 48.38%, lung E max The average value decreased by 41.73%, which proves the effectiveness of the composite reactive resonant shielded coil proposed in this invention.

[0075] V. Analysis of the impact of input parameters (MOAT method); In order to qualitatively measure the impact of input parameters and select those that affect the above four indicators (system energy transfer efficiency), η Human body Emax brain E max lungs E max For input variables with significant impact, this invention employs the MOAT (Morris-One-at-a-Time) method, substituting it into the Attention-Unet uncertain surrogate model for solution. The results are as follows: Figure 14 As shown in (a)-(d). From Figure 14 It is known that the capacitance of the composite reactive resonant shielded coil is the most influential factor on the system's energy transmission efficiency; while for the human body (including the torso, brain, and lungs), the vehicle's parking location has the greatest impact. Therefore, drivers or passengers should maintain a certain distance around the EV-WPT system. Furthermore, since the composite reactive resonant shielded coil can enhance the coupling between the coils in the EV-WPT system, it can effectively reduce the LMF (Limited Metal Flow Factor), which also demonstrates the effectiveness of the composite reactive resonant shielded coil proposed in this invention.

[0076] This invention proposes a composite reactive resonant shielded coil structure for EV-WPT systems. This structure comprises a reactive resonant shielded coil, a nanocrystalline layer, and an aluminum plate, achieving a reduction in magnetic flux density in non-working areas while maintaining high system energy transfer efficiency. To mitigate the impact of uncertainties on human electromagnetic exposure safety, this invention proposes a fast multi-objective optimization method based on Attention-Unet deep learning for robust optimization design of the composite reactive resonant shielded coil structure. Simulation tests verify that after applying this composite reactive resonant shielded coil structure, the maximum induced electric field intensity in the human body and lungs (…) is reduced. E max The electromagnetic radiation dose (LMF) of the EV-WPT system decreased by 61.72% and 40.39% respectively, while the system energy transfer efficiency improved by 11.79%. Simultaneously, this invention established an uncertainty surrogate model for the relationship between the human body's internal electromagnetic exposure dose and the system's energy transfer efficiency, and integrated MOWOA to achieve robust optimization considering uncertainties, significantly improving optimization efficiency and reducing computation time from 1958.3 hours to 156.8 hours. To evaluate the impact of the EV-WPT system's LMF on the safety of human electromagnetic exposure, this invention constructed a refined human body simulation model and quantified the impact of input parameters on the system's energy transfer efficiency and the efficiency of different parts of the human body using the MOAT method. E max The degree of influence. Simulation experiments show that, considering uncertainties, the composite reactive resonant shielded coil structure can still protect the human body. E max The average value decreased by 48.38%. Furthermore, without using the composite reactive resonant shielded coil structure, E maxThe probability of exceeding the ICNIRP 2010 limit is 43.09%; however, after applying this structure, E max The probability of exceeding the standard was reduced to 0%, and the average system energy transmission efficiency was improved by 6.46%, further demonstrating the effectiveness and robustness of the composite reactive resonant shielded coil structure.

[0077] The above are merely preferred embodiments of the present invention. It should be noted that those skilled in the art can make several modifications and improvements without departing from the concept of the present invention, and these should also be considered within the scope of protection of the present invention. These modifications and improvements will not affect the effectiveness of the implementation of the present invention or the practicality of the patent.

Claims

1. A composite reactive resonant shielded coil structure, applied in a wireless charging system for electric vehicles, characterized in that, The system includes a reactive resonant shielding coil, a nanocrystalline layer, and an aluminum plate. The nanocrystalline layer is adhered to the periphery of the reactive resonant shielding coil, and the aluminum plate is adhered to the nanocrystalline layer and located outside the reactive resonant shielding coil. The reactive resonant shielding coil generates a reverse canceling magnetic field through leakage magnetic field to weaken the electromagnetic field strength in the non-working area of ​​the electric vehicle wireless charging system. The nanocrystalline layer and aluminum plate are used to attract the leakage electromagnetic field to the vicinity of the reactive resonant shielding coil through their high magnetic permeability characteristics, thereby increasing the strength of the reverse magnetic field and reducing the strength of the external magnetic field, thus reducing the electromagnetic exposure dose to the human body.

