Wireless charging system and method for electric vehicle

Abstract

The invention relates to the technical field of electric vehicle wireless charging, in particular to an electric vehicle wireless charging system and method, and the system comprises a hierarchical architecture construction module which is used for constructing a hierarchical architecture comprising perception collection, data synchronization, local decision, edge diagnosis, cloud analysis and control execution layers, and interfaces of the layers interact; the differential sensing acquisition module is used for differentially arranging a magnetic field sensor, an impedance spectroscopy probe and a temperature and environment sensor based on the distribution of an electromagnetic field and a thermosensitive area of a key component, and acquiring a multi-modal signal; and the signal synchronous processing module is used for carrying out anti-interference processing by using the multi-mode signal, unifying a time sequence through clock synchronization, checking integrity and reconstructing abnormity. According to the electric vehicle wireless charging system, the layered architecture module and the differential sensing acquisition module are constructed, so that the problem that fault diagnosis is not timely due to information isolation and response lag caused by the fact that most traditional electric vehicle wireless charging systems adopt single-layer control and single sensing means is solved.

Description

Technical Field

[0001] This invention relates to the field of wireless charging technology for electric vehicles, and more particularly to a wireless charging system and method for electric vehicles. Background Technology

[0002] With the rapid development of electric vehicles, wireless charging technology has received widespread attention due to its convenience and safety. Existing wireless charging systems typically employ electromagnetic induction or magnetic resonance to transfer electrical energy from the charging station to the vehicle's receiver, achieving contactless charging. Electromagnetic induction uses an alternating magnetic field generated by a transmitting coil to induce a current in the receiving coil, completing energy transfer; this method is suitable for low-power and short-distance charging scenarios. Magnetic resonance, on the other hand, utilizes a high-quality resonant coil for energy transfer, enabling high-efficiency charging over relatively long distances while having lower requirements for the alignment of the vehicle and the charging station. Currently, wireless charging systems are gradually being adopted in scenarios such as home garages, public parking lots, and highway service areas, and are increasingly being integrated with smart grids and vehicle energy management systems to achieve charging scheduling and load management.

[0003] Traditional wireless charging systems for electric vehicles mostly employ single-layer control and single-sensor methods. Due to isolated information and delayed response, this leads to problems with untimely fault diagnosis. Summary of the Invention

[0004] To overcome the above shortcomings, this invention provides a wireless charging system and method for electric vehicles, aiming to improve the problem that traditional wireless charging systems for electric vehicles mostly adopt single-layer control and single sensing methods, resulting in untimely fault diagnosis due to isolated information and delayed response.

[0005] In a first aspect, the present invention provides the following technical solution: a wireless charging system for electric vehicles includes: A layered architecture module is constructed to build a layered architecture that includes sensing and acquisition, data synchronization, local decision-making, edge diagnostics, cloud analysis, and control execution layers, with interfaces between each layer interacting. The differential sensing acquisition module is used to collect multimodal signals by differentially arranging magnetic field sensors, impedance spectrum probes, temperature and environmental sensors based on the distribution of electromagnetic fields and thermosensitive areas of key components. The signal synchronization processing module is used for anti-interference processing of multi-mode signals, unifying timing through clock synchronization, verifying integrity, and reconstructing anomalies. The hierarchical collaborative diagnostic module is used for hierarchical diagnosis. The local decision-making layer responds quickly to faults according to rules; the edge diagnostic layer integrates models and data-driven analysis to generate fault confidence; and the cloud maps parameters with the environment using digital twins, optimizes the model through distributed learning, and predicts health trends. The two-level fusion decision module is used to convert multimodal signals into a unified feature vector and synthesize multi-source diagnostic conclusions through evidence fusion algorithms; The adaptive calibration module is used to perform adaptive control and sensor periodic calibration based on multi-source diagnostic conclusions, and dynamically adjust the resonant frequency, compensation parameters or power; the sensors in the differential sensing acquisition module are calibrated through built-in benchmarks or closed-loop self-tests, and the calibration results are used as the two-level fusion weights.

[0006] By adopting the above technical solution, and by constructing a hierarchical architecture module, a differential sensing acquisition module, a signal synchronization processing module, a hierarchical collaborative diagnosis module, a two-level fusion decision module, and an adaptive calibration module, efficient acquisition, processing, and diagnosis of multimodal signals can be achieved. This improves the problem that traditional electric vehicle wireless charging systems mostly use single-layer control and single sensing methods, resulting in untimely fault diagnosis due to information isolation and response lag.

