Synchronous detection method of RF power signal based on real-time spectrum analysis

Abstract

The present application provides a method for synchronous detection of RF power supply signals based on real-time spectrum analysis, and relates to the field of RF power supply control technology. The method includes: performing feature analysis on RF power supply signals based on spectrum feature indicators to obtain key spectrum parameters; determining parameter deviation data, and performing control analysis on parameter deviation data using a multivariate control channel to generate a synchronous control signal; performing adaptive control based on the synchronous control signal, performing control quality analysis on the control RF power supply signal according to the expected control indicator, and determining the control deviation data to optimize the multivariate control channel. The present application can solve the technical problems of the existing RF power supply synchronous control method, which cannot efficiently and accurately detect and analyze the RF power supply signal, resulting in poor accuracy and timeliness of RF power supply control, and poor operational stability and reliability of RF power supply, and can significantly improve the overall performance and operational stability of the RF power supply.

Description

Technical Field

[0001] The present application relates to the field of radio frequency power supply control technology, and in particular to a method for synchronously detecting radio frequency power supply signals based on real-time spectrum analysis. Background Art

[0002] In RF communication, RF power supplies provide the necessary RF energy for wireless communication devices to achieve information transmission. However, the output signal of RF power supplies often fluctuates and deviates, causing wireless communication devices to not work properly. Therefore, the synchronous control technology of RF power supplies came into being, aiming to achieve precise control of RF power supplies by real-time monitoring and analysis of RF power supply signals and timely feedback adjustment.

[0003] At present, the existing RF power synchronization control method is unable to perform efficient and accurate detection and analysis of the RF power signal, resulting in the inability to timely and accurately correct the control deviation, resulting in poor accuracy and timeliness of RF power control, and causing technical problems such as poor operating stability and reliability of the RF power supply. Summary of the invention

[0004] The purpose of this application is to provide a method for synchronous detection of RF power supply signals based on real-time spectrum analysis, so as to solve the technical problem that the existing RF power supply synchronous control method cannot perform efficient and accurate detection and analysis of RF power supply signals, resulting in the inability to timely and accurately correct control deviations, resulting in poor accuracy and timeliness of RF power supply control, and causing poor operational stability and reliability of the RF power supply.

[0005] In view of the above problems, the present application provides a method for synchronous detection of RF power signals based on real-time spectrum analysis, wherein the method includes: using a spectrum analyzer to monitor and capture RF power signals, and performing signal denoising based on a predetermined processing scheme to determine a standard RF power signal; based on spectrum feature indicators, combining spectrum analysis technology to perform feature analysis on the standard RF power signal to obtain key spectrum parameters; performing deviation analysis on the key spectrum parameters according to the standard spectrum parameters to determine parameter deviation data, and using a multivariate control channel to perform control analysis on the parameter deviation data to generate a synchronous control signal; based on the synchronous control signal, performing adaptive control of the RF power supply, tracking and monitoring the regulated RF power signal, performing control quality analysis on the regulated RF power signal according to expected control indicators, and determining control deviation data; recording the parameter deviation data and the control deviation data, constructing control correction data, and performing model optimization on the multivariate control channel according to the control correction data within a preset period.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0007] The standard RF power signal is determined by monitoring and capturing the RF power signal with a spectrum analyzer, and performing signal denoising based on a predetermined processing scheme; based on spectrum characteristic indicators, the standard RF power signal is characterized by combining spectrum analysis technology to obtain key spectrum parameters; deviation analysis is performed on the key spectrum parameters according to the standard spectrum parameters to determine parameter deviation data, and the parameter deviation data is regulated and analyzed by a multivariate control channel to generate a synchronous control signal; based on the synchronous control signal, adaptive control of the RF power supply is performed, the RF power signal is tracked and monitored, and the control quality analysis of the regulated RF power signal is performed according to the expected control indicators to determine the control deviation data; the parameter deviation data and the control deviation data are recorded to construct control correction data, and the model of the multivariate control channel is optimized according to the control correction data within a preset period. That is to say, by performing signal denoising on the RF power signal based on a predetermined processing scheme to obtain a standard RF power signal, the accuracy and reliability of the standard RF power signal can be improved; then, using spectrum analysis technology, the standard RF power signal is characterized according to the spectrum characteristic index to obtain key spectrum parameters, which can improve the accuracy of the key spectrum parameters; further, the key spectrum parameters are analyzed for deviation according to the standard spectrum parameters to determine the parameter deviation data, and the parameter deviation data is regulated and analyzed through the multivariate control channel to obtain a synchronous control signal; then, based on the synchronous control signal, the RF power is adaptively controlled, and after the adaptive control, the RF power signal is tracked and monitored, and then the control quality analysis of the RF power signal is performed according to the expected control index to determine the control deviation data; finally, the parameter deviation data and the control deviation data are recorded to construct the control correction data, and the multivariate control channel is model optimized according to the control correction data within the preset period. The precision and accuracy of the RF power signal detection and analysis can be improved, and then the RF power control deviation can be corrected in time according to the signal detection result, and the accuracy and timeliness of the RF power control can be improved, thereby achieving the technical effect of significantly improving the overall performance and operation stability of the RF power supply.

