Intelligent wireless charging control system and method for electric vehicle

Abstract

The invention provides an electric vehicle intelligent wireless charging control system, which comprises a central control module, a wireless charging module, a charger module, a sensor module and a big data module, and is characterized in that the central control module is used for acquiring state data related to a battery pack, analyzing the state data and outputting an execution signal; the wireless charging module receives an execution signal of the central control module, executes and deploys corresponding charging parameters and charges a battery, the charger module has an information acquisition function and is used for carrying out information interaction with the central control module, and the charger module is used for acquiring a current state and transmitting the current state to the central control module; the sensor module is used for collecting battery pack state data, and the big data module is used for information interaction with the central control module. According to the invention, coordination processing can be carried out in multiple charging environments.

Description

Technical Field

[0001] The present invention relates to the fields of multi-charging coordination and artificial intelligence, and in particular to an intelligent wireless charging control system and method for an electric vehicle. Background Art

[0002] With the widespread popularity of electric bicycles, related charging technologies are also constantly improving. However, the traditional wired charging method is still the mainstream in the current market. Although this method is mature, it also exposes a series of problems. First, the frequent plugging and unplugging of the cable causes it to be easily worn, which not only shortens the service life of the equipment, but may also cause failures in long-term use. Secondly, poor contact is a common phenomenon in wired charging. This situation not only affects the charging efficiency, but also may cause incomplete battery charging, thereby affecting the endurance of the electric bicycle. In addition, wired charging also has safety hazards. For example, when used in a humid environment, short circuits or electric shock accidents may occur, posing potential risks to users.

[0003] At present, the research on wireless charging technology is mainly concentrated in the basic technical fields such as magnetic coupling resonance and zero voltage switching. Traditional control methods usually adopt fixed parameter PID control or simple open-loop control, lacking intelligent management of the charging process. In terms of communication, a single short-distance communication protocol is mostly used, and the reliability and real-time performance of data transmission are relatively poor. The optimization of charging parameters mainly depends on the preset charging curve, which is difficult to adapt to the changes of different battery types and usage environments. In terms of multi-charger collaborative management, the existing technologies mostly adopt simple time division multiplexing or fixed priority strategies, lacking flexible resource scheduling mechanisms. However, there are still many specific problems in the current wireless charging control technology: First, in the data acquisition and preprocessing links, the synchronization accuracy of sensor data is generally at the millisecond level, which is difficult to meet the high-precision control requirements; second, the existing noise suppression algorithm has poor adaptability to electromagnetic interference, and the filtering effect is unstable in complex electromagnetic environments; third, the optimization of charging parameters lacks consideration of the network topology, resulting in inefficient resource allocation; fourth, the state prediction model is not adaptable enough to the charging characteristics of different types of batteries, and the prediction accuracy is difficult to guarantee; finally, the collaborative control of multiple chargers does not fully consider the influence of physical space distance, affecting the real-time and scalability of the system. These technical problems seriously restrict the large-scale application and intelligent upgrading of wireless charging systems. Summary of the invention

[0004] The technical problem to be solved by the present invention is to overcome the defects of the prior art. The present invention proposes an intelligent wireless charging control system and method for electric vehicles. The system can adapt to the complex and changeable charging environment and realize the intelligent management of the charging process. From data collection to the final execution control, each link adopts a targeted optimization algorithm to form a closed-loop control system, which effectively realizes the coordinated control of multiple charging.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is: an electric vehicle intelligent wireless charging control system, comprising:

[0006] A central control module, the central control module is used to obtain status data related to the battery pack, analyze the status data, and output an execution signal; the central control module also stores the status data;

[0007] The wireless charging module receives and executes the execution signal of the central control module, adjusts the corresponding charging parameters, and charges the battery;

[0008] A charger module is adapted to the battery pack, and the battery is charged by wirelessly connecting the charger module to the wireless charging module; the charger module has an information collection function, and is used to exchange information with the central control module; the charger module is used to obtain status data related to the battery pack in the current state, and transmit it to the central control module;

[0009] A sensor module, used to collect battery pack status data and related environmental data;

[0010] The big data module is used to exchange information with the central control module.

[0011] 2. An electric vehicle intelligent wireless charging control system as claimed in claim 1, characterized in that:

[0012] The central control module includes a central control processing module, a central data storage module, a 4G module, and an API interface;

[0013] The wireless charging module includes a ground unit status monitoring module, a ground unit control processing module, a charging module, and a wireless communication module.

[0014] The charger module includes a battery status monitoring module, a battery data processing module, and a connector module;

[0015] The sensor module includes a voltage sensor, a current sensor, a temperature sensor array, an electromagnetic field sensor, and a position sensor.

[0016] Further, the following steps are included:

[0017] S1. Acquire real-time status data of the wireless charging control system, perform noise reduction processing on it, and obtain microsecond-level status data;

[0018] S2. Intelligently optimize charging parameters based on the processed status data;

[0019] S3, pre-build hybrid neural network and multi-objective optimization algorithm model, predict charging status and control the generation of charging strategy based on optimized charging parameters;

[0020] S4. Based on the charging strategy obtained in step S3, dynamic balancing of charging load is achieved through periodic charging load distribution, thereby achieving coordinated control of multiple chargers.