[0007] Furthermore, the construction of the layered architecture module, which includes the sensing and acquisition, data synchronization, local decision-making, edge diagnostics, cloud analysis, and control execution layers, includes the following steps: Define the functions of each layer, including the inputs and outputs of the perception and acquisition layer, data synchronization layer, local decision-making layer, edge diagnostic layer, cloud analysis layer, and control and execution layer; Each layer of functionality is implemented as an independent module, forming a data processing unit that can be called through an interface; Define the data format, transmission protocol, communication frequency, and synchronization method to enable data exchange between different layers; Unified timing management is implemented for data transmission between modules at each layer, using wired or wireless communication methods to ensure complete information transmission; Redundant hardware is used for the local decision-making layer and the edge diagnostic layer, and cross-layer data transmission is encrypted and authenticated; Integrate the modules at each layer to verify the implementation of data flow, interface, and timing synchronization.

[0008] Furthermore, the acquisition of multimodal signals in the difference sensing acquisition module includes the following steps: Determine the placement and sampling sequence of the magnetic field sensor, impedance spectroscopy probe, temperature sensor, and environmental sensor; The magnetic field sensor, impedance spectrum probe, temperature sensor, and environmental sensor were initialized and calibrated to a reference standard, respectively. Use a unified trigger signal or a synchronous clock to simultaneously start sampling from each sensor; The analog signals collected by the sensor are converted into digital signals, and basic filtering and noise reduction processing is performed. The digital signals from various sensors are encoded in a unified format to generate a multimodal signal data structure, and timestamps and sensor identifiers are added. The multimodal signals are buffered at the acquisition node and prepared for transmission to the data synchronization layer or edge diagnostic layer according to the interface protocol.

[0009] Furthermore, the integrity verification and anomaly reconstruction in the signal synchronization processing module includes the following steps: The multimodal signals acquired by the difference sensing acquisition module are supplemented with cyclic redundancy check codes or checksums; At the signal receiving or processing node, integrity checks are performed on the multimodal signal to determine whether there are transmission or acquisition errors. For detected abnormal signals, interpolation or reconstruction processing is performed based on adjacent time steps or historical continuous sampling data; Update the reconstructed signal to the signal buffer or transmission queue and attach a status flag; Record the refactoring operations.

[0010] Furthermore, the hierarchical diagnosis in the hierarchical collaborative diagnosis module includes the following steps: Local decision-making layer processing: Receives multimodal signals output by the signal synchronization processing module, performs threshold judgment and state determination on the signals according to a preset rule table, and generates fault trigger signals; Edge diagnostic layer processing: Perform parameter drift analysis on multimodal signals, identify abnormal parameters using model-driven methods, and input the signals into a lightweight AutoEncoder model to generate anomaly scores; Edge layer fusion: The model-driven analysis results are fused with the data-driven anomaly score to form preliminary fault confidence data; Cloud analytics: Upload edge diagnostic results to the cloud and map the correlation between wireless charging system parameters and environmental data through a digital twin model; Distributed learning optimization: The diagnostic model is iteratively optimized in a distributed manner in the cloud, and the model parameters are updated. Health trend generation: Generate system health status data and long-term trend data based on the optimized cloud model.

[0011] Furthermore, the synthesis of multi-source diagnostic conclusions through evidence fusion algorithm in the two-level fusion decision module includes the following steps: The multimodal signals output from the edge diagnostic layer and the local decision layer are transformed into a unified feature vector; The feature vectors are weighted according to their source credibility. A basic probability assignment is constructed for each feature vector based on evidence theory; The Dempster-Shafer evidence fusion rule is used to combine and calculate the basic probability assignments of each feature vector to generate a comprehensive confidence score. Normalize or redistribute highly conflicting evidence that emerges during the fusion process; The multi-source diagnostic conclusion is output based on the integrated confidence level after fusion and then sent to the adaptive calibration module.

[0012] Furthermore, the adaptive control and sensor periodic calibration in the adaptive calibration module, which dynamically adjusts the resonant frequency, compensation parameters, or power, includes the following steps: S1. Receive multi-source diagnostic conclusions output by the two-level fusion decision module, including fault confidence and system state feature vector; S2. Analyze the deviation of power device, compensation network and coil parameters based on the diagnostic conclusions; S3. Determine the target values ​​and magnitudes for adjusting the resonant frequency, compensation parameters, or charging power; S4. Send commands to the power converter and compensation network through the controller to perform resonant frequency adjustment, compensation parameter switching or charging power adjustment; S5. Perform built-in benchmark or closed-loop self-test calibration on the sensors in the difference sensing acquisition module to correct the zero point and range parameters. S6. Use the calibrated sensor parameters or health scores as weight inputs for the two-level fusion mechanism; S7. Repeat S1 to S6 according to the preset cycle.

