Optimization Design Method for the Surrounding Shielding Coil of a Wireless Charging Device for Electric Vehicles

Abstract

The present invention is applicable to the technical field of wireless power transmission, and provides an optimized design method for a surrounding shielding coil of an electric vehicle wireless charging device, including: designing a surrounding shielding coil structure; on the basis of the surrounding shielding coil structure, using an extreme learning neural network to build a surrogate model between the coil structure parameters and the optimization objectives, and then embedding a multi-objective NSGA-II algorithm to optimize the coil structure. The present invention proposes a surrounding active shielding coil structure for the electromagnetic exposure safety problem of an electric vehicle wireless power transmission system, effectively reducing the leakage magnetic field intensity outside the system coupling mechanism, and at the same time reducing the negative impact on the system transmission performance; the surrogate model constructed by the present invention realizes the accurate prediction of the optimization target value of the wireless power transmission system; the present invention combines the multi-objective NSGA-II algorithm to realize the synchronous improvement of multiple performance indicators of the wireless power transmission system.

Description

Technical Field

[0001] The present invention belongs to the technical field of wireless power transmission, and particularly relates to an optimized design method for a surrounding shielding coil of an electric vehicle wireless charging device. Background Technique

[0002] With the increasing marketization of high-power wireless charging devices for electric vehicles, their electromagnetic exposure safety issues have attracted people's attention and become the focus of research institutions at home and abroad. Improving the electromagnetic exposure environment of wireless power transmission for electric vehicles is crucial for the application and popularization of this technology. Therefore, it is particularly necessary to conduct in-depth research on electromagnetic shielding technology for wireless power transmission systems. To ensure the safe use of electric vehicle wireless charging devices, some international organizations such as the International Commission on Non-Ionizing Radiation Protection (ICNIRP) have formulated strict standards for relevant electromagnetic exposure indicators.

[0003] In the field of electromagnetic shielding technology for wireless power transmission, the traditional active shielding principle is to connect a single coil in series with the main coil of wireless power transmission in the reverse direction, and use the current reverse-phase characteristic of the reverse-connected coil to cancel the leakage magnetic field. Although this method reduces the leakage magnetic field outside the system coupling mechanism, it also weakens the magnetic field in the transmission area, which has an adverse impact on the transmission performance of the system. In addition, most of the existing research on the optimized design of shielding structures only focuses on the improvement of single-target performance, which is of little significance for improving the comprehensive performance of the system. For this reason, the present invention proposes an optimized design method for a surrounding shielding coil of an electric vehicle wireless charging device. Summary of the Invention

[0004] The purpose of the present invention is to provide an optimized design method for a surrounding shielding coil of an electric vehicle wireless charging device, aiming to solve the problems proposed in the above background technique.

[0005] The purpose of the present invention is achieved through the following technical solutions:

[0006] An optimized design method for a surrounding shielding coil of an electric vehicle wireless charging device includes the following steps:

[0007] Step 1: Build a surrounding shielding coil;

[0008] Build a surrounding active shielding coil structure in electromagnetic numerical calculation software. The surrounding active shielding coil structure is formed by placing four independently powered shielding coils on the four sides of the transmitting coil of the wireless charging device, and the current of the shielding coil is in the same frequency and opposite phase as the current of the transmitting coil.

[0009] Step 2: Establish a surrogate model of the shielding coil structure;

[0010] Build a neural network framework based on the extreme learning machine, taking the shielding coil structure parameters as the optimization variables, and the wireless power transfer efficiency and the maximum leakage magnetic flux density on the peripheral observation surface of the system coupling mechanism as the optimization objectives, and establish a surrogate model between the optimization variables and the optimization objectives;

[0011] Step 3: Multi-objective optimization design of the shielding coil structure;

[0012] On the basis of the constructed surrogate model, embed the multi-objective NSGA-II algorithm to search for the optimal solution of the shielding coil structure parameters in the optimization variable interval.