[0008] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented according to the contents of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically cited below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0010] Figure 1 This is a flow chart of a method for synchronously detecting radio frequency power supply signals based on real-time spectrum analysis in the present application;

[0011] Figure 2 This is a schematic diagram of a flow chart for determining spectrum characteristic indicators in a method for synchronous detection of radio frequency power supply signals based on real-time spectrum analysis in the present application. DETAILED DESCRIPTION

[0012] This application provides a method for synchronous detection of RF power signals based on real-time spectrum analysis, which solves the problem that the existing RF power synchronous control method cannot accurately and timely correct the control deviation due to the inability to efficiently and accurately detect and analyze the RF power signal, resulting in poor accuracy and timeliness of RF power control, and poor operational stability and reliability of RF power. The precision and accuracy of RF power signal detection and analysis can be improved, and then the RF power control deviation can be corrected in time according to the signal detection results, improving the accuracy and timeliness of RF power control, thereby achieving the technical effect of significantly improving the overall performance and operational stability of the RF power supply.

[0013] Below, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited to the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application. It should also be noted that, for the convenience of description, only the parts related to the present application are shown in the accompanying drawings, rather than all of them.

[0014] For examples, please see the attached Figure 1 The present application provides a method for synchronous detection of radio frequency power supply signals based on real-time spectrum analysis, and the method specifically comprises the following steps:

[0015] Step 1: Use a spectrum analyzer to monitor and capture the RF power signal, and perform signal denoising based on a predetermined processing scheme to determine a standard RF power signal.

[0016] Specifically, a high-performance spectrum analyzer is a device used to measure and analyze RF signals. It can capture and display the frequency components and power levels of signals. During the operation of the RF power supply, the spectrum analyzer is used to monitor and capture the RF power supply signal to obtain the RF power supply signal. For example, the output signal of the RF power supply is connected to the input port of the spectrum analyzer to ensure the quality of the connecting cable and interface to reduce signal loss and interference. Then, the RF power supply signal is denoised based on a predetermined processing scheme to obtain a standard RF power supply signal, wherein the preprocessing scheme includes filtering, gain adjustment, time domain or frequency domain denoising and other steps, such as applying signal denoising techniques such as median filtering, wavelet denoising, adaptive filtering, etc. to reduce the impact of noise on the signal. By performing signal denoising based on a predetermined processing scheme, the accuracy and reliability of the standard RF power supply signal can be improved, providing support for the subsequent synchronous and precise control of the RF power supply.

[0017] Step 2: Based on the spectrum characteristic index, the spectrum analysis technology is combined to perform characteristic analysis on the standard RF power signal to obtain key spectrum parameters.

[0018] Specifically, spectrum characteristic analysis is performed based on sample data to determine spectrum characteristic indicators, where spectrum characteristic indicators refer to spectrum characteristics that are highly correlated with the operating quality of the RF power supply, such as frequency deviation, harmonic components, etc. Then, spectrum analysis technology is used to perform characteristic analysis on the standard RF power supply signal according to the spectrum characteristic indicators to obtain key spectrum parameters, where spectrum analysis technology is a technology used to determine the frequency components of the signal. It involves converting the signal from the time domain to the frequency domain in order to more intuitively understand the frequency components of the signal, for example: Fourier transform, wavelet transform and other methods can be used for characteristic analysis. By obtaining key spectrum parameters, a basis is provided for the subsequent parameter deviation analysis and synchronous regulation of the RF power supply.

[0019] Step three: perform deviation analysis on the key spectrum parameters according to the standard spectrum parameters, determine parameter deviation data, use a multi-element control channel to perform control analysis on the parameter deviation data, and generate a synchronous control signal.

[0020] Specifically, first, according to the design specifications and performance standards of the RF power supply, a series of standard spectrum parameters are set, including standard center frequency, standard bandwidth, standard harmonic distortion, standard phase noise, standard signal-to-noise ratio, etc., wherein the standard spectrum parameters are the same as the parameters in the spectrum characteristic indicators; then, the extracted key spectrum parameters are compared with the standard spectrum parameters, and the deviation values ​​are calculated, wherein the deviation analysis can adopt statistical methods, such as calculating the average deviation, standard deviation, etc.; then, according to the results of the deviation analysis, the parameter deviation data is determined, wherein the parameter deviation data includes the deviation value between each key spectrum parameter and the standard parameter.