[0013] Secondly, the present invention provides the following technical solution: a method for wireless charging of electric vehicles, the method comprising the following steps: Construct a layered architecture that includes sensing and acquisition, data synchronization, local decision-making, edge diagnostics, cloud analysis, and control execution layers, with interfaces between each layer interacting. Based on the distribution of electromagnetic fields and thermosensitive regions of key components, magnetic field sensors, impedance spectrum probes, temperature and environmental sensors are arranged differently to collect multimodal signals; Multimodal signals are used for anti-interference processing, clock synchronization is used to unify timing, integrity is verified, and anomalies are reconstructed. Layered diagnosis: the local decision-making layer responds quickly to faults according to rules; the edge diagnosis layer integrates models and data-driven analysis to generate fault confidence; the cloud uses digital twins to map parameters and associate them with the environment, and optimizes the model through distributed learning to predict health trends. Multimodal signals are transformed into unified feature vectors, and multi-source diagnostic conclusions are synthesized through evidence fusion algorithms; Adaptive control and sensor periodic calibration are performed based on multi-source diagnostic conclusions, dynamically adjusting the resonant frequency, compensation parameters, or power; the sensors in the differential sensing acquisition module are calibrated through built-in benchmarks or closed-loop self-tests, and the calibration results are used as the two-level fusion weights.

[0014] The present invention has the following beneficial effects: 1. In this invention, by constructing a hierarchical architecture module, a differential sensing acquisition module, a signal synchronization processing module, a hierarchical collaborative diagnosis module, a two-level fusion decision module, and an adaptive calibration module, efficient acquisition, processing, and diagnosis of multimodal signals are achieved. This improves the problem that traditional electric vehicle wireless charging systems mostly use single-layer control and single sensing methods, resulting in untimely fault diagnosis due to information isolation and response lag.

[0015] 2. In this invention, by clearly defining the functions of the hierarchical architecture modules and implementing each module independently, combined with the synchronous processing of multimodal sensing signals and hierarchical collaborative diagnosis, real-time fault determination and health trend prediction of the wireless charging system can be achieved. This improves the problem that traditional systems mostly use centralized monitoring, which makes it difficult to grasp the system status in real time due to data transmission delay and limited processing capabilities.

[0016] 3. In this invention, a two-level fusion decision module uses an evidence fusion algorithm to generate multi-source diagnostic conclusions, and combines an adaptive calibration module to dynamically adjust the resonant frequency, compensation parameters, and charging power, thereby realizing sensor self-calibration and system adaptive control. This improves the problem that traditional wireless charging systems mostly use fixed parameters and manual calibration, and due to the lack of intelligent feedback mechanisms, charging efficiency and safety cannot be optimized in real time. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the architecture of a wireless charging system for electric vehicles proposed in this invention. Figure 2 This is a schematic flowchart of a wireless charging method for electric vehicles proposed in this invention. Detailed Implementation

[0018] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1: In a first embodiment of the present invention, the present invention provides a wireless charging system for electric vehicles, such as... Figure 1As shown, a layered architecture module is constructed to build a layered architecture including sensing and acquisition, data synchronization, local decision-making, edge diagnostics, cloud analysis and control execution layers, with interfaces between each layer interacting. Furthermore, the construction of a layered architecture module, which includes layers for perception and acquisition, data synchronization, local decision-making, edge diagnostics, cloud analysis, and control execution, includes the following steps: Define the functions of each layer, including the inputs and outputs of the perception and acquisition layer, data synchronization layer, local decision-making layer, edge diagnostic layer, cloud analysis layer, and control and execution layer; Each layer of functionality is implemented as an independent module, forming a data processing unit that can be called through an interface; Define the data format, transmission protocol, communication frequency, and synchronization method to enable data exchange between different layers; Unified timing management is implemented for data transmission between modules at each layer, using wired or wireless communication methods to ensure complete information transmission; Redundant hardware is used for the local decision-making layer and the edge diagnostic layer, and cross-layer data transmission is encrypted and authenticated; Integrate the modules at each layer to verify the implementation of data flow, interface, and timing synchronization.