[0021] Furthermore, step S1 further comprises:

[0022] S11, collect battery voltage data V, battery current data I, coil temperature data Tc, ambient temperature data Te, electromagnetic field strength data Em, alignment deviation data D and communication quality data Q through sensors; synchronize the sampling time to generate a time series data matrix X; perform time stamp alignment on the time series data matrix X and output a synchronized data set X';

[0023] S12, read the synchronous data set X', calculate the dynamic noise covariance matrix R(t) of the data,

[0024] The function formula is, R(t) = R0 + ΔR·exp(-λ||e(t)||²),

[0025] Among them, R0 is the basic covariance matrix, e(t) is the estimation error, λ is the adaptive coefficient, and ΔR is the dynamic adjustment matrix; based on the obtained dynamic noise covariance matrix, the synchronous data set X' is adaptively Kalman filtered to output the filtered data set Y;

[0026] S13, receiving the data set Y, using the nonlinear mapping function to transform the features, the function formula is:

[0027] φ(x)=tanh(wx+b)·exp(-αx²), where w is the weight coefficient, b is the bias term, and α is the attenuation coefficient. Based on the transformed data, the feature correlation matrix is ​​calculated to obtain the feature correlation matrix C. The feature contribution is calculated based on the feature correlation matrix C, and the features whose contribution is greater than the preset threshold are selected to form the fused feature data F.

[0028] Furthermore, step S2 further comprises:

[0029] S21, receiving fused feature data F, constructing an edge weight matrix w(e)=β1·P(e)+β2·D(e)+β3·Q(e) according to power transmission efficiency, transmission distance and link quality, where P(e) is power transmission efficiency, D(e) is transmission distance, Q(e) is link quality, β1, β2, β3 are dynamic weight coefficients; establishing a network topology map based on the edge weight matrix, and outputting a network topology matrix G;

[0030] S22. Read the network topology matrix G and the pre-stored historical state data H, and use the state transition probability update formula P(s'|s,a) = P0(s'|s,a)·exp(γ·ΔV(s,s')) to calculate the new transition probability, where P0 is the initial transition probability, γ is the learning rate, and ΔV is the state value function difference; generate the state transfer matrix P according to the updated state transition probability; optimize the power, frequency and phase parameters based on the state transfer matrix P, and output the optimized parameter set K.

[0031] Furthermore, step S3 further comprises:

[0032] S31, receiving the optimized parameter set K and the real-time state data R, respectively inputting them into the LSTM network, the GRU network and the attention network to obtain three sets of prediction results; using the prediction function y(t+1) = f(LSTM(x(t))) +g(GRU(x(t)))+h(Attention(x(t))) to fuse the three sets of prediction results, where f, g and h are weight functions; outputting the fused state prediction value Y;

[0033] S32. Read the state prediction value Y and the preset target parameter T, and construct a multi-objective optimization function J = w1·E +w2·T+w3·S, where E is energy consumption, T is charging time, S is the safety risk index, and w1, w2, and w3 are dynamic weights; calculate the optimization target matrix J; generate a hierarchical charging strategy based on the optimization target matrix J, and output the control strategy set S.

[0034] Furthermore, step S4 further comprises:

[0035] S41. Receive the control strategy set S, perform collaborative calculation based on the current state update rule xᵢ(k+1)=xᵢ(k)+∑aᵢⱼ(xⱼ(k)- xᵢ(k))·exp(-μ·d(i, j)), where aᵢⱼ is the adjacency matrix element, d(i, j) is the distance between nodes, and μ is the attenuation coefficient; generate the collaborative state matrix C;

[0036] Step S42, read the collaborative state matrix C, and use the load distribution function L(i)=L0(i)+ΔL·sin(ωt +φᵢ) to calculate the load value of each charger, where L0(i) is the basic load, ΔL is the adjustment amplitude, ω is the cycle parameter, and φᵢ is the phase difference; generate a load distribution vector L; formulate an execution instruction according to the load distribution vector L, and output an execution instruction set E.

[0037] Compared with the prior art, the beneficial effects of the present invention include: building a complete intelligent wireless charging control system for electric vehicles. The system can adapt to complex and changeable charging environments, realize intelligent management of the charging process, and adopt targeted optimization algorithms in each link from data collection to final execution control, forming a closed-loop control system, effectively realizing coordinated control of multiple charging. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 , is a hub diagram of the intelligent wireless charging control system for electric vehicles in Example 1 of the present invention.

[0039] Figure 2 , is a flow chart of step S1 in the electric vehicle intelligent wireless charging control method of the present invention.

[0040] Figure 3 , is a flow chart of step S2 in the electric vehicle intelligent wireless charging control method of the present invention.

[0041] Figure 4 , is a flow chart of step S3 in the electric vehicle intelligent wireless charging control method of the present invention.