[0021] A multivariate control channel is constructed, wherein the multivariate control channel includes multiple power supply conditions and corresponding multiple control analysis channels, wherein the control analysis channel includes multiple control analysis units, and the construction principle of each control analysis unit is different, such as: BP neural network, random decision forest, etc.; first, real-time power supply conditions are collected and acquired, wherein the real-time power supply conditions include key parameters such as output power, frequency, phase, temperature, voltage, etc.; then the real-time power supply conditions are input into the multivariate control channel for matching, and a matching control channel is determined, wherein the matching control channel may be a combination of one or more channels, which can provide an optimal control strategy; then the parameter deviation data is input into the matching control channel for control analysis, and a synchronous control signal is output, wherein the synchronous control signal is the control parameter of the RF power supply, including power adjustment, frequency adjustment, phase adjustment, etc.

[0022] By constructing a multi-element control channel, it provides support for the intelligent control and analysis of parameter deviation data, while improving the accuracy and efficiency of the control analysis and providing a basis for the synchronous and precise control of the RF power supply.

[0023] Step 4: Execute adaptive control of the RF power supply based on the synchronous control signal, track and monitor the regulated RF power supply signal, perform control quality analysis on the regulated RF power supply signal according to the expected control index, and determine the control deviation data.

[0024] Specifically, according to the synchronous control signal output by the matching control channel, the adaptive control of the RF power supply is performed, wherein the control includes power adjustment, frequency adjustment, phase adjustment, etc., to correct the deviation and optimize the RF power supply output; then the spectrum analyzer is used to track and monitor the RF power supply signal after control, including key parameters such as output power, frequency, and phase. The expected control index is obtained, and the expected control index can be set according to actual needs, including signal stability index, frequency deviation index, etc.; further, according to the expected control index, the quality analysis of the RF power supply signal after control is performed, such as: comparing standard parameters, evaluating signal stability, checking frequency deviation, etc., and determining the control deviation data according to the quality analysis results, wherein the deviation data includes the difference between the actual control parameters and the expected control index. By obtaining the control deviation data, the control effect can be understood in real time, and the control strategy can be adjusted and optimized accordingly to achieve a better control effect.

[0025] Step 5: Record the parameter deviation data and the control deviation data, construct control correction data, and optimize the model of the multivariate control channel according to the control correction data within a preset period.

[0026] Specifically, the control parameters are traced and analyzed according to the control deviation data to obtain standard control parameters. Standard control parameters refer to control parameters that can eliminate control deviations and achieve the best control effect. Then the parameter deviation data and standard control parameters are recorded as control correction data; the control correction data set within the preset period is obtained, wherein the preset period is the model optimization update period, which can be set according to the actual situation, for example: the model optimization update period is set to 24 hours; the control correction data set reflects the control effect and deviation of the RF power supply under different working conditions; then the model of the multivariate control channel is optimized according to the control correction data within the preset period, wherein the model optimization method includes adjusting control parameters, improving control algorithms, introducing new control strategies, etc., and the optimized model is further verified using the new control data set, such as verifying the predictive performance and control effect of the model to ensure the effectiveness of the model in practical applications; finally, subsequent synchronous control is performed according to the optimized multivariate control channel. Through the above steps, the performance and efficiency of RF power supply control can be further improved to ensure the stability and reliability of its output signal.

[0027] By providing a method for synchronous detection of radio frequency power supply signals based on real-time spectrum analysis, the existing radio frequency power supply synchronous control method can solve the problem that the control deviation cannot be corrected in time and accurately due to the inability to efficiently and accurately detect and analyze the radio frequency power supply signal, the accuracy and timeliness of the radio frequency power supply control are poor, and the technical problem of poor operation stability and reliability of the radio frequency power supply is caused. First, the radio frequency power supply signal is monitored and captured by a spectrum analyzer, and the signal is denoised based on a predetermined processing scheme to determine the standard radio frequency power supply signal; then, based on the spectrum characteristic index, the standard radio frequency power supply signal is characterized by combining the spectrum analysis technology to obtain key spectrum parameters; then, the key spectrum parameters are analyzed for deviation according to the standard spectrum parameters to determine the parameter deviation data, and the parameter deviation data is regulated and analyzed by a multivariate control channel to generate a synchronous control signal; next, the adaptive control of the radio frequency power supply is performed based on the synchronous control signal, the radio frequency power supply signal is tracked and monitored, and the control quality analysis of the control radio frequency power supply signal is performed according to the expected control index to determine the control deviation data; finally, the parameter deviation data and the control deviation data are recorded, the control deviation data is constructed, and the model of the multivariate control channel is optimized according to the control deviation correction data within a preset period. It can improve the precision and accuracy of RF power signal detection and analysis, and then timely correct RF power control deviations according to signal detection results, improve the accuracy and timeliness of RF power control, thereby achieving the technical effect of significantly improving the overall performance and operational stability of the RF power supply.