[0020] Specifically, the layered architecture is constructed by defining the functions of the perception and acquisition layer, data synchronization layer, local decision-making layer, edge diagnostic layer, cloud analysis layer, and control execution layer. Each layer is independently modularized to form a data processing unit that can be called via an interface. First, the input and output signal sets for each layer are determined. With output results ,in Indicates the first The layer receives multimodal data or state parameters. The signal originates from a sensor, communication interface, or upper-layer module. This signal is acquired by the difference sensing acquisition module and undergoes preliminary filtering. For the first The output result after layer processing Indicates the first The layer processing parameter set includes thresholds, weight coefficients, and filtering coefficients; secondly, the functions of each layer are encapsulated into independent data processing units, each processing unit communicating through an interface. Communicate with adjacent layers to ensure input It is possible Mapped to lower layer ,Right now This indicates data format conversion, transmission protocol processing, and timing synchronization operations; furthermore, to ensure cross-layer data consistency, a unified clock is used. Timing management is performed on data transmission at each layer, and the data synchronization function is expressed as follows: Includes data alignment, buffering, and redundancy check operations, output As input to the lower-level processing; in addition, redundant hardware is introduced for the local decision layer and the edge diagnostic layer, and the redundant output is fused through a fusion function. Generate consistent results Weighted average or majority voting strategies can be used to output... For use by the edge diagnostics or control execution layer, while cross-layer communication data is encrypted via a function. Encryption and authentication are performed; finally, the modules are integrated, and the data flow between modules is verified through interfaces and time synchronization. Ensuring integrity and interactivity, and outputting unified multimodal processing results. The results can be used as input for edge diagnostics, cloud analysis and control execution layer, for subsequent multi-source diagnostic decision-making, fault prediction and adaptive control, to realize hierarchical collaborative processing and information transmission of the system.

[0021] By constructing a layered architecture and enabling independent operation and interface interaction of each layer module, unified transmission, timing synchronization, and redundancy processing of multimodal data can be achieved, providing a reliable data foundation for subsequent diagnostic analysis and control execution.

[0022] The differential sensing acquisition module is used to collect multimodal signals by differentially arranging magnetic field sensors, impedance spectrum probes, temperature and environmental sensors based on the distribution of electromagnetic fields and thermosensitive areas of key components. Furthermore, the acquisition of multimodal signals in the difference sensing acquisition module includes the following steps: Determine the placement and sampling sequence of the magnetic field sensor, impedance spectroscopy probe, temperature sensor, and environmental sensor; The magnetic field sensor, impedance spectrum probe, temperature sensor, and environmental sensor were initialized and calibrated to a reference standard, respectively. Use a unified trigger signal or a synchronous clock to simultaneously start sampling from each sensor; The analog signals collected by the sensor are converted into digital signals, and basic filtering and noise reduction processing is performed. The digital signals from various sensors are encoded in a unified format to generate a multimodal signal data structure, and timestamps and sensor identifiers are added. The multimodal signals are buffered at the acquisition node and prepared for transmission to the data synchronization layer or edge diagnostic layer according to the interface protocol.

[0023] Specifically, the difference sensing acquisition module obtains data by analyzing the electromagnetic field intensity distribution of key components. Temperature distribution in the thermosensitive region The analysis determined the placement of the magnetic field sensor, impedance spectroscopy probe, temperature sensor, and environmental sensor. With sampling order During sensor initialization, reference values ​​are set respectively. And calibrate the gain coefficient A unified trigger signal is used during the sampling process. Or synchronous clock Synchronously activate all sensors and collect analog signals. Converted into digital signals by analog-to-digital converter And apply the basic filter To achieve noise reduction processing; among which As the filter core, the processed digital signal is encoded according to a unified data structure. For timestamps, Multimodal signals generated for sensor identification Cached in the collection node Internally, according to the interface protocol The data is prepared for transmission to the data synchronization layer or edge diagnostic layer for subsequent signal synchronization processing, hierarchical diagnostics, and fusion decision-making. Input data includes real-time measurement signals from each sensor. Arrangement With sampling order The output is a multimodal signal data structure. This data structure provides a unified, synchronous, and traceable foundation for subsequent anti-interference processing, feature extraction, and fault diagnosis.

[0024] By strategically deploying multiple types of sensors and synchronously acquiring and encoding multimodal signals, comprehensive perception of the status of key components can be achieved, providing a unified and reliable data foundation for subsequent signal synchronization, feature extraction, and fault diagnosis.

[0025] The signal synchronization processing module is used for anti-interference processing of multi-mode signals, unifying timing through clock synchronization, verifying integrity, and reconstructing anomalies. Furthermore, the integrity verification and anomaly reconstruction in the signal synchronization processing module includes the following steps: The multimodal signals acquired by the difference sensing acquisition module are supplemented with cyclic redundancy check codes or checksums; At the signal receiving or processing node, integrity checks are performed on the multimodal signal to determine whether there are transmission or acquisition errors. For detected abnormal signals, interpolation or reconstruction processing is performed based on adjacent time steps or historical continuous sampling data; Update the reconstructed signal to the signal buffer or transmission queue and attach a status flag; Record the refactoring operations.

