Search

CN-122026633-A - Self-adaptive impedance matching system based on hybrid optimization strategy and working method

CN122026633ACN 122026633 ACN122026633 ACN 122026633ACN-122026633-A

Abstract

The invention discloses a self-adaptive impedance matching system based on a hybrid optimization strategy and a working method thereof, and belongs to the technical field of megahertz wireless power transmission. Based on the impedance measuring unit, the invention can adjust the T-shaped matching network and the control unit, and utilizes the mode of combining the improved white whale optimization algorithm with the BP neural network to carry out parameter on-line optimization on the matching network, and control the on-off of a relay of the matching network so as to transform the load impedance to the required source impedance, thereby completing the self-adaptive impedance matching. The method of the invention uses the neural network to improve the initial population quality of the optimization algorithm, introduces a chained updating mechanism on the basis of the original beluga optimization algorithm and simplifies the optimizing strategy. The scheme is simple in design, and is an efficiency optimization scheme with high matching precision and high speed. The invention can adaptively process the impedance mismatch problem in the megahertz wireless power transmission system, thereby improving the energy transmission efficiency of the system.

Inventors

  • FAN XINGMING
  • XU SHUO
  • ZHANG XIN

Assignees

  • 南宁桂电电子科技研究院有限公司

Dates

Publication Date
20260512
Application Date
20260130

Claims (10)