[0028] Further, determine the spectrum characteristic indicators, such as Figure 2 As shown, step 2 of this application includes:

[0029] Construct a spectrum feature space, wherein the spectrum feature space includes at least frequency deviation, bandwidth, harmonic component, phase noise, signal-to-noise ratio and sideband; take the radio frequency power source attribute information as a constraint, perform homologous search based on industrial big data, and obtain a sample data set; based on the sample data set, sequentially perform impact evaluation on multiple characteristic indicators in the spectrum feature space and the operating quality of the radio frequency power source, and generate a characteristic indicator sequence according to the evaluation results; extract a preset number of characteristic indicators in the characteristic indicator sequence and set them as the spectrum characteristic indicators.

[0030] Specifically, first, a spectrum feature space is constructed, where the spectrum feature space includes multiple spectrum feature indicators. The spectrum feature indicators refer to indicators that are related to the operating quality of the RF power supply, including at least frequency deviation, bandwidth, harmonic components, phase noise, signal-to-noise ratio and sideband. Then, the RF power supply attribute information is used as a search constraint, where the RF power supply attribute information includes output power, frequency range, bandwidth, etc., which can be set according to the design specifications and performance standards of the RF power supply. Then, an industrial big data platform or database is used to perform a homologous search based on the attribute information of the RF power supply, including historical operating data, fault records, maintenance records, etc. of the RF power supply; sample data that can reflect the performance and operating quality of the RF power supply is further extracted from the search data to obtain a sample data set.

[0031] Then, based on the sample data set, the impact of multiple characteristic indicators in the spectrum feature space and the operating quality of the RF power supply is evaluated in turn, where the operating quality evaluation indicators include power stability, frequency stability, harmonic distortion, etc., that is, the correlation between the characteristic indicators and the operating quality is analyzed, where the greater the correlation, the higher the impact, and multiple impacts of multiple characteristic indicators are obtained; then, the multiple characteristic indicators are arranged from large to small according to the impact, and a characteristic indicator sequence is generated; finally, a preset number of characteristic indicators in the characteristic indicator sequence are extracted and set as the spectrum characteristic indicators, and the preset number can be set according to actual needs. The higher the required accuracy, the larger the preset number, where the spectrum characteristic indicators are key features that can effectively reflect the performance and operating quality of the RF power supply.

[0032] Further, to generate a characteristic indicator sequence, the present application also includes the following steps:

[0033] A power quality evaluation index set is obtained, wherein each power quality evaluation index includes an index importance; a first characteristic index and a first power quality evaluation index are randomly selected, and based on a single variable analysis strategy, the first characteristic index and the first power quality evaluation index are evaluated for correlation using the sample data set to generate a first index correlation coefficient; multiple index correlation coefficients of the first characteristic index are determined by iterative analysis, and the multiple correlation coefficients are comprehensively evaluated according to the index importance to obtain a first correlation coefficient; multiple correlation coefficients of multiple characteristic indicators are obtained in turn, and the multiple characteristic indicators are arranged from large to small according to the correlation coefficient to obtain the characteristic indicator sequence.

[0034] Specifically, first, a power quality evaluation index set is obtained, wherein the power quality evaluation index set includes multiple power quality evaluation indicators, and each power quality evaluation indicator includes an indicator importance, wherein the power quality evaluation indicators include power stability, frequency stability, harmonic components and power noise, and then each power quality evaluation indicator is evaluated for its importance to determine its weight in the evaluation. Commonly used evaluation methods include expert scoring, entropy weight method and other methods. According to the indicator importance evaluation results, the weight of each indicator is determined, and the weight reflects the relative importance of the indicator in the evaluation.

[0035] Then, any one of the multiple characteristic indicators is randomly selected as the first characteristic indicator, and any one of the multiple power quality evaluation indicators is randomly selected as the first power quality evaluation indicator. Then, based on the single variable analysis strategy, the sample data set is used to evaluate the correlation between the first characteristic indicator and the first power quality evaluation indicator. The single variable analysis strategy is a statistical analysis method that focuses on the influence or correlation of one variable on another variable. In this analysis, usually only one variable is considered, and its role in explaining or predicting another variable is evaluated, such as the Pearson correlation coefficient, the Spearman rank correlation coefficient, etc., which are used to evaluate the linear correlation between the two variables; using the selected single variable analysis method, the first characteristic indicator and the first power quality evaluation indicator are evaluated for correlation, and the first indicator correlation coefficient will be generated according to the evaluation result, indicating the degree of correlation between the two variables. The higher the correlation, the stronger the correlation between the two variables, and the larger the first indicator correlation coefficient.