[0013] Furthermore, in the above step 1, the power transfer coil of the wireless charging device is a rectangular coil, and the analytical formula of the spatial magnetic induction intensity of the rectangular coil in the rectangular coordinate system is as follows:

[0014] ;

[0015] In the formula: respectively represent the components of the magnetic induction intensity at the spatial coordinate point in the x, y, z direction; represents the magnetic permeability in vacuum; I represents the amplitude of the alternating linear current; z represents z coordinate value; represents the distance from the corner of the rectangular linear current to the spatial coordinate point; i represents the serial number of the corner of the rectangular linear current; represents the function of the half side length of the rectangular coil and the x coordinate value; represents the function of the half side length of the rectangular coil and the y coordinate value;

[0016] In the formula is further expressed as:

[0017] ;

[0018] Formula 6: , , , ;

[0019] In the formula: respectively represent the distances from the four corners of the rectangular linear current to the analyzed spatial coordinate point P ; both represent the functions of the geometric parameters of the rectangular coil and the spatial coordinate parameters; respectively represent the two half side length values of the rectangular linear current in the XOY plane; x and y respectively represent x ,y Coordinate value;

[0020] Assume that in a wireless power transfer coupling mechanism, the number of turns of the coil groups on the transmitting side and the receiving side are n and m respectively. Then, the magnetic field intensity generated by each single-turn coil at the spatial point P is expressed as B t,1 , B t,2 , B t,n and B r,1 , B r,2 , B r,m . Its total field strength can be expressed as:

[0021] ;

[0022] In the formula: represents the total field strength, represents the magnetic field intensity generated by a single-turn coil on the transmitting side at the spatial point P , represents the magnetic field intensity generated by a single-turn coil on the receiving side at the spatial point P .

[0023] Furthermore, in the circumferential active shielding coil structure, assume that the magnetic flux component generated by the transmitting coil of the wireless charging device is Φ 3, Φ 6, the magnetic flux component generated by one side shielding coil is Φ 1, Φ 2, and the magnetic flux component generated by the other side shielding coil is Φ 4, Φ 5. Then, there are:

[0024] ;

[0025] ;

[0026] ;

[0027] ;

[0028] In the formula: and represent the total magnetic field intensity generated by one side shielding coil and the transmitting coil in the transmission area and the peripheral area respectively, and represent the total magnetic field intensity generated by the other side shielding coil and the transmitting coil in the transmission area and the peripheral area respectively.

[0029] Further, the specific process of step 2 is as follows: taking the length l , width ω , and thickness d of the shielding coil as optimization variables, and taking the system transmission efficiency η , the maximum leakage magnetic flux density B 1 on the observation surface 1 m away from the wireless charging device on the side of the vehicle body, and the maximum leakage magnetic flux density B 2 on the observation surface 1 m away from the wireless charging device at the rear of the vehicle body as optimization objectives; selecting leakage magnetic field observation positions on the side and at the rear of the vehicle body, solving using electromagnetic numerical simulation finite element method, and based on different variable combinations { l , ω , d}, extracting the corresponding η , B 1, and B 2 sample sets; constructing a surrogate model based on the extreme learning machine neural network framework, and training the surrogate model using the extracted sample sets.

[0030] Further, for each input sample , the activation value of the neurons on the hidden layer of the extreme learning machine is expressed as:

[0031] ;

[0032] In the formula: represents the activation function, represents the transpose of the input weight vector of the l th neuron, represents the corresponding bias term; the extreme learning machine calculates the optimal weight value from the hidden layer to the output layer, and its output is expressed in the form of a linear combination:

[0033] ;

[0034] ;

[0035] In the formula: represents the weight matrix; represents the weights of each node in the hidden layer; represents the network output value; L represents the total number of neuron nodes in the hidden layer; represents the weight of the l th node; represents the transpose of the weight matrix; represents the hidden layer node activation value matrix;

[0036] The optimal weight matrix is derived according to the following formula , to minimize the overall prediction error value of the sample:

[0037] ;

[0038] In the formula: H represents the hidden layer activation value matrix, T represents the target output matrix, represents H the Moore-Penrose generalized inverse matrix of

[0039] Furthermore, the specific process of step 3 is as follows: Based on the trained surrogate model, the multi-objective NSGA-II algorithm is embedded to optimize the coil structure variable range, obtain the Pareto optimal solution set, and find the optimal solution with the maximum transmission efficiency and the minimum electromagnetic exposure index according to the Pareto optimization result.

[0040] Furthermore, combining the surrogate model and the multi-objective NSGA-II algorithm, the multi-objective optimization framework of the shielding coil structure is expressed as:

[0041] ;

[0042] In the formula: represents the total objective function, represents the length of the shielding coil, represents the width of the shielding coil, represents the thickness of the shielding coil, respectively represent the objective functions of transmission efficiency, the maximum leakage magnetic flux density on the side observation surface of the vehicle body, and the maximum leakage magnetic flux density on the rear observation surface of the vehicle body, respectively represent the distribution weights of

[0043] Assume that the optimized design variable group is , then the optimization range of each variable is expressed as follows:

[0044] ;

[0045] ;

[0046] ;

[0047] ;

[0048] In the formula: respectively represent the optimization ranges of the length, width, and thickness of the shielding coil.

