Intelligent control method and system for underwater cavitation type net washing machine

By constructing a standard texture feature vector classification decision tree and a three-dimensional simulation model, combined with Bayesian dynamic timing and trend regression model, the problem of poor cleaning effect of underwater cavitation net washing machine is solved, and efficient and energy-saving pollutant removal is achieved .

CN118444625BActive Publication Date: 2024-11-01SOUTH CHINA SEA FISHERIES RES INST CHINESE ACAD OF FISHERY SCI +1
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202410846899.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-06-27
Publication Date
2024-11-01
Estimated Expiration
2044-06-27

AI Technical Summary

Technical Problem

The existing underwater cavitation net washing machines are difficult to achieve accurate analysis of attachments on the mesh, resulting in poor cleaning effect, many pollutant residues, and high cleaning rate and energy consumption, which fail to achieve rational cleaning.

Method used

By obtaining the actual texture feature vectors of the web to be cleaned, a standard texture feature vector classification decision tree is constructed, the adsorption characteristic index is analyzed, the range to be cleaned is planned, a three-dimensional simulation model is constructed for simulation simulation, and a Bayesian dynamic timing model and trend regression model are generated , regulate the underwater cavitation net washing machine to improve cleaning efficiency and quality.

Benefits of technology

Intelligent regulation of underwater cavitation network washing machines is realized, cleaning rate and quality is improved, energy consumption is reduced, and pollutant removal is ensured.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118444625B_ABST
    Figure CN118444625B_ABST
Patent Text Reader

Abstract

The present invention relates to the technical field of underwater net washing equipment, and in particular to an intelligent control method and system for an underwater cavitation net washing machine. A standard texture feature vector classification decision tree is constructed to identify texture feature vectors, so as to obtain the adsorption characteristic index of each attached pollutant on the net to be cleaned for the manufacturing material; the cleaning range is reasonably planned according to the coordinates of the first type of attached pollutants and the second type of attached pollutants; based on the three-dimensional simulation model of the underwater cavitation net washing machine, the difference analysis between the bubble volume simulation mean value and the ideal bubble volume value of each attached pollutant in all the areas to be cleaned is simulated and analyzed to control the underwater cavitation net washing machine; a Bayesian dynamic time series model is constructed, and the bubble density simulation value is calculated through the Bayesian dynamic time series model for control; the present invention can intelligently control the bubble cavitation effect of the underwater cavitation net washing machine, thereby improving the cleaning rate and quality, and making the underwater cavitation net washing machine more energy-efficient.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of underwater net washing equipment, and in particular to an intelligent control method and system for an underwater cavitation type net washing machine. Background Art

[0002] An underwater cavitation net washing machine is a device used to clean nets. Its working principle mainly relies on the impact force generated by the cavitation effect (i.e. the formation and rupture of bubbles) to remove dirt and various attachments, thereby completing the cleaning of marine aquaculture nets and achieving a cleaning effect. With the continuous advancement of science and technology, the current underwater cavitation net washing machines can intelligently control various parameters of bubbles to clean nets, but there are still some underwater cavitation net washing machines that cannot accurately analyze the types of attachments on the nets and the adsorption characteristics of the attachments on the nets, making it difficult for underwater cavitation net washing machines to generate bubbles of corresponding specifications and density distribution to remove pollutants, resulting in a large amount of pollutants remaining after cleaning, poor cleaning effect, and difficulty in achieving rational cleaning, which greatly reduces the cleaning rate and increases energy consumption output, which is not conducive to the far-reaching and sustainable development of the industry. Summary of the invention

[0003] The present invention overcomes the deficiencies of the prior art and provides an intelligent control method and system for an underwater cavitation type net washing machine.

[0004] To achieve the above object, the technical solution adopted by the present invention is:

[0005] The first aspect of the present invention provides an intelligent control method for an underwater cavitation type net washing machine, comprising the following steps:

[0006] Acquire the actual texture feature vectors of all attached pollutants on the net to be cleaned, and construct a standard texture feature vector classification decision tree, and identify the actual texture feature vectors in the standard texture feature vector classification decision tree to obtain the adsorption characteristic index of each attached pollutant for the manufacturing material;

[0007] Calculating information gain based on the adsorption characteristic index for analysis to determine the first type of attached pollutants and the second type of attached pollutants in the pattern distribution coordinate diagram, and calculating and planning based on the coordinates of the first type of attached pollutants and the second type of attached pollutants to obtain all the to-be-cleaned ranges of the underwater cavitation type net washing machine for the nets to be cleaned;