  1. 1. An adaptive impedance matching system based on a hybrid optimization strategy, comprising: (1) The impedance measuring unit is used for detecting the input impedance of the wireless power transmission system in real time; (2) An adjustable T-shaped matching network connected with the impedance measuring unit and used for transforming the input impedance to the required source impedance; (3) The control unit is connected with the impedance measurement unit and the adjustable T-shaped matching network, is used for receiving the input impedance information, carries out on-line optimization on the parameters of the adjustable T-shaped matching network based on a mixed optimization strategy of combining an improved white whale optimization algorithm with the BP neural network, and generates a control signal to adjust the parameters of the matching network so as to realize self-adaptive impedance matching; The adjustable T-shaped matching network is of an L-C-L-shaped structure and comprises two inductance arrays and a capacitance array, each array is formed by connecting a plurality of discrete inductance or capacitance elements through a relay, and the equivalent inductance value or capacitance value of the array is changed by controlling the on-off combination of the relay; The control unit is a Micro Control Unit (MCU) configured to execute the hybrid optimization strategy, and the flow comprises: Step1, predicting an initial average value of parameters of a matching network according to the input impedance measured currently by utilizing a BP neural network which is trained offline in advance; step 2, generating an initial population for improving the white whale optimization algorithm through probability sampling based on the initial average value; Step 3, performing iterative optimization by running an improved beluga optimizing algorithm, wherein the algorithm comprises a chained updating stage, an exploring stage, a developing stage and a whale falling stage; and 4, converting the optimized optimal parameter combination into a relay control signal, and driving the adjustable T-shaped matching network to complete impedance transformation.
  2. 2. The hybrid optimization strategy-based adaptive impedance matching system of claim 1, wherein the impedance measurement unit comprises a dual directional coupler for separating the incident wave and the reflected wave signals and an AD8302 phase amplitude detection module for detecting the amplitude ratio and the phase difference of the incident wave and the reflected wave signals and calculating an input impedance.
  3. 3. The adaptive impedance matching system based on the hybrid optimization strategy according to claim 1, wherein the BP neural network has an input of a real part and an imaginary part of an input impedance, and an output of the BP neural network is a predicted mean value of three adjustable parameters in the matching network, and the improved beluga optimization algorithm introduces a chained update mechanism based on an original algorithm and simplifies an update strategy in a exploration and development stage.
  4. 4. The adaptive impedance matching system based on a hybrid optimization strategy according to claim 3, wherein the improved beluga optimization algorithm introduces a chained update mechanism based on the original algorithm and simplifies the update strategy, and the algorithm execution flow comprises the following steps: Step 1, generating an initial population of an improved beluga optimization algorithm by combining a sampling mode of normal distribution with a optimizing parameter mean value and a set standard deviation obtained by a BP neural network; Step 2, after initialization is completed, introducing a chain updating rule: ; Wherein, the Representing the first of the population The number of individuals who are to be treated, Representing the individuals with the best fitness value in the current population, Representing random numbers uniformly distributed over the interval 0,1, Representing individuals who are likely to be updated; step 3, defining boundary constraint and individual updating principle in the population, which is used for updating the optimal value after chained updating: ; ; Wherein, the Representation of First, the The values of the individual parameter components are chosen, Representation of First, the The upper bound of the individual parameter components, Representation of First, the The lower bound of the individual parameter components, Representation of Is used for the adaptation value of the (a), Representation of Is a fitness value of (a); Step 4, carrying out individual exploration or development behaviors in the population according to the required probability: ; ; Wherein, the Representing that the random numbers distributed uniformly are taken in the interval [0,1], Representing population quantity Taking the whole random number from the uniform distribution upwards in interval [0,1], and Not equal to , Representing the current number of iterations of the algorithm, Representing the total iteration number of the algorithm; Step 5, entering a whale stage according to probability after the exploration or development is finished: ; Wherein, the The probability of whale falling is indicated, Is shown in the interval Orally taking random numbers distributed uniformly; And step 6, after the current iteration times are greater than the set iteration times, ending the algorithm and outputting an optimizing result.
  5. 5. The adaptive impedance matching system based on the hybrid optimization strategy according to claim 1, wherein the prediction module of the BP neural network adopts a fully-connected network structure, and comprises an input layer, a hidden layer and an output layer, wherein the input layer comprises 2 neurons, and the input layer corresponds to input impedance respectively The real part of (2) And imaginary part The hidden layer is provided with a plurality of layers and a plurality of neurons according to the requirement, and a tansig activation function is adopted, and the expression is as follows: ; the output layer comprises 3 neurons, which respectively correspond to three adjustable parameters of the matching network , And Using purelin activation functions, the expression of which is: 。
  6. 6. The adaptive impedance matching system based on a hybrid optimization strategy according to claim 5, wherein the training data of the BP neural network is derived from WPT system simulation or experimental measurement, and the training process comprises the following steps: step 1, network initialization, randomly initializing connection weights and offsets among layers, setting learning rate, momentum factors and maximum training times, wherein the random initialization weights Bias and method of making same The expression is as follows: ; Wherein, the The index layer is represented as such, Initializing standard deviation; Step 2, forward propagation, inputting training samples into a grid, and calculating output of each layer, wherein the first layer The forward propagation of the layer is expressed as follows: ; Wherein, the , Is an activation function; step 3, calculating errors of network output and expected output, and adopting reflection coefficients as loss functions: ; Wherein, the In order to obtain the number of samples, Is the desired output; and 4, back propagation, namely reversely adjusting weights and offsets of all layers according to errors, wherein the formula is as follows: ; Wherein, the Is the learning rate; And 5, performing iterative training, and repeating the steps 2-4 until the maximum training times or errors are smaller than a set threshold value.
  7. 7. A method for adaptive impedance matching based on a hybrid optimization strategy, applied to the system as claimed in any one of claims 1-6, comprising the steps of: step 1, acquiring input impedance of a system in real time through an impedance measuring unit; step 2, the control unit receives the input impedance and runs a hybrid optimization strategy, and the strategy combines the rapid prediction of the BP neural network with the global optimizing capability of the improved beluga optimization algorithm to calculate the optimal parameter combination of the adjustable T-shaped matching network on line; Step 3, generating a control signal according to the optimal parameter combination, adjusting the on-off state of the relay of the adjustable T-shaped matching network, changing the network equivalent parameter, and matching the input impedance to the required source impedance; And 4, calculating the reflection coefficient after matching, if the reflection coefficient is lower than a set threshold value, judging that the matching is successful, otherwise, returning to the step 3 to perform parameter optimization and matching again.
  8. 8. The adaptive impedance matching method according to claim 6, wherein the performing of the hybrid optimization strategy includes outputting an initial estimate of the matching network parameters from the input impedance using the trained BP neural network, generating an initial population of the improved white whale optimization algorithm based on the initial estimate, performing parameter optimization by an iterative process including chain updating, exploring, developing and whale falling phases, and outputting the matching network parameters that minimize the reflection coefficient as an optimal solution.
  9. 9. The adaptive impedance matching method according to claim 7 or 8, wherein the improved white whale optimization algorithm directs the optimization process with a fitness function, the fitness function being a reflection coefficient, and the expression is: ; Wherein, the In order to match the source impedance to be desired, Representing the reflection coefficient, the smaller the reflection coefficient, the better the matching effect.
  10. 10. The adaptive impedance matching method based on the hybrid optimization strategy according to claim 7, wherein the parameter optimizing process of the adjustable T-type matching network is limited by the actual adjustable range of each inductor and capacitor array, and the optimizing parameters are two equivalent inductance values 、 And an equivalent capacitance value 。