[0049] Compared with the prior art, the beneficial effects of the present invention are:

[0050] Aiming at the electromagnetic exposure safety problem of the wireless power transfer system for electric vehicles, the present invention proposes a surrounding active shielding coil structure, which effectively reduces the leakage magnetic field intensity outside the system coupling mechanism and simultaneously reduces the negative impact on the transmission performance. On this basis, the present invention constructs an optimization surrogate model of the shielding coil structure based on the extreme learning machine, which realizes the accurate prediction of the optimization target value of the wireless power transfer system. Then, combined with the multi-objective NSGA-II algorithm, the optimization design of the shielding coil structure is carried out, realizing the synchronous improvement of multiple performance indicators of the wireless power transfer system. Description of the Drawings

[0051] Figure 1 It is the flowchart of the method of the present invention.

[0052] Figure 2 It is the rectangular line current model.

[0053] Figure 3 It is the schematic diagram of the arrangement of the surrounding active shielding coils.

[0054] Figure 4 It is the schematic diagram of the cross-sectional magnetic flux line distribution.

[0055] Figure 5 It is the area of concern for the leakage magnetic field around the vehicle body.

[0056] Figure 6 It is the multi-objective optimization design process of the shielding coil structure.

[0057] Figure 7 It is the simulation model of the wireless power transfer for electric vehicles.

[0058] Figure 8 It is the distribution of the leakage magnetic flux density on Observation Plane 1; where (a) is the leakage magnetic flux density on Observation Plane 1 without shielding, and (b) is the leakage magnetic flux density on Observation Plane 1 with active shielding.

[0059] Figure 9 It is the distribution of the leakage magnetic flux density on Observation Plane 2; where (a) is the leakage magnetic flux density on Observation Plane 2 without shielding, and (b) is the leakage magnetic flux density on Observation Plane 2 with active shielding.

[0060] Figure 10 It is the Pareto optimization result.

[0061] Figure 11 It is the comparison of the probability distributions of the transmission efficiency before and after multi-objective optimization in the actual scenario.

[0062] Figure 12 It is the comparison of the probability distributions of the maximum leakage magnetic flux density on Observation Plane 1 before and after multi-objective optimization in the actual scenario.

[0063] Figure 13Comparison of the probability distribution of the maximum leakage magnetic flux density on the observation surface 2 before and after multi-objective optimization in the actual scenario. Detailed implementation mode

[0064] For a clearer understanding of the technical features, objectives, and beneficial effects of the present invention, the technical solution of the present invention will be described in detail below, but it should not be construed as a limitation on the implementable scope of the present invention.

[0065] The following describes the specific implementation of the present invention in detail with reference to specific embodiments.

[0066] An embodiment of the present invention provides an optimized design method for the surrounding shielding coil of an electric vehicle wireless charging device, and its flowchart is as Figure 1 shown, and the method includes the following steps:

[0067] Step 1: Build a surrounding shielding coil.

[0068] Build a surrounding active shielding coil structure in the electromagnetic numerical simulation software. Its basic architecture consists of four shielding coils placed on the four sides of the transmitting coil of the wireless charging device. The shielding coils are in the form of independent power supplies, and the shielding current is in the same frequency and opposite phase as the current of the transmitting coil.

[0069] Step 2: Establish a surrogate model of the shielding coil structure.

[0070] Build a neural network framework based on the extreme learning machine, with the length l , width ω and thickness d of the shielding coil as the optimization variables, and the transmission efficiency η , the maximum leakage magnetic flux density on the observation surface 1 m away from the wireless charging device on the side of the vehicle body B 1 and the maximum leakage magnetic flux density on the observation surface 1 m away from the wireless charging device at the rear of the vehicle body B 2 as the optimization objectives; select the leakage magnetic field observation positions on the side and rear of the vehicle body, use the electromagnetic numerical simulation finite element solution, and based on different variable combinations { l , ω , d}, extract the corresponding η , B 1 and B 2 sample sets; use the extreme learning machine neural network to build a surrogate model, and use the extracted sample sets to train the surrogate model.