[0008] Constructing a three-dimensional simulation model of an underwater cavitation-type net washing machine, performing net washing simulation in the three-dimensional simulation model and obtaining the bubble volume simulation mean of each attached pollutant in all the areas to be cleaned, and regulating the underwater cavitation-type net washing machine through difference analysis between the bubble volume simulation mean and the ideal bubble volume value to obtain a first regulation scheme;

[0009] Based on the first control scheme, a secondary screen washing simulation is performed in a three-dimensional simulation model to obtain a number of cavitation effects generated when all attached pollutants are cleaned at each uniform time series node, a Bayesian dynamic time series model is constructed according to the several cavitation effects, a bubble density simulation value is deduced through the Bayesian dynamic time series model, the bubble density simulation value is analyzed, and a second control scheme is generated;

[0010] The water flow flux values ​​and cavitation effect migration trajectories passing through the underwater cavitation type net washing machine at different time nodes are obtained, and a trend regression model is constructed. Based on the water flow flux values ​​and cavitation effect migration trajectories, the final formation area of ​​the water flow migration cavitation high pressure zone is analyzed in the trend regression model. The underwater cavitation type net washing machine is regulated according to the final formation area to obtain a third regulation scheme.

[0011] Furthermore, in a preferred embodiment of the present invention, the actual texture feature vectors of all attached pollutants on the net to be cleaned are obtained, and a standard texture feature vector classification decision tree is constructed, and the adsorption characteristic index of each attached pollutant for the manufacturing material is obtained by identifying the actual texture feature vector in the standard texture feature vector classification decision tree, specifically including the following steps:

[0012] The nets to be cleaned currently fixed on the underwater cavitation net washing machine are photographed by an underwater camera to obtain image information of the nets to be cleaned;

[0013] The GLCM algorithm is introduced to perform grayscale quantization calculation on the image information of the net to be cleaned, and a grayscale co-occurrence matrix is ​​generated. The pixels of the grayscale co-occurrence matrix are traversed based on a preset direction and distance and normalized to obtain the actual texture feature vectors of all attached pollutants on the net to be cleaned;

[0014] Obtaining the actual fishing environment factor set where the net to be cleaned was last fished, searching in a big data network based on the actual fishing environment factor set to obtain the types of attached pollutants associated with the net to be cleaned, and obtaining a number of standard texture feature vectors of each attached pollutant type under the conditions of the actual fishing environment factor set;

[0015] A segmentation threshold is preset based on each type of attached pollutant, and N tree trunk nodes of each type of attached pollutant are generated. Each tree trunk node is segmented by a number of standard texture feature vectors to form M groups of standard texture feature vector branch nodes until the segmentation threshold is reached, thereby obtaining a standard texture feature vector classification decision tree;

[0016] The actual texture feature vectors of all attached pollutants on the net to be cleaned are defined as identification samples, and the identification samples are introduced into a standard texture feature vector classification decision tree for identification. If a leaf node at the end of a branch node can be identified, the identification is stopped, and the type information of each attached pollutant on the net to be cleaned is output;

[0017] The material of the net to be cleaned is obtained, and a comprehensive knowledge graph of fishery is obtained through a big data network. The type information and material information of each attached pollutant are imported into the comprehensive knowledge graph of fishery to obtain the adsorption characteristic index of each attached pollutant for the material of the net.

[0018] Furthermore, in a preferred embodiment of the present invention, the information gain is calculated based on the adsorption characteristic index for analysis to determine the first type of attached pollutants and the second type of attached pollutants in the pattern distribution coordinate diagram, and the coordinates of the first type of attached pollutants and the second type of attached pollutants are calculated and planned to obtain all the to-be-cleaned ranges of the underwater cavitation type net washing machine for the nets to be cleaned, which specifically includes the following steps:

[0019] Constructing a two-dimensional domain coordinate system, importing the image information of the net to be cleaned into the two-dimensional domain coordinate system for registration, generating a pattern distribution framework, and fitting the texture feature vectors of all the pollutants attached to the net to be cleaned in the pattern distribution framework to obtain a pattern distribution coordinate map of all the pollutants attached to the net to be cleaned;

[0020] Based on the trade-off algorithm, the characteristic entropy and conditional entropy of each adsorption characteristic index are calculated, and the information gain of each adsorption characteristic index is determined according to the characteristic entropy and conditional entropy. If the information gain is greater than a preset information gain, the attached pollutants greater than the preset information gain in the pattern distribution map are marked as Class I attached pollutants, and the remaining attached pollutants are marked as Class II attached pollutants;