2. The composite reactive resonant shielded coil structure according to claim 1, characterized in that, The relative permeability of the nanocrystalline layer is 22000 under operating conditions at room temperature and 85 kHz.

3. The composite reactive resonant shielded coil structure according to claim 1, characterized in that, The reactive resonant shielding coil is provided with four coils, including two transverse reactive resonant shielding coils and two longitudinal reactive resonant shielding coils, which are symmetrically arranged on the transverse and longitudinal sides of the receiving coil, respectively; both the transverse and longitudinal reactive resonant shielding coils are connected to a compensation capacitor.

4. The composite reactive resonant shielded coil structure according to claim 3, characterized in that, The distance between the transverse reactive resonant shielding coil and the edge of the receiving coil is 0-0.05m; the distance between the longitudinal reactive resonant shielding coil and the edge of the receiving coil is 0-0.05m. The compensation capacitor of the transverse reactive resonant shielding coil is 500-1500nF; the compensation capacitor of the longitudinal reactive resonant shielding coil is 500-1500nF.

5. A robust optimization method for the composite reactive resonant shielded coil structure according to any one of claims 1-4, characterized in that, include: An Attention-Unet uncertainty surrogate model is constructed, with design variables and uncertainty factors as inputs, and the system energy transmission efficiency, the maximum induced electric field strength of the human body, the maximum induced electric field strength of the brain, and the maximum induced electric field strength of the lungs as outputs. The design variables include the distance between the lateral reactive resonant shielding coil and the edge of the receiving coil, the distance between the longitudinal reactive resonant shielding coil and the edge of the receiving coil, and the compensation capacitance of the lateral and longitudinal reactive resonant shielding coils. The uncertainty factors include the lateral offset of the human body, the longitudinal offset of the human body, the lateral offset of the vehicle, the longitudinal offset of the vehicle, the distance between the transmitting coil and the receiving coil, and the human body orientation deflection angle. Based on the output of the Attention-Unet uncertain surrogate model, a multi-objective whale optimization algorithm is used to perform multi-objective optimization, solve the Pareto front, and obtain robust optimal parameters.

6. The robust optimization method according to claim 5, characterized in that, The core framework of the Attention-Unet uncertain agent model is the Attention-Unet network, which is formed by introducing the Attention Gate mechanism on the basis of the U-Net network. The U-Net network is a fully convolutional deep network model based on encoder-decoder. The input data is normalized before entering the Attention-Unet network; First, the data passes through an encoder, which consists of convolutional layers and downsampling layers. The convolutional operation automatically extracts local spatial or sequence features from the input data, and a ReLU activation function is applied after convolution. After passing through the convolutional layers, the downsampling layer... MaxPool The algorithm slides through windows, retaining the maximum value within each window, progressively compressing information to preserve the most critical features. After processing through a downsampling layer, the information enters the Attention Gate, which controls the transmission of encoder features to the decoder. The decoding path first passes through an upsampling layer, then concatenates with the output of the Attention Gate, and finally performs convolution and ReLU activation. Finally, it is mapped to the output dimension through another convolution to obtain the final output. The set of is : ； In the formula, This is the set of predicted outputs of the model; W final These are the final convolution kernel parameters; The feature input before entering the convolutional mapping; ρ For bias.

7. The robust optimization method according to claim 5, characterized in that, The multi-objective whale optimization algorithm updates the population position through three stages: prey search, prey encirclement, and bubble web attack. Its position update formula includes: Hunting phase: ； ； In the formula, X ( t +1) indicates the updated whale location; X r ( t The location of the whale is randomly selected by the population. A and C For coefficient vectors; D The distance between the individual whale and the optimal solution; t This represents the current iteration number; X ( t () indicates the current location of the whale; The stage of surrounding the prey: ； ； In the formula, z It is a constant that defines the shape of the spiral; X z ( t () represents the optimal whale location obtained so far; Bubble web attack phase: ； ； In the formula, l It is a random floating-point number in the interval (-1, 1).

Images