[0036] Then, the same method as that used to obtain the first indicator correlation coefficient is used for iterative analysis to determine the correlation coefficients of multiple indicators of the first characteristic indicator, and the multiple correlation coefficients are weighted according to the indicator importance, and the weighted calculation result is used as the first correlation coefficient. By obtaining the first correlation coefficient by weighting according to the indicator importance, the accuracy and reliability of the setting of the first correlation coefficient can be improved. Multiple correlation coefficients of multiple characteristic indicators are obtained in sequence, and the multiple characteristic indicators are arranged from large to small according to the correlation coefficient to obtain the characteristic indicator sequence.

[0037] Further, to construct a multi-modulation channel, step three of this application includes:

[0038] Taking the RF power attribute information as a constraint, query the power control log to determine multiple sample power conditions and multiple sample spectrum parameter deviation sets, wherein the sample power conditions and the sample spectrum parameter deviation sets have a corresponding relationship; randomly select a first sample power condition, obtain a first sample spectrum parameter deviation set of the first sample power condition, and construct a first sample parameter space based on the first sample spectrum parameter deviation set; randomly generate multiple first sample parameters based on the first sample parameter space, wherein the first sample parameters satisfy a predetermined uniform constraint and a predetermined quantity constraint; for the purpose of satisfying multiple first sample parameters, optimize the control parameters respectively based on the control evaluation function, determine multiple optimal control parameters, and construct a first training set based on the multiple first sample parameters and the multiple optimal control parameters; use the first training set to perform supervised training and cross-validation on the initial control channel to obtain a first control channel that meets the expected convergence constraint; construct the multivariate control channel based on the mapping association between the first sample power condition and the first control channel.

[0039] Specifically, first, with the RF power source attribute information as a constraint condition, query the control log of the same RF power source, that is, extract the sample power source conditions and sample spectrum parameter deviations related to the RF power source attribute information from the control log, and retrieve multiple sample power source conditions and multiple sample spectrum parameter deviation sets, wherein the sample power source conditions and the sample spectrum parameter deviation sets have a corresponding relationship. Then, randomly select any sample power source condition from the multiple sample power source conditions as the first sample power source condition, and obtain the first sample spectrum parameter deviation set of the first sample power source condition; further, use the minimum sample spectrum parameter deviation in the first sample spectrum parameter deviation set as the lower limit of the sample parameter space, and use the maximum sample spectrum parameter deviation in the first sample spectrum parameter deviation set as the upper limit of the sample parameter space to construct the first sample parameter space.

[0040] Configure a predetermined uniform constraint and a predetermined quantity constraint. The uniform constraint is used to ensure that the first sample parameter is evenly distributed in the parameter space, such as setting the similarity of any two sample parameters to be less than a preset threshold, etc., to meet the uniformity constraint. The quantity constraint determines the size of the generated first sample parameter set; then, based on the predetermined uniform constraint and the predetermined quantity constraint, a plurality of first sample parameters are randomly generated in the first sample parameter space. Further, for the purpose of meeting a plurality of first sample parameters, the control parameters are optimized based on the control evaluation function, wherein the control evaluation function has a plurality of evaluation indicators, such as power stability, frequency stability, harmonic components, power noise, etc.; first, determine the possible value range of the control parameter, including the power adjustment parameter threshold, the frequency adjustment parameter threshold, the phase adjustment parameter threshold, etc.; then, based on the possible value range of the control parameter, randomly generate a plurality of control parameters, and use the optimization algorithm to optimize the control parameters, such as: genetic algorithm, particle swarm optimization algorithm, etc.; after the optimization process, determine a plurality of optimal control parameters, the optimal control parameter refers to the control parameter that can achieve the best control effect under the given control evaluation function; then construct the first training set based on the mapping relationship between the plurality of first sample parameters and the plurality of optimal control parameters.

[0041] The first training set is further used as sample training data to perform supervised training and cross-validation on the initial control channel, wherein the loss function and gradient descent algorithm can be used for training to obtain a first control channel that meets the expected convergence constraints; then, multiple control channels of multiple sample power conditions are obtained in turn, and the multivariate control channel is constructed based on the mapping association between the multiple sample power conditions and the multiple control channels.