[0042] Figure 5 , is a flow chart of step S4 in the electric vehicle intelligent wireless charging control method of the present invention. DETAILED DESCRIPTION

[0043] It is easy to understand that according to the technical solution of the present invention, without changing the essential spirit of the present invention, a person skilled in the art can propose a variety of interchangeable structural modes and implementation modes. Therefore, the following specific implementation modes and drawings are only exemplary descriptions of the technical solution of the present invention, and should not be regarded as the whole of the present invention or as a limitation or restriction to the technical solution of the present invention.

[0044] Example 1

[0045] Based on the problems mentioned in the background technology, the present application proposes an intelligent wireless charging control system for electric vehicles, including a central control module, a wireless charging module, a charger module, a sensor module, and a big data module. The central control module is used to obtain status data related to the battery pack, analyze the status data, and output an execution signal. The central control module also stores the status data. The wireless charging module receives the execution signal of the central control module and executes it, adjusts the corresponding charging parameters, and charges the battery. The charger module is adapted to the battery pack and is wirelessly connected to the wireless charging module through the charger module to charge the battery. The charger module has an information collection function and is used to exchange information with the central control module. The charger module is used to obtain status data related to the battery pack in the current state and transmit it to the central control module. The sensor module is used to collect battery pack status data and related environmental data. The big data module is used to exchange information with the central control module.

[0046] In this solution, the central control module includes a central control processing module, a central data storage module, a 4G module, and an API interface; the wireless charging module includes a ground unit status monitoring module, a ground unit control processing module, a charging module, and a wireless communication module; the charger module includes a battery status monitoring module, a battery data processing module, and a connector module; the sensor module includes a voltage sensor, a current sensor, a temperature sensor array, an electromagnetic field sensor, and a position sensor.

[0047] In this solution, the wireless charging module can be specifically assigned to a ground wireless charging unit, and the number can be arranged as required to match the charger module. A charging area is set up near the ground wireless charging unit. When the charger module is located in the charging area, the charger module is first charged and turned on to obtain the status data of the battery pack and the environmental status data sensed by the sensor module. The charger module processes the data and sends it to the central control module for data analysis, adjusting the charging strategy, deploying the wireless charging module, and charging the battery pack through the charger module. At the same time, the data is stored and uploaded to the big data module.

[0048] In order to better illustrate the present solution, this embodiment is described in combination with specific components, and its working process is as follows:

[0049] like Figure 1 As shown in the figure, it is a schematic diagram of the electric vehicle wireless charging station hub of this embodiment. The charging station includes several electric vehicle wireless charging systems, each of which is composed of at least a vehicle charger, a ground wireless charging unit and a central controller. The charger and the ground wireless charging unit exchange information through LoRa, and the ground wireless charging unit and the central controller exchange information through the CAN bus. The central controller is provided with an API interface to facilitate user control and query.

[0050] The charger module can be specifically an on-board charger, which at least includes a battery status monitoring module, a battery data processing module, a charger wireless communication (LoRa) module, and a sensor module. The battery status monitoring module is used to monitor the battery charging process and battery status. The collected data is calculated by the battery data processing module and then transmitted to the ground wireless charging unit by the wireless communication module. The sensor module is used to detect current data, such as pressure, temperature, etc.

[0051] The wireless charging module can be specifically a ground wireless charging unit, which at least includes a wireless charging status monitoring module, a wireless charging control processing module and its device, a CAN bus and a wireless communication (LoRa) module. The wireless charging status monitoring module is used to monitor the working status of the system in all directions, the wireless charging control processing module and its device are used to process system data and respond to various instructions issued by the central controller, the CAN bus is used to exchange information with the central controller, and the wireless communication module is used to exchange information with the vehicle charger.

[0052] The specific location of the central control module is the central controller, which at least includes the CAN bus, central control processing module, data storage module and 4G module. The CAN bus is used to exchange information with the ground wireless charging unit; the control processing module is used to process all data in the charging station and the instructions issued by the cloud platform; the data storage module is used to store all operating data in the charging station; and the 4G module is used to exchange information with the big data cloud platform.

[0053] This is a schematic diagram of the normal operation process of the electric vehicle wireless charging system. The normal operation process of the system is as follows:

[0054] This is the idle stage of the ground wireless charging unit. When idle, the ground wireless charging unit will be in a state of waiting for the vehicle charger to enter. In this state, the ground wireless charging unit will cyclically determine whether there is a vehicle charger in place and respond to the instructions issued by the central controller.

[0055] The ground wireless charging unit requests to start up: When the vehicle charger is in place, the specific device of the ground wireless charging unit detection module is triggered. The ground wireless charging unit confirms that the vehicle charger is in place and requests to start up from the central controller.

[0056] Micro-power startup phase: After making a judgment, the central controller issues a startup command to the ground wireless charging unit. After receiving the startup command, the ground wireless charging unit executes the system micro-power startup and ensures constant output power.

[0057] Handshake phase: After the ground wireless charging unit is turned on, the on-board charger obtains power and starts up. After battery status detection and data processing, it performs LoRa handshake with the ground wireless charging unit and sends a charging request to the ground wireless charging unit.