[0026] Specifically, the signal synchronization processing module processes the multimodal signals output by the difference sensing acquisition module. Interference suppression and clock synchronization management are implemented to achieve signal integrity verification and anomaly reconstruction; the input data consists of digital signal sequences from magnetic field sensors, impedance spectrum probes, temperature sensors, and environmental sensors. The sampling timestamp is The signal amplitude is The signal is obtained through a unified trigger signal or a synchronous clock; firstly, a cyclic redundancy check code is added to each signal. or checksum Used for integrity verification, the received signal is calculated at the receiving end. of or And judge Used to detect transmission or acquisition anomalies; if Then, interpolation or reconstruction algorithms are used to recover the abnormal signal. Linear interpolation can be expressed as: If continuous historical data is available, a sliding window average is used. Generate the reconstructed digital signal sequence ; then Update to the signal buffer or transmission queue and attach a status flag. This is used to identify whether the signal has been reconstructed and to record the operation log. Used for subsequent tracing and diagnostic analysis; the output is a unified time-series, multimodal signal that has undergone integrity verification and anomaly correction. The next step is to provide this signal to the hierarchical collaborative diagnostic module for fault detection and edge analysis, so as to achieve comprehensive perception of system status and health trend assessment.

[0027] By performing integrity verification, anomaly detection, and reconstruction processing on multimodal signals and achieving unified timing synchronization, reliable signal transmission and continuity are ensured, providing accurate data for subsequent fault diagnosis and health assessment.

[0028] The hierarchical collaborative diagnostic module is used for hierarchical diagnosis. The local decision-making layer responds quickly to faults according to rules; the edge diagnostic layer integrates models and data-driven analysis to generate fault confidence; and the cloud maps parameters with the environment using digital twins, optimizes the model through distributed learning, and predicts health trends. Furthermore, the hierarchical diagnosis in the hierarchical collaborative diagnosis module includes the following steps: Local decision-making layer processing: Receives multimodal signals output by the signal synchronization processing module, performs threshold judgment and state determination on the signals according to a preset rule table, and generates fault trigger signals; Edge diagnostic layer processing: Perform parameter drift analysis on multimodal signals, identify abnormal parameters using model-driven methods, and input the signals into a lightweight AutoEncoder model to generate anomaly scores; Edge layer fusion: The model-driven analysis results are fused with the data-driven anomaly score to form preliminary fault confidence data; Cloud analytics: Upload edge diagnostic results to the cloud and map the correlation between wireless charging system parameters and environmental data through a digital twin model; Distributed learning optimization: The diagnostic model is iteratively optimized in a distributed manner in the cloud, and the model parameters are updated. Health trend generation: Generate system health status data and long-term trend data based on the optimized cloud model.

[0029] Specifically, the hierarchical collaborative diagnostic module assesses system health status through the collaborative processing of the local decision-making layer, the edge diagnostic layer, and the cloud analysis layer. The local decision-making layer receives multimodal signals output by the signal synchronization processing module. Indicates the first Signals collected by each sensor are based on a preset rule table. Perform threshold judgment on the signal Generate a fault trigger signal: Edge diagnostic layer receives And calculate the parameter drift for each signal. It is the historical moving average, and at the same time... Input lightweight autoencoder model Generate reconstruction error Then shift the parameters With reconstruction error Fusion Forming a preliminary fault confidence vector The edge diagnostic results are uploaded to the cloud, where a digital twin model is used to map system parameters with environmental data. Model parameters are iteratively optimized through distributed learning algorithms. For environmental feature vectors, For loss function, For actual health labels, The learning rate is used to generate system health status data after iterative convergence. and long-term trend data Trend TrendAnalysis (Health) results are used to guide the adaptive calibration module in adjusting power device parameters, compensation network and coil parameters to maintain stable operation of the wireless charging system.

[0030] The hierarchical collaborative diagnostic module enables multi-level fault identification and health status assessment. It combines model and data-driven analysis to generate fault confidence and optimizes the system's long-term health trend through cloud optimization, thereby improving the continuous monitoring and adaptive management capabilities of the wireless charging system.

[0031] The two-level fusion decision module is used to convert multimodal signals into a unified feature vector and synthesize multi-source diagnostic conclusions through evidence fusion algorithms; Furthermore, the synthesis of multi-source diagnostic conclusions through evidence fusion algorithms in the two-level fusion decision module includes the following steps: The multimodal signals output from the edge diagnostic layer and the local decision layer are transformed into a unified feature vector; The feature vectors are weighted according to their source credibility. A basic probability assignment is constructed for each feature vector based on evidence theory; The Dempster-Shafer evidence fusion rule is used to combine and calculate the basic probability assignments of each feature vector to generate a comprehensive confidence score. Normalize or redistribute highly conflicting evidence that emerges during the fusion process; The multi-source diagnostic conclusion is output based on the integrated confidence level after fusion and then sent to the adaptive calibration module.