[0071] Step 3: Multi-objective optimization design of the shielding coil structure.

[0072] Based on the trained proxy model, the multi-objective NSGA-II algorithm is incorporated to optimize the variable range, obtain the Pareto optimal solution set, and find the optimal solution with the maximum transmission efficiency and the minimum electromagnetic exposure index.

[0073] The specific process is as follows:

[0074] Step 1: Build a surrounding shielding coil.

[0075] 1.1 Spatial magnetic field distribution characteristics of the wireless charging coil

[0076] The magnetic energy coil group is the main medium for transmitting electrical energy in a wireless charging system. It mainly relies on the spatial alternating magnetic field to achieve wireless power transmission. Therefore, the magnetic energy coil group largely determines the magnetic field distribution in the transmission channel and the leakage magnetic field distribution in the surrounding environment. In the application of magnetic coupling wireless power transmission, the coil group is usually composed of multiple independent insulated stranded wires wound together. According to the magnetic field distribution pattern, it can be mainly divided into two categories: unipolar coils and bipolar coils. The structure of unipolar coils is relatively simple, such as circular coils and rectangular coils, etc., while the structure of bipolar coils is relatively complex, such as DD coils and BBP coils, etc.

[0077] In this study, a rectangular coil is used as the research object. Assume that the two half-side lengths of the rectangular line current in the XOY plane are respectively a 、 b , I represents the amplitude of the line current. Then, for any point P ( x , y , z ) in the x and y directions, the magnetic vector potential expressions are:

[0078] ;

[0079] ;

[0080] In the formula: and respectively represent the magnetic vector potential of point P ( x , y , z ) in the x and y directions; represents the magnetic permeability in vacuum; respectively represent the distances from the four corners of the rectangular line current to the P point; x and y respectively represent the P point's x 、y Coordinate values; specifically, as shown in Figure 2 .

[0081] Generally, the relationship between the magnetic vector potential and the spatial magnetic induction intensity of a rectangular line current in the air domain can be expressed as follows:

[0082]

[0083] In the formula: respectively represent the components of the magnetic induction intensity at the spatial coordinate point in the direction; z represents z coordinate value;

[0084] By combining Equation 1 - Equation 3, the analytical formula for the magnetic induction intensity of a rectangular coil in the rectangular coordinate system can be obtained as follows:

[0085] ;

[0086] In the formula: represents the distance from the corner of the rectangular line current to the spatial coordinate point P for analysis, i represents the serial number of the corner of the rectangular line current; represents the function of the half - side length of the rectangular coil and the P point x coordinate value; represents the function of the half - side length of the rectangular coil and the P point y coordinate value;

[0087] In Equation 4, can be further expressed as:

[0088] ;

[0089] Equation 6: , , , ;

[0090] In the formula: all represent the functions of the geometric parameters of the rectangular coil and the spatial coordinate parameters of the P point.

[0091] Assume that in the wireless power transfer coupling mechanism, the number of turns of the coil groups on the transmitting side and the receiving side are n and m respectively. Then, the magnetic field intensity generated by each single - turn coil at the P point can be expressed as B t,1 , B t,2 , Bt,n and B r,1 , B r,2 , B r,m , the magnetic flux density at a certain spatial coordinate point generated by a single-turn coil under alternating current power supply is the basis for calculating the total field strength at this point, and can be expressed as follows:

[0092] ;

[0093] In the formula: represents the total field strength, represents the magnetic field strength generated by the single-turn coil on the transmitting side at the spatial point P , represents the magnetic field strength generated by the single-turn coil on the receiving side at the spatial point P .

[0094] 1.2 Surrounding active shielding coil;

[0095] During the application of wireless power transfer for electric vehicles, a large amount of leakage magnetic field will be generated around the coupling mechanism. In order to effectively protect against electromagnetic safety, the present invention places a shielding coil on each of the four sides of the transmitting coil of the wireless charging device, as shown in Figure 3 . All four shielding coils adopt the form of active feeding, and the current direction of the shielding coil is opposite to that of the wireless power transfer transmitting coil. Through this coil layout method, a specific target magnetic flux line direction is generated. Taking the cross-section of the side view of the system coupling mechanism as an example, the overall magnetic flux line distribution after adding the shielding coil is shown in Figure 4 .