[0058] Parameter adjustment and power supply stage: After receiving the charging request, the ground wireless charging unit enters the parameter adjustment and power supply stage. In this stage, the ground wireless charging unit adjusts the system operating parameters to make the system enter the normal charging stage. After this stage is completed, the ground wireless charging unit reports the start charging data to the central controller.

[0059] Normal power supply stage: The system enters the normal power supply stage until charging is completed.

[0060] The above process is the normal operation process of the system, which contains multiple interrupt programs to deal with system failures. When a system failure occurs, a fault signal is output through the corresponding hardware circuit. When the system receives the fault signal, the system power is immediately turned off through the interrupt program.

[0061] This is the control flow chart of the ground wireless charging unit in idle state. The specific control process is as follows:

[0062] S101: System initialization, including resetting system-related flags.

[0063] S102: Read the output of the vehicle charger in-place detection module. When the vehicle charger is in place, the detection module will output a corresponding signal, such as using a pressure sensor or a Hall switch.

[0064] S103: Determine whether the onboard charger is in place. If so, the ground wireless charging unit uploads a power-on request to the central controller and terminates the current control module. If the onboard charger is not in place, execute the next step.

[0065] S104: Determine whether the instruction sent by the central controller is received. If the instruction is received, execute the relevant instruction, and upload the execution result to the central controller through the CAN bus after the execution is completed. Then continue to read the output of the vehicle charger in-place detection module. If the instruction sent by the central controller is not received, directly continue to read the output of the vehicle charger in-place detection module.

[0066] If the central controller does not issue a power-on command, only the microcontroller and detection module of the charging unit with extremely low power consumption will work at this stage, avoiding the ineffective power loss of the charging unit and the safety problem of space electromagnetic radiation generated by the transmitting coil when there is no on-board charger.

[0067] like Figure 4As shown in FIG. 1 , it is a control flow chart of the ground wireless charging unit in the power-on request phase. The specific control process is as follows:

[0068] S201: Upload power-on request: When the on-board charger is reliably in place, the ground wireless charging unit uploads the power-on request to the central controller. The central controller queues the power-on requests of multiple charging units at the same time and issues power-on commands one by one to ensure that only one device in the current system is in the power-on stage, avoiding the bus current shock problem caused by the simultaneous startup of multiple devices, which helps to further reduce the overall system loss and extend the life of the device.

[0069] S202: Confirming flags: Before the system receives a power-on command, check all system flags to confirm that the system-related flags are correct.

[0070] S203: Query command: Check whether there is a power-on command issued by the central processor on the CAN bus.

[0071] S204: Determination: If a power-on instruction is obtained, the system-related flag bits are changed to prepare for the system power-on; otherwise, the correctness of the system flag bits is continuously checked.

[0072] like Figure 5 As shown in FIG. 1 , it is a control flow chart of the ground wireless charging unit in the micro-power startup stage. The specific control process is as follows:

[0073] S301: Non-resonant start: The inverter starts PWM at a non-resonant frequency, so that the system works in a non-resonant state. At this time, the output power of the system can be calculated by the system parameters to be a very small value. It is only necessary to ensure that the on-board charger microcontroller works normally and sends a response signal.

[0074] S302: Determine the number of cycles. If the number of cycles is greater than 3, the system micro-power startup is determined to have failed. The ground wireless charging unit uploads the startup failure information to the central controller via the CAN bus and resets the relevant flags. If the number of cycles is less than 3, execute the next step.

[0075] S303: Delay for a fixed time to ensure that the on-board charger obtains enough power to start self-test and the LoRa module.

[0076] S304: The ground wireless charging unit reads the vehicle charger LoRa information.

[0077] S305: Determine whether the vehicle charger has a response signal. If no response signal is received, the number of cycles is increased by 1 and the process returns to step S302; otherwise, the process proceeds to the next step.

[0078] S306: Perform secondary confirmation on the received response signal to ensure that the connected on-board charger meets the standards.

[0079] S307: Determine whether the response of the on-board charger is normal. If the response is abnormal, upload the startup failure information to the central controller via the CAN bus and reset the relevant flag bit; otherwise, execute the next step.

[0080] S308: Upload the vehicle charger access success information to the central controller via the CAN bus.

[0081] The ground wireless charging unit and the on-board charger perform information interaction in the LoRa handshake phase. The specific control process is as follows:

[0082] S401: After the on-board charger receives power and starts, it adjusts its own LoRa module to broadcast mode and cyclically sends a request handshake signal "a".

[0083] S402: After receiving the handshake request signal, the ground wireless charging unit sends an approval handshake signal "b" and the communication address "c" of the ground wireless charging unit LoRa to it.

[0084] S403: After receiving the information "c" including the consent handshake signal "b" and the communication address, the on-board charger sets its own communication address to "c" and starts self-checking the battery status of the electric vehicle.

[0085] S404: After the on-board charger completes the self-check of the battery status, it generates information “d” and transmits it to the ground wireless charging unit via LoRa.

[0086] S405: After receiving the information "d", the ground wireless charging unit determines whether to enter the normal charging stage.