[0032] Specifically, each feature vector is composed of normalized measurements from each sensor. Composition; among which For the first The sample at the () The measured values ​​from the sensors are acquired by the difference sensing acquisition module and then processed by filtering, noise reduction, and standardization to obtain a unified feature vector. As input; for each By source credibility Weighting yields weighted eigenvectors Constructing a basic probability assignment based on evidence theory For each possible state Calculate trust quality; where and The Dempster-Shafer fusion rule was used to calculate the overall confidence level. Normalization processing is performed on highly conflicting evidence. Or conflict reassignment to generate the final fusion result. Output the overall confidence level after fusion. This leads to the adaptive calibration module, which guides the adjustment of resonant frequency, compensation parameters, or charging power, as well as sensor calibration. Each parameter, such as... and Based on sensor measurements and historical diagnostic data, the overall confidence level is... This indicates the confidence level of each fault or health status, which the adaptive calibration module will then determine based on. Perform system adjustments and periodic calibrations to maintain the stability and health status of the wireless charging system.

[0033] By transforming multimodal signals into a unified feature vector and using an evidence fusion algorithm to generate a comprehensive confidence score, collaborative diagnosis of multi-source information can be achieved, improving the accuracy of fault identification and providing a basis for adaptive calibration.

[0034] The adaptive calibration module is used to perform adaptive control and sensor periodic calibration based on multi-source diagnostic conclusions, and dynamically adjust the resonant frequency, compensation parameters or power; the sensors in the differential sensing acquisition module are calibrated through built-in benchmarks or closed-loop self-tests, and the calibration results are used as two-level fusion weights. Furthermore, the adaptive calibration module, including adaptive control and sensor periodic calibration, dynamically adjusts the resonant frequency, compensation parameters, or power, and includes the following steps: S1. Receive multi-source diagnostic conclusions output by the two-level fusion decision module, including fault confidence and system state feature vector; S2. Analyze the deviations in power device, compensation network, and coil parameters based on the diagnostic conclusions; S3. Determine the target values ​​and magnitudes for adjusting the resonant frequency, compensation parameters, or charging power; S4. Send commands to the power converter and compensation network through the controller to perform resonant frequency adjustment, compensation parameter switching or charging power adjustment; S5. Perform built-in benchmark or closed-loop self-test calibration on the sensors in the differential sensing acquisition module to correct the zero point and range parameters. S6. Use the calibrated sensor parameters or health scores as weight inputs for the two-level fusion mechanism; S7. Repeat S1 to S6 according to the preset cycle.

[0035] Specifically, the adaptive calibration module receives multi-source diagnostic conclusions from the two-level fusion decision module, including fault confidence. With system state feature vector For power device parameters Compensation network parameters With coil parameters Perform deviation analysis Calculate and adjust the resonant frequency based on the diagnostic conclusion. Adjustment of compensation parameters With charging power adjustment And issue commands through the controller. The power converter and compensation network are dynamically adjusted, while the sensors in the differential sensing module undergo built-in benchmark or closed-loop self-calibration to obtain calibrated parameters. And use it as the weight of the two-level fusion mechanism. Input is used to update the diagnostic fusion results, and the module operates on a preset cycle. Repeatedly execute S1 to S6 above to achieve continuous adaptive control and sensor calibration. The process executes S1 to S6 once per cycle, and the results are used to dynamically optimize the resonant state of the wireless charging system, compensate for network matching and power output, improve system stability, and provide reliable input data for subsequent health trend prediction and fault protection.

[0036] The adaptive calibration module enables dynamic adjustment of the resonant frequency, compensation parameters, and power of the wireless charging system. It also continuously optimizes the system's operating status through periodic sensor calibration, improving operational stability and supporting health trend prediction and fault protection.