[0096] Assume that the magnetic flux component generated by the wireless power transfer transmitting coil is Φ 3, Φ 6, the magnetic flux component generated by one side shielding coil is Φ 1, Φ 2, and the magnetic flux component generated by the other side shielding coil is Φ 4, Φ 5. It can be seen that within the transmission area between the transceiver ends of the wireless charging device, the magnetic field strength shows a superposition effect, which can supplement the transmission performance. While in the peripheral protection area, the magnetic field strength shows a cancellation effect, which can effectively weaken the leakage magnetic field, specifically as shown in Equations 8 - 11:

[0097]

[0098] In the formula: and represent the total magnetic field strength generated by one side shielding coil and the transmitting coil in the transmission area and the peripheral area respectively, and It represents the total magnetic field intensity generated by the shielding coil on the other side and the transmitting coil in the transmission area and the peripheral area respectively.

[0099] Step 2: Establish a proxy model for the shielding coil structure;

[0100] In actual situations, there is a strong non - linear mapping relationship between the structural parameters and electrical performance parameters of a wireless power transfer system. To address the above - mentioned problems, the present invention uses an extreme learning machine neural network to construct a proxy model, achieving accurate prediction of the optimization target values of the wireless power transfer system, mainly including the wireless power transfer efficiency and the leakage magnetic field intensity outside the system coupling mechanism. The selection positions of the leakage magnetic field observation plane are set at the rear of the vehicle body (such as Figure 5 ② in the figure) and the side of the vehicle body (such as Figure 5 ① in the figure), which are areas where people are likely to move in actual scenarios, as specifically shown in Figure 5 the figure.

[0101] Different from the traditional back - propagation neural network, the weights of the hidden layer of the extreme learning machine do not need to be adjusted during the training process. Therefore, it has advantages such as high solution efficiency and strong generalization performance, and is very suitable for fitting complex non - linear function mapping relationships. For each input sample , the activation value of the neurons on the hidden layer of the extreme learning machine can be expressed as:

[0102] ;

[0103] In the formula: represents the activation function, represents the transpose of the input weight vector of the l th neuron, represents the corresponding bias term; these parameter values are all randomly initialized and remain unchanged during the network training and learning process. The primary task of the extreme learning machine is to calculate the optimal weight value from the hidden layer to the output layer, and its output can be expressed in the form of a linear combination:

[0104]

[0105] In the formula: represents the weight matrix; represents the weights of each node in the hidden layer; represents the network output value; L represents the total number of neuron nodes in the hidden layer; represents the l th node's weight; represents the transpose of the weight matrix; represents the matrix of hidden layer node activation values.

[0106] To minimize the overall prediction error value of the samples, the optimal weight matrix can be derived according to the following formula :

[0107]

[0108] In the formula: H represents the hidden layer activation value matrix, T represents the target output matrix, represents H the Moore - Penrose generalized inverse matrix of

[0109] In the surrogate model of the wireless power transfer shielding coil structure, the length l width ω and thickness d of the shielding coil are set as input variables; the output variables are η , B 1, B 2], where η represents the transmission efficiency, B 1 represents the maximum leakage magnetic flux density on the observation plane 1 m away from the wireless charging device on the side of the vehicle body, B 2 represents the maximum leakage magnetic flux density on the observation plane 1 m away from the wireless charging device at the rear of the vehicle body, η, B 1, B 2 are both actual optimization objectives. In this study, the parametric sweep module built into the electromagnetic numerical simulation software COMSOL is used to obtain the sample sets of l , ω , d} corresponding to η , B 1 and B 2, and these are used as the training samples of the surrogate model. According to the row - column dimensions of the optimization variables and objectives, the number of nodes in both the input layer and the output layer is set to 3 in the extreme learning machine network layer.

[0110] Step 3: Multi - objective optimization design of the shielding coil structure;

[0111] The present invention intends to further optimize the design of the surrounding - type shielding coil structure by using the non - dominated sorting genetic algorithm (NSGA - II). The NSGA - II algorithm overcomes the defects existing in the early NSGA algorithm, such as high computational cost and slow execution speed. It has been improved mainly in the following three aspects:

[0112] 1) The NSGA - II algorithm adopts a fast non - dominated sorting method, reducing the complexity of the original NSGA algorithm from Ɵ ( mN 3 ) to Ɵ (mN 2 ) significantly reduces the operation time of the algorithm, where Ɵ represents the complexity, m represents the number of objective functions, N represents the number of populations.