[0087] In order to avoid interference in the communication between the on-board charger and the ground wireless charging unit in stages S401 and S402, the power-on command can be controlled by the central controller in step S3 to ensure that there is only one ground wireless charging unit in the charging station that is in the micro-power startup stage.

[0088] The control process of the ground wireless charging unit in the parameter adjustment and power supply stage is as follows:

[0089] S501: Set the step value of the system parameter (such as frequency, phase, etc.) and perform the first step.

[0090] S502: Detect the working state of the ground wireless charging unit and obtain working state parameters of the ground wireless charging unit.

[0091] S503: Determine whether the working state of the ground wireless charging unit is normal. If the parameters are abnormal, upload the current abnormal parameters to the central controller via the CAN bus, and reset the relevant flags to shut down the system. Otherwise, proceed to the next step.

[0092] S504: Send a query command to the vehicle charger via LoRa to query the battery status of the electric vehicle and read the battery status parameters.

[0093] S505: Determine whether the battery status of the electric vehicle is normal. If the status is abnormal, upload the current abnormal parameters to the central controller via the CAN bus, and reset the relevant flags to shut down the system. Otherwise, proceed to the next step.

[0094] S506: Execute parameter stepping.

[0095] S507: Determine whether the parameters (frequency, phase, etc.) have reached the resonance condition. If there is still no resonance, return to S502. Otherwise, execute the next step.

[0096] S508: Upload the power supply to the central controller through the CAN bus to complete the parameter adjustment and upload related parameters.

[0097] The control process of the ground wireless charging unit in the normal charging stage is as follows:

[0098] S601: Start the counter. When the counter reaches a threshold, the ground wireless charging unit reports working data to the central controller.

[0099] S602: Detect the working state of the ground wireless charging unit and obtain working state parameters of the ground wireless charging unit.

[0100] S603: Determine whether the working state of the ground wireless charging unit is normal. If the parameters are abnormal, upload the current abnormal parameters to the central controller via the CAN bus, and reset the relevant flags to shut down the system. Otherwise, proceed to the next step.

[0101] S604: Determine whether an instruction issued by the central controller is received, if an instruction is received, execute the relevant instruction and report the execution result to the central controller. Otherwise, execute the next step.

[0102] S605: Determine whether the battery status of the electric vehicle is normal. If the status is abnormal, upload the current abnormal parameters to the central controller via the CAN bus, and reset the relevant flags to shut down the system. Otherwise, proceed to the next step.

[0103] S606: Determine whether the counter reaches the threshold. If the counter reaches the threshold, the ground wireless charging unit reports the working data to the central controller via the CAN bus, clears the counter and executes the next step. Otherwise, execute the next step.

[0104] S607: Determine whether to continue charging, if necessary, return to S602. Otherwise, execute the next step.

[0105] S608: Charging is completed. The ground wireless charging unit reports the charging completion to the central controller via the CAN bus and shuts down the system.

[0106] Example 2

[0107] In Example 1, based on the intelligent wireless charging control system for electric vehicles, this application also proposes a wireless charging control method, which realizes intelligent coordination of multiple chargers during charging through a core algorithm. The main scheme is as follows:

[0108] Step 1: Acquire the real-time status data of the wireless charging control system, perform noise reduction processing on it, and obtain microsecond-level status data;

[0109] Step 2: Based on the processed status data, intelligently optimize the charging parameters;

[0110] Step 3: pre-build a hybrid neural network and multi-objective optimization algorithm model to predict the charging state and control the generation of charging strategy based on the optimized charging parameters;

[0111] Step 4: Based on the charging strategy, the dynamic balancing of the charging load is achieved through periodic charging load distribution, so as to achieve coordinated control of multiple chargers.

[0112] In view of the above steps, this application combines innovative technologies such as multi-dimensional data acquisition, adaptive preprocessing, topology optimization, hybrid intelligent prediction and distributed collaborative control to build a complete intelligent wireless charging control solution for electric vehicles. This solution can adapt to complex and changeable charging environments and realize intelligent management of the charging process. From data acquisition to final execution control, each link adopts targeted optimization algorithms to form a closed-loop control system.

[0113] Based on the above technical solution, this application provides a specific implementation scheme:

[0114] Step 1: Through multi-dimensional data collection and adaptive preprocessing mechanism, high-precision data quality assurance is achieved, as follows:

[0115] S11, respectively obtain raw data from the voltage sensor, current sensor, temperature sensor array, electromagnetic field sensor, position sensor and communication module; perform range calibration on each sensor data to generate a calibration data set D1; perform linear correction on the calibration data set D1 according to the sensor characteristic curve, and output a corrected data set D2;

[0116] S12, read the corrected data set D2, segment the data using the sliding window method, and obtain the segmented data set D3; calculate the statistical characteristics of each segmented data, including mean, variance, skewness and kurtosis, and generate a feature data set D4; perform outlier detection based on the feature data set D4, and output the filtered data set D5.