[0042] Further, to construct a control evaluation function, the present application also includes the following steps:

[0043] The expression of the regulation evaluation function is:

[0044] Among them, P(i) represents the fitness of the i-th control parameter, w1 is the indicator importance of power stability, EY i is the power variance after the i-th control parameter is adjusted, w2 is the importance of the frequency stability indicator, EV i is the frequency variance after the i-th control parameter is adjusted, w3 is the harmonic distortion weight, S i is the harmonic content after the i-th control parameter is adjusted, w4 is the power noise weight, C i is the power noise content after the i-th control parameter is adjusted.

[0045] Specifically, in the control evaluation function, P(i) represents the fitness of the i-th control parameter, where the greater the fitness, the better the control effect; w1 is the indicator importance of power stability, EY iis the power variance after the i-th control parameter is adjusted, and the power variance is inversely proportional to the power stability; w2 is the importance of the frequency stability indicator, EV i is the frequency variance after the i-th control parameter is adjusted, and the frequency variance is inversely proportional to the frequency stability; w3 is the harmonic distortion weight, S i is the harmonic content after the i-th control parameter is adjusted, w4 is the power noise weight, C i is the power noise content after the i-th control parameter is adjusted.

[0046] Further, to obtain the first control channel, the present application also includes the following steps:

[0047] Configure a control analysis operator, and build multiple initial analysis units based on the control analysis operator, wherein the control analysis operator includes at least a BP neural network, a random decision forest, and a long short-term memory network; divide the first training set into N equal parts, select N times without replacement to construct the first training set, iterate and select N times to obtain N training sets, wherein N is the number of control analysis operators; use the N training sets to perform supervised training and cross-validation on multiple initial analysis units respectively, and obtain multiple convergent analysis units that meet expected convergence constraints; determine multiple convergence coefficients of the multiple convergent analysis units, and build an analysis output layer according to the multiple convergence coefficients, wherein the output of the analysis output layer is a weighted result of the multiple convergent analysis units; build a control analysis layer in parallel based on the multiple convergent analysis units, and generate the first control channel in combination with the analysis output layer.

[0048] Specifically, first, configure the regulation and analysis operator, wherein the regulation and analysis operator includes at least BP neural network, random decision forest and long short-term memory network, and additional regulation and analysis operators can also be added according to actual conditions. The more regulation and analysis operators there are, the higher the accuracy and reliability of regulation and analysis; wherein the BP neural network is a feedforward neural network, which is widely used in classification and regression problems. It contains an input layer, a hidden layer and an output layer, and is trained through a back propagation algorithm, and can be used to deal with nonlinear problems; random decision forest is an integrated learning method, which consists of multiple decision trees, and improves the accuracy and stability of prediction by constructing multiple decision trees and taking the average or majority vote; long short-term memory network is a recurrent neural network that can learn long-term dependencies, contains special cell states and gating structures, and can process sequence data. Then, multiple initial analysis units are built based on the regulation and analysis operator.

[0049] The first training set is divided into N parts to obtain N data sets, and then N times are selected from the N data sets without replacement to construct the first training set, and N training sets are selected iteratively N times, where N is the number of regulatory analysis operators; then the N training sets are used to supervise the training of multiple initial analysis units, where the loss function and gradient descent algorithm can be used for supervised training, and after the preset number of training times is met, the training sets are exchanged to perform verification training on the multiple initial analysis units to obtain multiple convergent analysis units that meet the expected convergence constraints, where the expected convergence constraints are the expected accuracy. Further determine multiple convergence coefficients of the multiple convergent analysis units, where the convergence coefficient is the output accuracy of the convergent analysis unit; then build an analysis output layer based on the multiple convergence coefficients, where the output of the analysis output layer is the weighted result of the multiple convergent analysis units. On the other hand, a regulatory analysis layer is built in parallel based on multiple convergent analysis units, and the first regulatory channel is generated in combination with the analysis output layer.

[0050] Further, to generate a synchronous control signal, step three of the present application includes:

[0051] Collect real-time operating data of the RF power supply, input the real-time operating data into a multivariate control channel for similarity matching, and determine an associated control channel, wherein the similarity between the operating data of the associated control channel and the real-time operating data meets a predetermined similarity threshold; use the associated control channel to control and analyze the parameter deviation data, and output the synchronous control signal.

[0052] Specifically, the real-time working condition data of the RF power supply is collected, wherein the real-time working condition data includes output power, frequency, phase, temperature, voltage, etc.; then the real-time working condition data is input into the multivariate control channel, and the similarity between the real-time working condition data and the working condition data of the control channel is calculated, for example: the cosine similarity algorithm, Euclidean distance, etc. can be used for similarity analysis, and the control channels with similarity greater than the predetermined similarity threshold are selected as associated control channels, wherein there are one or more associated control channels. The parameter deviation data is further controlled and analyzed using the associated control channel, and the synchronous control signal is output. If there are multiple associated control channels, the synchronous control signal is the average of the output results of the multiple associated control channels.