[0113] 2) The elite strategy is adopted to merge the parent individuals and offspring individuals and perform non-dominated sorting, expanding the search space. When generating the next generation of populations, individuals with higher priority are selected according to the sorting, effectively increasing the probability of retaining excellent individuals.

[0114] 3) The NSGA-II algorithm uses the crowding distance strategy to replace the fitness sharing strategy of specifying the sharing radius in the original NSGA algorithm, enhancing the diversity of individuals in the population. This method is beneficial to the selection, crossover, and compilation behaviors of individuals in the entire interval.

[0115] Combined with the extreme learning machine surrogate model and the multi-objective NSGA-II algorithm, the multi-objective optimization framework of the shielding coil structure can be expressed as:

[0116]

[0117] In the formula: represents the total objective function, represents the length of the shielding coil, represents the width of the shielding coil, represents the thickness of the shielding coil, represent the objective functions of transmission efficiency, the maximum leakage magnetic flux density on the side observation surface of the vehicle body, and the maximum leakage magnetic flux density on the rear observation surface of the vehicle body respectively, represent respectively the allocation weights of.

[0118] Assume that the optimization design variable group is , then the optimization interval of each variable can be expressed as follows:

[0119]

[0120]

[0121] In the formula: represent the optimization intervals of the length, width, and thickness of the shielding coil respectively.

[0122] The present invention uses the coil structure variables and optimization target values extracted from electromagnetic numerical simulation software as a training sample set to train an extreme learning machine neural network. When the training error reaches the required accuracy, the training process of the surrogate model is completed. The surrogate model can accurately predict the optimization target values corresponding to different variable groups. Then, test samples are extracted from the coil structure variable group. Based on the extreme learning machine modeling, combined with the multi-objective NSGA-II algorithm, the optimization interval of the shielding coil structure variables is optimized to obtain a Pareto optimal solution set, so as to find the shielding coil structure parameter group with the maximum transmission efficiency and the minimum electromagnetic exposure. The overall process is as Figure 6 shown.

[0123] Example 1: The present invention uses the electromagnetic numerical simulation software COMSOL to build a simulation model of an electric vehicle wireless power transmission system, as Figure 7 shown. The body size parameters are as follows: length 4.5 meters, width 2 meters, height 1.5 meters, which is basically the same as the volume size of common household cars on the market. In the simulation model, the body material is set as aluminum metal, the tire material is set as rubber, and the window material is set as tempered glass. In this study, the wireless power transmission system realizes energy transmission based on the principle of double-coil coupled resonance on the transmitting side and the receiving side. The working frequency is set to 85 kHz. In the simulation model, the coil material is set as copper. Referring to the vertical height of most household car chassis from the ground, the vertical transmission gap between the transmitting and receiving coil groups in the coupling mechanism is set to 0.2 m.

[0124] To test the electromagnetic safety protection performance of the surrounding active shielding coil, this study builds a shielding coil model in the COMSOL software, and sets up leakage magnetic field observation planes 1 m to the side and rear of the wireless charging device, which are respectively set as observation plane 1 and observation plane 2, to observe the distribution of the leakage magnetic field around the vehicle body, and at the same time compare with the simulation calculation results without the surrounding active shielding coil. The results are as Figure 8 and Figure 9 shown. It can be seen that without the protection of the shielding coil, the magnetic field leakage around the coupling mechanism is relatively serious, and the leakage magnetic flux density on the observation plane around the vehicle body is large. After taking shielding protection measures, the electromagnetic exposure around the coupling mechanism has been greatly reduced. Among them, referring to Figure 8 (a) and (b) in it, the maximum leakage magnetic flux density on observation plane 1 drops from 80 μT to 10 μT, a decrease of about 88%; referring to Figure 9 (a) and (b) in it, the maximum leakage magnetic flux density on observation plane 2 drops from 90 μT to 12 μT, a decrease of about 87%. The above results verify that the surrounding active shielding coil proposed by the present invention has excellent electromagnetic shielding effect and can effectively suppress the magnetic field leakage around.

[0125] In the part of the optimized design of the shielding coil, let l , ω , d The optimization intervals of are [330 mm, 350 mm], [90 mm, 100 mm], [3 mm, 8 mm] respectively. In this study, 5,000 sample points were collected within the specified intervals to train the extreme learning machine surrogate model. Based on the trained model, combined with the multi-objective NSGA-II algorithm, the Pareto optimization solution distributions of the system transmission efficiency, the maximum leakage magnetic flux density on observation plane 1 and observation plane 2 were finally obtained, as shown in Figure 10 . The optimal solution of the structure design of the loop-type active shielding coil can be obtained, as shown by the fork-shaped mark in Figure 10 . It can be seen that the optimal solution has higher transmission efficiency and lower leakage magnetic field intensity, meeting the actual needs of electric vehicle users.