[0117] S13, receiving the filtered data set D5, extracting the timestamp information of each data, and establishing a time index table T; interpolating and aligning the data according to the time index table T to obtain an aligned data set D6; calculating the sampling consistency index of the data, and generating a time series data quality matrix Q;

[0118] S14, read the aligned data set D6 and the time series data quality matrix Q, weight the data according to the indicators in the quality matrix, and obtain the weighted data set D7; use wavelet transform to perform multi-scale decomposition on the weighted data set D7, and output the decomposed data set D8; reconstruct the decomposed data set D8 to generate the final synchronization data set X';

[0119] Here, the nonlinear error problem of sensors under different working conditions is solved by performing range calibration and linear correction on various types of sensor data; the segmented statistical feature extraction method based on sliding window can not only detect data anomalies in time, but also retain the statistical characteristics of the data; the precise synchronization of multi-source heterogeneous data is achieved through the time index table and interpolation alignment algorithm; the multi-scale decomposition and reconstruction method based on wavelet transform is adopted to effectively suppress noise while retaining signal characteristics.

[0120] S15, read the synchronous data set X', calculate the short-term variance of the data, and generate the variance matrix V1; construct the data reliability weight based on the variance matrix V1, and output the weight matrix W1; adjust the basic covariance matrix R0 according to the weight matrix W1 to obtain the initial covariance matrix R1;

[0121] S16, receive the synchronous data set X' and the initial covariance matrix R1, use the exponential weighted moving average method to calculate the estimated error e(t), and generate the error sequence E1; calculate the adaptive coefficient λ according to the error sequence E1, and output the adaptive coefficient matrix A1; based on S17, read the initial covariance matrix R1 and the adjustment matrix M1, calculate the dynamic noise covariance matrix according to R(t)=R0+ΔR·exp(-λ||e(t)||²), and obtain the dynamic covariance matrix R2; construct the Kalman gain matrix, and output the gain matrix K1;

[0122] S17, input the synchronization data set X', the dynamic covariance matrix R2 and the gain matrix K1 into the adaptive Kalman filter, perform state estimation, and obtain the estimated state set S1; calculate the confidence interval of the estimated state set S1, and generate the confidence matrix C1; modify the estimated state set S1 according to the confidence matrix C1, and output the filtered data set Y;

[0123] S18. Receive the filtered data set Y, and use the nonlinear mapping function φ(x)=tanh(wx+b)·exp(-αx²) to perform feature transformation, where w is the weight coefficient, b is the bias term, and α is the attenuation coefficient; calculate the feature correlation matrix based on the transformed data to obtain the feature correlation matrix C; calculate the feature contribution based on the feature correlation matrix C, and select the features whose contribution is greater than the preset threshold to form the fused feature data F.

[0124] The design of step 1 above, by introducing segmented processing and statistical feature extraction based on sliding windows, enables the system to effectively identify and filter abnormal data points; the multi-scale decomposition and reconstruction method of wavelet transform not only retains the important features of the signal, but also effectively suppresses high-frequency noise; at the same time, through the dynamic weight adjustment mechanism of the time series data quality matrix, the adaptive fusion of data is realized. This method is particularly suitable for the situation where the quality of sensor data is unstable in the electric vehicle charging scenario. For example, when electromagnetic interference occurs during the charging process and causes distortion of some sensor data, the system can automatically reduce the weight of these data, thereby ensuring the reliability of data processing.

[0125] Step 2: By introducing topological optimization and Markov decision process based on graph theory, intelligent optimization of charging parameters is achieved. The specific plan is as follows:

[0126] S21. Construct an edge weight matrix w(e)=β1·P(e)+β2·D(e)+β3·Q(e) according to power transmission efficiency, transmission distance and link quality, where P(e) is power transmission efficiency, D(e) is transmission distance, and β1, β2, and β3 are dynamic weight coefficients; receive fused feature data F, extract power transmission efficiency P(e), transmission distance D(e) and link quality Q(e), and generate a feature combination matrix T1; use principal component analysis to calculate feature importance and output importance vector I1; calculate dynamic weight coefficients β1, β2, and β3 based on importance vector I1 to obtain weight coefficient set W2;

[0127] S22, read the feature combination matrix T1 and the weight coefficient set W2, construct the edge weight calculation function w(e), calculate the weight value of each edge in the network, and generate the edge weight matrix E2; establish a minimum spanning tree based on the edge weight matrix E2, and output the tree structure matrix T2;

[0128] S23, receiving the tree structure matrix T2, calculating the degree centrality and betweenness centrality of each node, and obtaining the centrality matrix C2; optimizing the network structure according to the centrality matrix C2, and generating the optimized structure matrix T3; extracting the topological features of the optimized structure matrix T3, and outputting the network topology matrix G;

[0129] S24. Read the network topology matrix G and the pre-stored historical state data H, and use the state transition probability update formula P(s'|s,a)=P0(s'|s,a)·exp(γ·ΔV(s,s')) to calculate the new transition probability, where P0 is the initial transition probability, γ is the learning rate, and ΔV is the state value function difference; generate the state transfer matrix P according to the updated state transition probability; optimize the power, frequency and phase parameters based on the state transfer matrix P, and output the optimized parameter set K.