Description

Self-adaptive impedance matching system based on hybrid optimization strategy and working method Technical Field The invention belongs to the technical field of megahertz wireless power transmission, and particularly relates to a self-adaptive impedance matching system based on a hybrid optimization strategy and a working method thereof. Background The wireless power transmission (Wireless Power Transfer, WPT) technology is taken as an innovative energy transmission mode which gets rid of the constraint of the traditional wired power supply, and by virtue of the core advantages of safety, convenience and reliability, the rapid development and wide application are realized in the past decades. From wireless charging of electric automobiles, energy supply of biomedical implanted devices (such as cardiac pacemakers and nerve stimulators), to the fields of non-contact power supply of household appliances in smart home, autonomous charging of mobile robots in industrial scenes and the like, the wireless power transmission technology gradually changes the traditional mode of energy supply, and becomes one of key support technologies for promoting the upgrading of related industries. Among the numerous wireless power transmission bands, the megahertz (MHz) band has demonstrated great potential for development by virtue of its unique technical characteristics. Compared with a low frequency band (such as a kHz level), the wireless power transmission system in the megahertz frequency band has higher spatial freedom, can realize power transmission without strict coil alignment, greatly improves the use convenience, and meanwhile, the frequency band can support the design of a transmitting and receiving coil with smaller size, is favorable for the miniaturization and integration of equipment, and is particularly suitable for scenes with strict requirements on volume and installation space, such as miniature implantable medical equipment, portable electronic terminals and the like. In addition, the energy transmission efficiency of the megahertz frequency band is excellent in a middle-short distance scene, the transmission distance and the energy loss can be balanced, and the actual requirements of most civil and industrial scenes are met, so that the method becomes a research hot spot and a key development direction in the current wireless electric energy transmission technical field. However, in practical applications, the megahertz wireless power transmission system faces a serious impedance mismatch problem, which has become a core bottleneck that restricts the improvement of the energy transmission efficiency of the system. The generation of the system impedance mismatch is caused by the comprehensive influence of various factors, namely, on one hand, the inverter in the system has obvious load sensitivity, when the load demand changes (such as battery SOC state change, household appliance working mode switching and the like in the electric automobile charging process), the load impedance changes, so that the output impedance of the inverter cannot be matched with the load impedance, and on the other hand, the coupling coefficient of the wireless power transmission system is easily influenced by the external environment and the use condition, such as distance deviation, angle deflection and relative position change between a transmitting coil and a receiving coil, interference of surrounding metal barriers and the like, the coupling coefficient can be fluctuated, and the original impedance matching state of the system is further destroyed. When impedance mismatch occurs, part of transmission energy returns to the transmitting end in the form of reflected waves, so that the energy transmission efficiency is greatly reduced, voltage and current distortion of a circuit of the transmitting end can be possibly caused, heating loss of a device is increased, even core power devices such as an inverter can be damaged when serious, and the stability and the service life of a system are influenced. In order to solve the problem of impedance mismatch, the energy transmission efficiency and stability of the system are improved, and an adaptive impedance matching technology is generated. According to the technology, the adjustable matching network and the measurement control module are introduced, the impedance state of the system is monitored in real time, and the matching network parameters are dynamically adjusted, so that negative effects caused by impedance mismatch are restrained, and the technology is an effective means for improving the performance of the megahertz wireless power transmission system currently. The optimization of the matching network parameters is a core link of the self-adaptive impedance matching system, and the performance of an optimization algorithm directly determines the matching speed and the matching precision, so that the actual use effect of the whole wireless power transmission system is aff