[0053] Further, the model of the multi-regulatory channel is optimized, and step five of this application includes:

[0054] Acquire multiple control deviation data within a preset period, perform control parameter traceability analysis based on the multiple control deviation data, and determine multiple standard control parameters; generate a control correction data set based on the multiple parameter deviation data and the multiple standard control parameters, divide the control correction data set according to the working conditions, and determine multiple working condition correction data sets; based on the incremental learning principle, use the multiple working condition correction data sets to perform mapping optimization training on the multiple control channels within the multivariate control channel.

[0055] Specifically, firstly, a plurality of control deviation data within a preset period are obtained, and then control parameter traceability analysis is performed on the plurality of control deviation data to obtain a plurality of standard control parameters, wherein the standard control parameter refers to a control parameter that can eliminate the control deviation and achieve the best control effect. Then, a control correction data set is generated according to the plurality of parameter deviation data and the plurality of standard control parameters, and the control correction data set is divided according to the working conditions to determine a plurality of working condition correction data sets.

[0056] Incremental learning is a machine learning method that allows the learning system to continuously learn new knowledge from new samples while retaining most of the previously learned knowledge. It allows the model to continuously learn new data based on existing knowledge. It is suitable for scenarios where data is constantly updated and can avoid retraining the entire model. Then, based on the principle of incremental learning, multiple working condition correction data sets are used to perform mapping optimization training on multiple control channels within the multivariate control channel, that is, the working condition correction data set is used as sample data to optimize the control channels with mapping associations. By optimizing the control channels based on incremental learning, the adaptability and performance of the model can be continuously improved, thereby further improving the accuracy and reliability of the synchronous control of the RF power supply.

[0057] In summary, the method for synchronous detection of RF power supply signals based on real-time spectrum analysis provided by the present application has the following technical effects:

[0058] By performing signal denoising on the RF power signal based on a predetermined processing scheme to obtain a standard RF power signal, the accuracy and reliability of the standard RF power signal can be improved; then, using spectrum analysis technology, the standard RF power signal is analyzed according to the spectrum characteristic index to obtain key spectrum parameters, which can improve the accuracy of the key spectrum parameters; further, the key spectrum parameters are analyzed according to the standard spectrum parameters to determine the parameter deviation data, and the parameter deviation data is analyzed and regulated through the multivariate control channel to obtain a synchronous control signal; then, based on the synchronous control signal, the RF power is adaptively controlled, and after the adaptive control, the RF power signal is tracked and monitored, and then the control quality analysis of the RF power signal is performed according to the expected control index to determine the control deviation data; finally, the parameter deviation data and the control deviation data are recorded to construct the control correction data, and the multivariate control channel is optimized according to the control correction data within the preset period. The precision and accuracy of the RF power signal detection and analysis can be improved, and then the RF power control deviation can be corrected in time according to the signal detection result, and the accuracy and timeliness of the RF power control can be improved, thereby achieving the technical effect of significantly improving the overall performance and operation stability of the RF power supply.

[0059] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

[0060] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technology, the present application is also intended to include these modifications and variations.