[0126] Considering that there are some potential uncertain factors in the wireless charging scenario of electric vehicles, such as the spatial misalignment of the wireless power transmission coil group caused by improper driver operation, which will have an unignorable impact on the performance indicators of the charging system. Therefore, this study also carried out uncertainty quantification analysis to further verify the effectiveness of the proposed shielding scheme in solving practical problems. Combining the actual scenario, in this embodiment of the present invention, Latin hypercube sampling is adopted to collect samples of relevant uncertain variables, specifically including: the horizontal offset Δx , Δy of the main transmitting and receiving coils of the coupling mechanism, the vertical interval h of the main transmitting and receiving coils, and the deflection angle α of the coil group, which are specifically listed in Table 1, where U represents a uniform distribution.

[0127]

[0128] According to the distribution types and distribution intervals of the random input variables set in Table 1, the comparison of the probability density distributions of the performance objectives of the wireless power transmission system before and after multi-objective optimization is calculated, as shown in Figures 11 - 13 . The average values of the transmission efficiency before and after optimization are 0.78 and 0.93 respectively, with an effective increase of 19.2% ( Figure 11 ); the average values of the maximum leakage magnetic flux density on observation plane 1 before and after optimization are 9.35 μT and 6.07 μT respectively, with an effective reduction of 35.1% ( Figure 12 ); the average values of the maximum leakage magnetic flux density on observation plane 2 before and after optimization are 11.37 μT and 7.66 μT respectively, with an effective reduction of 32.6% ( Figure 13). From the above results, it can be concluded that when the wireless power transmission process is interfered by external uncertain factors, the transmission efficiency of the system after multi-objective optimization has a higher probability of being at a high level, and at the same time, the electromagnetic exposure risk also decreases. Therefore, based on the optimized design method of the surrounding shielding coil of the wireless charging device for electric vehicles proposed in the present invention, it can ensure excellent system transmission performance and electromagnetic safety protection efficiency, providing a scientific and effective theoretical basis and solution for engineering designers.

[0129] The above is only the preferred embodiment of the present invention. It should be noted that for those skilled in the art, without departing from the concept of the present invention, several modifications and improvements can be made, which should also be regarded as the protection scope of the present invention, and these will not affect the implementation effect of the present invention and the practicability of the patent.

Claims

1. An optimization design method for a surrounding shielding coil of an electric vehicle wireless charging device, characterized in that: The following steps are involved: Step 1: Build a surrounding shielding coil; A surrounding active shielding coil structure is built in the electromagnetic numerical calculation software. The surrounding active shielding coil structure is formed by placing four shielding coils on the four sides of the transmitting coil of the wireless charging device. The shielding coil current is in the same frequency and opposite phase to the transmitting coil current. Step 2: Establish a proxy model of the shielding coil structure; A neural network framework based on extreme learning machine was built, with the shielding coil structural parameters as optimization variables, the wireless power transmission efficiency and the maximum leakage flux density on the peripheral observation surface of the system coupling mechanism as optimization targets, and a proxy model of optimization variables and optimization targets was established; Step 3: Multi-objective optimization design of shielding coil structure; Based on the constructed proxy model, the multi-objective NSGA-II algorithm is used to optimize the optimization variable interval to obtain the optimal solution of the shielding coil structural parameters. In step 1, the energy transmission coil of the wireless charging device is a rectangular coil, and the spatial magnetic induction intensity analytical formula of the rectangular coil in the rectangular coordinate system is as follows: ； Where: They represent the magnetic induction intensity at the spatial coordinate point x y z Directional weight; represents the magnetic permeability in vacuum; I Indicates the amplitude of the alternating line current; z express z Coordinate value; Indicates the distance from the corner of the rectangular line current to the spatial coordinate point; i Indicates the corner number of the rectangular line current; Represents the half side length of the rectangular coil and x Function of coordinate values; Represents the half side length of the rectangular coil and y Function of coordinate values; In the formula r p q Further expressed as: ; Formula 6: ; Where: Represents the spatial coordinate points of the four corners of the rectangular line current to the analysis P distance; All represent functions of rectangular coil geometric parameters and space coordinate parameters; They represent the two half-side lengths of the rectangular line current in the XOY plane; x and y Respectively x , y Coordinate value; Assume that in the wireless power transmission coupling mechanism, the number of turns of the transmitting and receiving coil groups are n and m , then each single-turn coil is at a point in space P The magnetic field strength generated at B t,1 , B t,2 , B t,n as well as B r,1 , B r,2 , B r,m , the total field strength can be expressed as: ; Where: is the total field strength, Indicates that the transmitting side single-turn coil is at the spatial point P The magnetic field strength generated at Indicates that the single-turn coil on the receiving side is at the spatial point P The magnetic field strength generated at .