[0130] In step 2, the optimal topology of the charging network is constructed by dynamically calculating edge weights and comprehensively considering multiple factors such as power transmission efficiency, transmission distance, and link quality. Then, the adaptive state transition probability matrix is ​​used to dynamically adjust the charging parameters. This method is particularly suitable for handling scenarios where multiple parking spaces at electric vehicle charging stations are charged simultaneously. When multiple vehicles are charging at the same time, the system can dynamically adjust the power allocation and charging sequence according to the actual situation of each charging location, avoiding the waste of resources and low charging efficiency caused by traditional fixed parameter solutions.

[0131] Step 3: By designing a hybrid neural network and a multi-objective optimization algorithm, accurate prediction of the charging state and intelligent generation of control strategies are achieved. The specific solution is as follows:

[0132] S31. Use LSTM network, GRU network and attention network to predict function y(t+1)=f(LSTM(x(t)))+g(GRU(x(t)))+h(Attention(x(t))), where f, g, h are weight functions. Input the optimization parameter set K and real-time status data R into the training modules of LSTM network, GRU network and attention network respectively. Use cross-validation method to divide the training set and validation set to obtain data set partition table D9. Train three networks based on data set partition table D9 to generate network parameter set P1. Calculate the performance index of each network on the validation set and output the performance index matrix M2.

[0133] S32, read the performance indicator matrix M2, use the fuzzy evaluation method to calculate the weight coefficients f, g, h of each network, obtain the network weight vector W3, input the test data into the trained network, obtain three sets of prediction results, generate the prediction result set Y1, perform weighted fusion on the prediction result set Y1 according to the network weight vector W3, and output the state prediction value Y;

[0134] S33. Read the state prediction value Y and the preset target parameter T, construct a multi-objective optimization function J=w1·E+w2·T+w3·S, where E is energy consumption, T is charging time, S is the safety risk index, w1, w2, w3 are dynamic weights, and calculate the optimization target matrix J. Based on the optimization target matrix J, a hierarchical charging strategy is generated, and the control strategy set S is output.

[0135] In step three, the LSTM, GRU and attention mechanism are organically combined to not only capture the long-term dependencies in the charging process, but also focus on key timing features; at the same time, by introducing a multi-objective optimization function that considers energy consumption, charging time and safety risks, the comprehensive optimization of the charging strategy is achieved. This method is particularly suitable for dealing with the diverse types of electric vehicle batteries and complex operating conditions.

[0136] Step 4: Through the consistency protocol and dynamic load balancing algorithm, the coordinated control of multiple chargers is realized. The specific design scheme is as follows:

[0137] S41. Use the state update rule xᵢ(k+1)=xᵢ(k)+∑aᵢⱼ(xⱼ(k)-xᵢ(k))·exp(-μ·d(i,j)) for collaborative computing, where aᵢⱼ is the adjacency matrix element, d(i,j) is the distance between nodes, and μ is the attenuation coefficient; generate the collaborative state matrix C;

[0138] Receive the control strategy set S, extract the initial state value x(k) of each node, and generate the state vector X1; construct the adjacency matrix a_ij, calculate the distance d(i,j) between nodes, and output the distance matrix D10; calculate the attenuation coefficient μ based on the distance matrix D10, and obtain the attenuation coefficient vector U1;

[0139] S42, read the state vector X1 and the attenuation coefficient vector U1, calculate the state increment of each node according to the state update rule, and obtain the increment matrix I2; update the node state value to generate the updated state vector X2; calculate the state consistency index and output the consistency index vector C3;

[0140] S43, evaluating the synergy effect according to the consistency index vector C3, if the consistency does not reach the preset threshold, returning to step S41 to continue iteration; when the preset threshold or the maximum number of iterations is reached, the updated state vector X2 is organized into a matrix form, and the synergy state matrix C is output;

[0141] S44. Read the collaborative state matrix C, and use the load distribution function L(i)=L0(i)+ΔL·sin(ωt+φᵢ) to calculate the load value of each charger, where L0(i) is the basic load, ΔL is the adjustment amplitude, ω is the cycle parameter, and φᵢ is the phase difference; generate a load distribution vector L; formulate an execution instruction according to the load distribution vector L, and output the execution instruction set E.

[0142] Step 4 introduces a state update rule that takes spatial distance into consideration, so that chargers with similar physical locations can work more closely together; at the same time, a dynamic balance of charging load is achieved through a periodic load distribution function. This method is particularly suitable for the management of large charging stations, and can effectively solve the problems of local overload and uneven resource allocation in traditional charging stations. When the load of a charging station changes suddenly, the system can complete load redistribution within milliseconds, avoiding system instability caused by local overload.

[0143] The technical scope of the present invention is not limited to the contents in the above description. Those skilled in the art can make various deformations and modifications to the above embodiments without departing from the technical idea of ​​the present invention, and these deformations and modifications should all fall within the protection scope of the present invention.