[0037] Example 2: Public charging stations for electric vehicles suffer from unstable charging efficiency, abnormal equipment temperature rise, and difficulty in timely detection of potential faults during high-power wireless charging. To address these issues, this invention provides a wireless charging method for electric vehicles, the process of which is as follows: Figure 2 As shown. The specific implementation process of this method is as follows: First, a layered architecture is constructed, including sensing and acquisition, data synchronization, local decision-making, edge diagnosis, cloud analysis, and control execution layers. The interfaces of each layer interact, enabling each layer to operate independently and exchange data, thereby ensuring the integrity of information transmission and time synchronization. Secondly, based on the distribution of electromagnetic fields and thermosensitive areas of key components, magnetic field sensors, impedance spectrum probes, temperature and environmental sensors are arranged differently to collect multi-mode signals, which can realize multi-dimensional monitoring of key parameters of the charging system and improve data coverage. Subsequently, multi-modal signals are used for anti-interference processing. The timing is unified by clock synchronization, integrity is verified and anomalies are reconstructed, so that the collected data is accurate and reliable, and the risk of signal error and loss is reduced. Next, layered diagnosis is implemented. The local decision-making layer responds quickly to faults according to rules, the edge diagnosis layer integrates models and data-driven analysis to generate fault confidence, and the cloud uses digital twin mapping parameters to associate with the environment and optimizes the model through distributed learning to predict health trends. This achieves early identification of potential faults and prediction of health status, enhancing system maintainability. Then, the multimodal signals are transformed into a unified feature vector, and multi-source diagnostic conclusions are synthesized through an evidence fusion algorithm to generate reliable system comprehensive diagnostic results to guide subsequent adjustments; Finally, based on the multi-source diagnostic results, adaptive control and sensor periodic calibration are performed to dynamically adjust the resonant frequency, compensation parameters or power. At the same time, the sensors in the differential sensing acquisition module are calibrated through built-in benchmarks or closed-loop self-tests. The calibration results serve as the weights for the two-level fusion, which can continuously optimize charging efficiency, reduce energy loss and temperature rise risks, and ensure the stable and safe operation of the charging system.

[0038] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A wireless charging system for electric vehicles, characterized in that, include: A layered architecture module is constructed to build a layered architecture that includes sensing and acquisition, data synchronization, local decision-making, edge diagnostics, cloud analysis, and control execution layers, with interfaces between each layer interacting. The differential sensing acquisition module is used to collect multimodal signals by differentially arranging magnetic field sensors, impedance spectrum probes, temperature and environmental sensors based on the distribution of electromagnetic fields and thermosensitive areas of key components. The signal synchronization processing module is used for anti-interference processing of multi-mode signals, unifying timing through clock synchronization, verifying integrity, and reconstructing anomalies. The hierarchical collaborative diagnostic module is used for hierarchical diagnosis. The local decision-making layer responds quickly to faults according to rules; the edge diagnostic layer integrates models and data-driven analysis to generate fault confidence; and the cloud maps parameters with the environment using digital twins, optimizes the model through distributed learning, and predicts health trends. The two-level fusion decision module is used to convert multimodal signals into a unified feature vector and synthesize multi-source diagnostic conclusions through evidence fusion algorithms; The adaptive calibration module is used to perform adaptive control and sensor periodic calibration based on multi-source diagnostic conclusions, and dynamically adjust the resonant frequency, compensation parameters or power; the sensors in the differential sensing acquisition module are calibrated through built-in benchmarks or closed-loop self-tests, and the calibration results are used as the two-level fusion weights.

2. The wireless charging system for electric vehicles according to claim 1, characterized in that, The construction of the layered architecture module, which includes the sensing and acquisition, data synchronization, local decision-making, edge diagnostics, cloud analysis, and control execution layers, includes the following steps: Define the functions of each layer, including the inputs and outputs of the perception and acquisition layer, data synchronization layer, local decision-making layer, edge diagnostic layer, cloud analysis layer, and control and execution layer; Each layer of functionality is implemented as an independent module, forming a data processing unit that can be called through an interface; Define the data format, transmission protocol, communication frequency, and synchronization method to enable data exchange between different layers; Unified timing management is implemented for data transmission between modules at each layer, using wired or wireless communication methods to ensure complete information transmission; Redundant hardware is used for the local decision-making layer and the edge diagnostic layer, and cross-layer data transmission is encrypted and authenticated; Integrate the modules at each layer to verify the implementation of data flow, interface, and timing synchronization.

3. The wireless charging system for electric vehicles according to claim 1, characterized in that, The acquisition of multimodal signals in the differential sensing acquisition module includes the following steps: Determine the placement and sampling sequence of the magnetic field sensor, impedance spectroscopy probe, temperature sensor, and environmental sensor; The magnetic field sensor, impedance spectrum probe, temperature sensor, and environmental sensor were initialized and calibrated to a reference standard, respectively. Use a unified trigger signal or a synchronous clock to simultaneously start sampling from each sensor; The analog signals collected by the sensor are converted into digital signals, and basic filtering and noise reduction processing is performed. The digital signals from various sensors are encoded in a unified format to generate a multimodal signal data structure, and timestamps and sensor identifiers are added. The multimodal signals are buffered at the acquisition node and prepared for transmission to the data synchronization layer or edge diagnostic layer according to the interface protocol.