[0021] Based on the pattern distribution coordinate map, the coordinate values ​​of two adjacent first-class attached pollutants are obtained, and the Euclidean distance between the coordinate values ​​of two adjacent first-class attached pollutants is calculated. If the Euclidean distance is greater than the preset Euclidean distance, the planning range threshold of each first-class attached pollutant is preset based on the coordinate value;

[0022] Taking each first-class attached pollutant as the center point, the surrounding second-class attached pollutants are planned to be included until the planning range threshold of each first-class attached pollutant is reached, and several to-be-cleaned ranges centered on each first-class attached pollutant are generated, which are defined as the first to-be-cleaned ranges;

[0023] If the Euclidean distance is less than the preset Euclidean distance, the adjacent first-class attached pollutants are used as the combination center point benchmark, and the surrounding second-class attached pollutants are included and planned based on the Euclidean distance until the preset Euclidean distance is met, and several adjacent first-class attached pollutants are used as the combination center to be cleaned ranges, which are defined as the second to-be-cleaned ranges;

[0024] A plurality of first to-be-cleaned areas and a plurality of second to-be-cleaned areas are combined to obtain all to-be-cleaned areas of the underwater cavitation type net washing machine for the nets to be cleaned.

[0025] Furthermore, in a preferred embodiment of the present invention, the three-dimensional simulation model of the underwater cavitation type net washing machine is constructed, the net washing simulation is performed in the three-dimensional simulation model and the bubble volume simulation mean of each attached pollutant in all the areas to be cleaned is obtained, and the underwater cavitation type net washing machine is regulated by the difference analysis between the bubble volume simulation mean and the ideal bubble volume value to obtain the first regulation scheme, which specifically includes the following steps:

[0026] The size parameters and corresponding adsorption characteristic indexes of all attached pollutants in each range to be cleaned are obtained, and are imported into the preset parameter control decision of the underwater cavitation type net washing machine for analysis, so as to generate the initial working parameters of the underwater cavitation type net washing machine in cleaning each range to be cleaned; wherein the initial working parameters include the initial vibration frequency, the initial amplitude and the initial water flow rate;

[0027] Constructing a three-dimensional simulation model of an underwater cavitation-type net washing machine, inputting the initial working parameters of the underwater cavitation-type net washing machine in each range to be cleaned into the three-dimensional simulation model to perform net washing simulation, so as to obtain the bubble volume simulation mean value for cleaning the first type of attached pollutants and the bubble volume simulation mean value for cleaning the second type of attached pollutants;

[0028] The Cardinal curve method is introduced, in which the quadratic polynomial coefficients between the size parameters of all attached pollutants and the corresponding adsorption characteristic index are calculated to obtain several quadratic polynomial coefficients, and several quadratic polynomial coefficients are fitted by spline interpolation to obtain several Cardinal curves;

[0029] A Cardinal surface is drawn based on the plurality of Cardinal curves, and an ideal bubble volume value required for cleaning the first type of attached pollutants and an ideal bubble volume value required for cleaning the second type of attached pollutants are obtained by aligning the Cardinal surface to the range to be cleaned for volume calculation;

[0030] Calculate the difference between the simulated mean value of the bubble volume for cleaning the first type of attached pollutants and the simulated mean value of the bubble volume for cleaning the second type of attached pollutants to obtain a simulated bubble difference; and simultaneously calculate the difference between the ideal bubble volume value required for cleaning the first type of attached pollutants and the ideal bubble volume value required for cleaning the second type of attached pollutants to obtain an ideal bubble difference;

[0031] A hash algorithm is introduced to calculate a hash function between the simulated bubble differences and the ideal bubble differences. If the hash function is greater than a preset hash function, the operating parameters of the underwater cavitation screen washing machine are regulated to obtain a first regulation scheme.

[0032] Furthermore, in a preferred embodiment of the present invention, the secondary screen washing simulation is performed in a three-dimensional simulation model based on the first control scheme to obtain a number of cavitation effects generated when all attached pollutants are cleaned at each uniform timing node, a Bayesian dynamic timing model is constructed according to the several cavitation effects, a bubble density simulation value is calculated through the Bayesian dynamic timing model, the bubble density simulation value is analyzed, and a second control scheme is generated, which specifically includes the following steps:

[0033] A number of uniform timing nodes are preset, and the first control scheme is executed on a number of uniform timing nodes by controlling the three-dimensional simulation model of the underwater cavitation type net washing machine to perform secondary net washing simulation, and the cavitation effect generated when cleaning all attached pollutants on each uniform timing node is recorded to obtain a number of cavitation effect coefficients;

[0034] The Dynamic Linear Model algorithm is introduced to create a state equation for cleaning all attached pollutants at each uniform timing node, and an observation equation is constructed based on several cavitation effect coefficients and the state equation. The next uniform timing node is predicted through the state equation to obtain the state equation and cavitation effect coefficient of the next uniform timing node.