2. The optimization design method of the surrounding shielding coil of the wireless charging device of an electric vehicle according to claim 1 is characterized in that: In the surrounding active shielding coil structure, it is assumed that the magnetic flux component generated by the transmitting coil of the wireless charging device is Φ 3. Φ 6. The magnetic flux component generated by the shielding coil on one side is Φ 1. Φ 2. The magnetic flux component generated by the shielding coil on the other side is Φ 4. Φ 5, then: ； ； ； ； Where: and It represents the sum of the magnetic field strengths generated by the shielding coil and the transmitting coil on one side in the transmission area and the peripheral area respectively. and It represents the sum of the magnetic field strengths generated by the shielding coil and the transmitting coil on the other side in the transmission area and the peripheral area respectively.

3. The optimization design method of the surrounding shielding coil of the wireless charging device of an electric vehicle according to claim 1 is characterized in that: The specific process of step 2 is: l ,width ω And thickness d As the optimization variable, the system transmission efficiency η , Maximum leakage magnetic flux density on the observation surface 1m away from the wireless charging device on the side of the vehicle body B 1 and the maximum leakage magnetic flux density on the observation surface 1m away from the wireless charging device behind the vehicle body B 2 as the optimization target; The leakage magnetic field observation positions are selected on the side and rear of the vehicle body, and the electromagnetic numerical simulation finite element solution is used. Based on different variable combinations { l , ω , d }, extract the corresponding η , B 1 and B 2; build a proxy model based on the extreme learning machine neural network framework, and use the extracted sample set to train the proxy model.

4. The optimization design method of the surrounding shielding coil of the wireless charging device of an electric vehicle according to claim 1 is characterized in that: For each input sample , the activation value of the neuron on the hidden layer of the extreme learning machine It is expressed as: ； Where: (·) represents the activation function, Indicates l The transpose of the input weight vector of the neuron, Represents the corresponding bias term; the extreme learning machine calculates the optimal weight value from the hidden layer to the output layer, and its output is expressed as a linear combination form: ； ； Where: represents the weight matrix; Represents the weight of each node in the hidden layer; Represents the network output value; L Represents the total number of neuron nodes in the hidden layer; Indicates l The weight of each node; represents the transpose of the weight matrix; Represents the hidden layer node activation value matrix; The optimal weight matrix is ​​derived as follows , in order to minimize the overall prediction error value of the sample: ; Where: H represents the hidden layer activation value matrix, T represents the target output matrix, express H The Moore-Penrose generalized inverse matrix of .

5. The optimization design method of the surrounding shielding coil of the wireless charging device of an electric vehicle according to claim 1 is characterized in that: The specific process of step 3 is as follows: based on the trained proxy model, the multi-objective NSGA-II algorithm is used to optimize the coil structure variable interval to obtain the Pareto optimization solution set, and the optimal solution with the maximum transmission efficiency and the minimum electromagnetic exposure index is found according to the Pareto optimization result.

6. The optimization design method of the surrounding shielding coil of the wireless charging device of an electric vehicle according to claim 1 is characterized in that: Combining the agent model and the multi-objective NSGA-II algorithm, the multi-objective optimization framework of the shielding coil structure is expressed as: ; Where: represents the overall objective function, represents the length of the shielding coil, represents the width of the shielding coil, Indicates the thickness of the shielding coil, They represent the objective functions of transmission efficiency, maximum leakage flux density on the side observation surface of the vehicle body, and maximum leakage flux density on the rear observation surface of the vehicle body, respectively. Respectively The allocation weight of Assume that the optimal design variable set is , then the optimization interval of each variable is expressed as follows: ； ； ； ； Where: They represent the optimization ranges of the shielding coil length, width and thickness respectively.

Images