Claims

1. A method for synchronously detecting radio frequency power supply signals based on real-time spectrum analysis, characterized in that: The method comprises: Using a spectrum analyzer to monitor and capture the RF power signal, and performing signal denoising based on a predetermined processing scheme to determine a standard RF power signal; Based on the spectrum characteristic index, the standard RF power signal is analyzed in combination with spectrum analysis technology to obtain key spectrum parameters; Performing deviation analysis on the key spectrum parameters according to the standard spectrum parameters to determine parameter deviation data, performing regulation and analysis on the parameter deviation data using a multi-element regulation channel to generate a synchronous regulation signal; Based on the synchronous control signal, perform adaptive control of the radio frequency power supply, track and monitor the regulated radio frequency power supply signal, perform control quality analysis on the regulated radio frequency power supply signal according to the expected control index, and determine the control deviation data; Record the parameter deviation data and the control deviation data, construct control correction data, and perform model optimization on the multivariate control channel according to the control correction data within a preset period; Among them, the construction of multiple regulatory channels includes: Taking the radio frequency power source attribute information as a constraint, querying the power source control log, determining a plurality of sample power source operating conditions and a plurality of sample spectrum parameter deviation sets, wherein the sample power source operating conditions and the sample spectrum parameter deviation sets have a corresponding relationship; Randomly selecting a first sample power supply condition, obtaining a first sample spectrum parameter deviation set for the first sample power supply condition, and constructing a first sample parameter space based on the first sample spectrum parameter deviation set; Randomly generating a plurality of first sample parameters based on the first sample parameter space, wherein the first sample parameters satisfy a predetermined uniformity constraint and a predetermined quantity constraint; For the purpose of satisfying a plurality of first sample parameters, optimizing the control parameters respectively based on the control evaluation function, determining a plurality of optimal control parameters, and constructing a first training set based on the plurality of first sample parameters and the plurality of optimal control parameters, wherein the control parameters include power stability, frequency stability, harmonic components, and power noise; Performing supervised training and cross-validation on the initial control channel using the first training set to obtain a first control channel that satisfies expected convergence constraints; Constructing the multivariate control channel based on the mapping association between the first sample power supply condition and the first control channel; Among them, constructing the regulation evaluation function includes: The expression of the regulation evaluation function is: ； in, Characterizes the fitness of the i-th control parameter, is the importance of the power stability indicator, is the power variance after the i-th control parameter is adjusted, is the index importance of frequency stability, is the frequency variance after the i-th control parameter is adjusted, is the harmonic distortion weight, is the harmonic content after the i-th control parameter is adjusted, is the power noise weight, is the power noise content after the i-th control parameter is adjusted; Wherein, obtaining the first regulation channel includes: Configuring a control analysis operator, and building a plurality of initial analysis units based on the control analysis operator, wherein the control analysis operator at least includes a BP neural network, a random decision forest, and a long short-term memory network; Divide the first training set into N equal parts, select N times without replacement to construct the first training set, and iterate and select N times to obtain N training sets, where N is the number of regulation and analysis operators; Using N training sets to perform supervised training and cross-validation on multiple initial parsing units, multiple convergent parsing units that meet expected convergence constraints are obtained; Determine multiple convergence coefficients of multiple convergence analysis units, and build an analysis output layer according to the multiple convergence coefficients, wherein the output of the analysis output layer is a weighted result of the multiple convergence analysis units; Building a control analysis layer in parallel based on multiple convergent analysis units, and generating the first control channel in combination with the analysis output layer; Among them, the model optimization of multiple regulatory channels includes: Acquire multiple control deviation data within a preset period, perform control parameter traceability analysis based on the multiple control deviation data, and determine multiple standard control parameters; Generate a control and correction data set according to multiple parameter deviation data and multiple standard control parameters, divide the control and correction data set according to working conditions, and determine multiple working condition correction data sets; Based on the principle of incremental learning, multiple working condition correction data sets are used to perform mapping optimization training on multiple control channels within the multivariate control channel.

2. The method for synchronously detecting radio frequency power supply signals based on real-time spectrum analysis according to claim 1, characterized in that: Determine the spectrum characteristic indicators, including: Constructing a spectrum feature space, wherein the spectrum feature space at least includes frequency deviation, bandwidth, harmonic components, phase noise, signal-to-noise ratio and sideband; Using the RF power source attribute information as a constraint, we conduct homology retrieval based on industrial big data to obtain a sample data set. Based on the sample data set, impact evaluation is performed on multiple characteristic indicators in the spectrum characteristic space and the operation quality of the radio frequency power supply in turn, and a characteristic indicator sequence is generated according to the evaluation results; A preset number of characteristic indicators are extracted from the characteristic indicator sequence and set as the spectrum characteristic indicators.

3. The method for synchronously detecting radio frequency power supply signals based on real-time spectrum analysis according to claim 2, characterized in that: Generate a sequence of characteristic indicators, including: Obtaining a power quality evaluation index set, wherein each power quality evaluation index includes an index importance; Randomly selecting a first characteristic indicator and a first power quality evaluation indicator, and based on a single variable analysis strategy, using the sample data set to evaluate the correlation between the first characteristic indicator and the first power quality evaluation indicator to generate a first indicator correlation coefficient; Iteratively analyzing and determining a plurality of first indicator correlation coefficients of the first characteristic indicator, and comprehensively evaluating the plurality of correlation coefficients according to the indicator importance to obtain a first correlation coefficient; A plurality of correlation coefficients of a plurality of first characteristic indicators are sequentially obtained, and the plurality of characteristic indicators are arranged from large to small according to the correlation coefficients to obtain the characteristic indicator sequence.

4. The method for synchronously detecting radio frequency power supply signals based on real-time spectrum analysis according to claim 1, characterized in that: Generate synchronous control signals, including: Collecting real-time operating condition data of the radio frequency power supply, inputting the real-time operating condition data into the multivariate control channel for similarity matching, and determining an associated control channel, wherein the similarity between the operating condition data of the associated control channel and the real-time operating condition data meets a predetermined similarity threshold; The associated control channel is used to perform control analysis on the parameter deviation data, and the synchronous control signal is output.

Images