4. The wireless charging system for electric vehicles according to claim 1, characterized in that, The integrity verification and anomaly reconstruction in the signal synchronization processing module includes the following steps: The multimodal signals acquired by the difference sensing acquisition module are supplemented with cyclic redundancy check codes or checksums; At the signal receiving or processing node, integrity checks are performed on the multimodal signal to determine whether there are transmission or acquisition errors. For detected abnormal signals, interpolation or reconstruction processing is performed based on adjacent time steps or historical continuous sampling data; Update the reconstructed signal to the signal buffer or transmission queue and attach a status flag; Record the refactoring operations.

5. The wireless charging system for electric vehicles according to claim 1, characterized in that, The hierarchical diagnosis in the hierarchical collaborative diagnosis module includes the following steps: Local decision-making layer processing: Receives multimodal signals output by the signal synchronization processing module, performs threshold judgment and state determination on the signals according to a preset rule table, and generates fault trigger signals; Edge diagnostic layer processing: Perform parameter drift analysis on multimodal signals, identify abnormal parameters using model-driven methods, and input the signals into a lightweight AutoEncoder model to generate anomaly scores; Edge layer fusion: The model-driven analysis results are fused with the data-driven anomaly score to form preliminary fault confidence data; Cloud analytics: Upload edge diagnostic results to the cloud and map the correlation between wireless charging system parameters and environmental data through a digital twin model; Distributed learning optimization: The diagnostic model is iteratively optimized in a distributed manner in the cloud, and the model parameters are updated. Health trend generation: Generate system health status data and long-term trend data based on the optimized cloud model.

6. The wireless charging system for electric vehicles according to claim 1, characterized in that, The synthesis of multi-source diagnostic conclusions through evidence fusion algorithm in the two-level fusion decision module includes the following steps: The multimodal signals output from the edge diagnostic layer and the local decision layer are transformed into a unified feature vector; The feature vectors are weighted according to their source credibility. A basic probability assignment is constructed for each feature vector based on evidence theory; The Dempster-Shafer evidence fusion rule is used to combine and calculate the basic probability assignments of each feature vector to generate a comprehensive confidence score. Normalize or redistribute highly conflicting evidence that emerges during the fusion process; The multi-source diagnostic conclusion is output based on the integrated confidence level after fusion and then sent to the adaptive calibration module.

7. The wireless charging system for electric vehicles according to claim 1, characterized in that, The adaptive calibration module, including adaptive control and sensor periodic calibration, dynamically adjusts the resonant frequency, compensation parameters, or power, comprising the following steps: S1. Receive multi-source diagnostic conclusions output by the two-level fusion decision module, including fault confidence and system state feature vector; S2. Analyze the deviation of power device, compensation network and coil parameters based on the diagnostic conclusions; S3. Determine the target values ​​and magnitudes for adjusting the resonant frequency, compensation parameters, or charging power; S4. Send commands to the power converter and compensation network through the controller to perform resonant frequency adjustment, compensation parameter switching or charging power adjustment; S5. Perform built-in benchmark or closed-loop self-test calibration on the sensors in the difference sensing acquisition module to correct the zero point and range parameters. S6. Use the calibrated sensor parameters or health scores as weight inputs for the two-level fusion mechanism; S7. Repeat S1 to S6 according to the preset cycle.

8. A method for wireless charging an electric vehicle, characterized in that, For a wireless charging system for an electric vehicle according to any one of claims 1-7, the method includes the following steps: Construct a layered architecture that includes sensing and acquisition, data synchronization, local decision-making, edge diagnostics, cloud analysis, and control execution layers, with interfaces between each layer interacting. Based on the distribution of electromagnetic fields and thermosensitive regions of key components, magnetic field sensors, impedance spectrum probes, temperature and environmental sensors are arranged differently to collect multimodal signals; Multimodal signals are used for anti-interference processing, clock synchronization is used to unify timing, integrity is verified, and anomalies are reconstructed. Layered diagnosis: the local decision-making layer responds quickly to faults according to rules; the edge diagnosis layer integrates models and data-driven analysis to generate fault confidence; the cloud uses digital twins to map parameters and associate them with the environment, and optimizes the model through distributed learning to predict health trends. Multimodal signals are transformed into unified feature vectors, and multi-source diagnostic conclusions are synthesized through evidence fusion algorithms; Adaptive control and sensor periodic calibration are performed based on multi-source diagnostic conclusions, dynamically adjusting the resonant frequency, compensation parameters, or power; the sensors in the differential sensing acquisition module are calibrated through built-in benchmarks or closed-loop self-tests, and the calibration results are used as the two-level fusion weights.

Images