Claims

1. An intelligent wireless charging control system for electric vehicles, characterized in that: include: A central control module, the central control module is used to obtain status data related to the battery pack, analyze the status data, and output an execution signal; The central control module also stores the status data; The wireless charging module receives and executes the execution signal of the central control module, adjusts the corresponding charging parameters, and charges the battery; A charger module is adapted to the battery pack, and the battery is charged by wirelessly connecting the charger module to the wireless charging module; the charger module has an information collection function, and is used to exchange information with the central control module; the charger module is used to obtain status data related to the battery pack in the current state, and transmit it to the central control module; A sensor module, used to collect battery pack status data and related environmental data; The big data module is used to exchange information with the central control module.

2. An electric vehicle intelligent wireless charging control system as claimed in claim 1, characterized in that: The central control module includes a central control processing module, a central data storage module, a 4G module, and an API interface; The wireless charging module includes a ground unit status monitoring module, a ground unit control processing module, a charging module, and a wireless communication module. The charger module includes a battery status monitoring module, a battery data processing module, and a connector module; The sensor module includes a voltage sensor, a current sensor, a temperature sensor array, an electromagnetic field sensor, and a position sensor.

3. An intelligent wireless charging control method for electric vehicles, characterized in that: The following steps are involved: S1. Acquire real-time status data of the wireless charging control system, perform noise reduction processing on it, and obtain microsecond-level status data; S2. Intelligently optimize charging parameters based on the processed status data; S3, pre-build hybrid neural network and multi-objective optimization algorithm model, predict charging status and control the generation of charging strategy based on optimized charging parameters; S4. Based on the charging strategy obtained in step S3, dynamic balancing of charging load is achieved through periodic charging load distribution, thereby achieving coordinated control of multiple chargers.

4. The electric vehicle intelligent wireless charging control method as claimed in claim 3, characterized in that: Step S1 further comprises: S11, collecting battery voltage data V, battery current data I, coil temperature data Tc, ambient temperature data Te, electromagnetic field strength data Em, alignment deviation data D and communication quality data Q through sensors; synchronizing the sampling time to generate a time series data matrix X; Align the time stamps of the time series data matrix X and output the synchronized data set X'; S12, read the synchronous data set X', calculate the dynamic noise covariance matrix R(t) of the data, The function formula is, , Among them, R0 is the basic covariance matrix, e(t) is the estimation error, λ is the adaptive coefficient, and ΔR is the dynamic adjustment matrix; based on the obtained dynamic noise covariance matrix, the synchronous data set X' is adaptively Kalman filtered to output the filtered data set Y; S13, receiving the data set Y, using the nonlinear mapping function to transform the features, the function formula is: , where w is the weight coefficient, b is the bias term, and α is the attenuation coefficient; based on the data obtained after the transformation, the feature correlation matrix is ​​calculated to obtain the feature correlation matrix C; based on the feature correlation matrix C, the feature contribution is calculated, and the features with contributions greater than the preset threshold are selected to form the fused feature data F.

5. The electric vehicle intelligent wireless charging control method according to claim 4, characterized in that: Step S2 further comprises: S21, receiving fused feature data F, constructing edge weight matrix according to power transmission efficiency, transmission distance and link quality , where P(e) is the power transmission efficiency, D(e) is the transmission distance, Q(e) is the link quality, β1, β2, β3 are dynamic weight coefficients; a network topology graph is established based on the edge weight matrix, and the network topology matrix G is output; S22, read the network topology matrix G and the pre-stored historical state data H, and use the state transition probability update formula Calculate the new transition probability, where P0 is the initial transition probability, γ is the learning rate, and ΔV is the state value function difference; generate the state transition matrix P according to the updated state transition probability; The power, frequency and phase parameters are optimized based on the state transfer matrix P, and the optimized parameter set K is output.

6. The electric vehicle intelligent wireless charging control method according to claim 5, characterized in that: Step S3 further includes: S31, receiving the optimized parameter set K and the real-time status data R, respectively inputting them into the LSTM network, the GRU network and the attention network to obtain three sets of prediction results; using the prediction function Fusion of three sets of prediction results, where f, g, and h are weight functions; output of the fused state prediction value Y; S32, read the state prediction value Y and the preset target parameter T, and construct a multi-objective optimization function J = w1·E +w2·T+w3·S, where E is energy consumption, T is charging time, S is a safety risk index, and w1, w2, and w3 are dynamic weights; The optimization target matrix J is calculated; a hierarchical charging strategy is generated based on the optimization target matrix J, and a control strategy set S is output.

7. An electric vehicle intelligent wireless charging control method as claimed in claim 6, characterized in that: Step S4 further includes: S41, receiving control strategy set S, and updating rules based on current status Perform collaborative calculations, where aᵢⱼ is the adjacency matrix element, d(i, j) is the distance between nodes, and μ is the attenuation coefficient; generate the collaborative state matrix C.

8. Step S42, read the collaborative state matrix C, and use the load distribution function Calculate the load value of each charger, where L0(i) is the basic load, ΔL is the adjustment amplitude, ω is the cycle parameter, and φᵢ is the phase difference; generate a load distribution vector L; formulate execution instructions based on the load distribution vector L, and output the execution instruction set E.

Images