[0035] According to the state equation and cavitation effect coefficient of the next uniform time series node and the observation equation, the corresponding cavitation effect coefficient on each uniform time series node is updated in a rolling manner until the states and covariances of several cavitation effect coefficients are updated to obtain a Bayesian dynamic time series model for secondary screen washing simulation;

[0036] A plurality of cavitation effect coefficients are introduced into the Bayesian dynamic time series model of the secondary screen washing simulation for Bayesian inference to obtain the bubble rate and bubble flux when cleaning all attached pollutants, and the bubble rate and bubble flux are combined for calculation to obtain the bubble density simulation value when cleaning all attached pollutants;

[0037] Determine whether the bubble density simulation value is in the ideal bubble density interval. If not, obtain the lower limit of the ideal bubble density interval, calculate the Mahalanobis distance between the bubble density simulation value and the lower limit of the interval, and adjust the working parameters of the underwater cavitation screen washing machine based on the Mahalanobis distance to obtain a second control scheme.

[0038] Furthermore, in a preferred embodiment of the present invention, the water flux value and the cavitation effect migration trajectory passing through the underwater cavitation type net washing machine at different time nodes are obtained, and a trend regression model is constructed. Based on the water flux value and the cavitation effect migration trajectory, the final formation area of ​​the water flow migration cavitation high pressure zone is analyzed in the trend regression model, and the underwater cavitation type net washing machine is regulated according to the final formation area to obtain a third regulation scheme, which specifically includes the following steps:

[0039] The underwater cavitation type net washing machine is regulated by the first regulation scheme or the second regulation scheme, and after the regulation, the water flow rate value and the cavitation effect migration trajectory passing through the underwater cavitation type net washing machine at different time nodes are obtained;

[0040] Based on the analysis of the water flux value and cavitation effect migration trajectory passing through the underwater cavitation screen washing machine at different time points, the water flux change trend and cavitation effect migration state are obtained;

[0041] The Spearson coefficient is introduced to calculate the correlation coefficient between the water flux change trend and the cavitation effect migration state, and the interference degree between the water flux change trend and the cavitation effect migration state is determined according to the correlation coefficient. If the interference degree is greater than the preset interference degree, a trend regression model is constructed based on the water flux change trend and the cavitation effect migration state.

[0042] The least square method is introduced to estimate the trend regression model to obtain the hysteresis term coefficient of the water flux change trend for the cavitation effect migration state in the trend regression model, and whether the hysteresis term coefficient is significantly 0 is tested to obtain the test result;

[0043] Generate a time series statistic of the water flux change trend for the cavitation effect migration state according to the test result, if the time series statistic is greater than a preset time series statistic, it proves that the lag term coefficient in the test result is significantly close to 0, and output the interference trend of the water flux change trend for the cavitation effect migration state according to the time series statistic;

[0044] Constructing a deep learning prediction model, making predictions in the deep learning prediction model according to the interference trend, obtaining the migration landing point of the cavitation high pressure zone of the underwater cavitation type screen washing machine, and analyzing the migration landing point of the cavitation high pressure zone to obtain the final formation area of ​​the cavitation high pressure zone;

[0045] The current range to be cleaned of the underwater cavitation type net washing machine is obtained, and the overlap rate between the current range to be cleaned and the final formation area of ​​the cavitation high-pressure zone is calculated. If the overlap rate is less than the preset overlap rate, the pressure parameters of the underwater cavitation type net washing machine are regulated based on the overlap rate to obtain a third regulation scheme.

[0046] A second aspect of the present invention provides an intelligent control system for an underwater cavitation type net washing machine, the intelligent control system for an underwater cavitation type net washing machine comprising a memory and a processor, the memory storing an intelligent control method program for an underwater cavitation type net washing machine, and when the intelligent control method program for an underwater cavitation type net washing machine is executed by the processor, the following steps are implemented:

[0047] Acquire the actual texture feature vectors of all attached pollutants on the net to be cleaned, and construct a standard texture feature vector classification decision tree, and identify the actual texture feature vectors in the standard texture feature vector classification decision tree to obtain the adsorption characteristic index of each attached pollutant for the manufacturing material;

[0048] Calculating information gain based on the adsorption characteristic index for analysis to determine the first type of attached pollutants and the second type of attached pollutants in the pattern distribution coordinate diagram, and calculating and planning based on the coordinates of the first type of attached pollutants and the second type of attached pollutants to obtain all the to-be-cleaned ranges of the underwater cavitation type net washing machine for the nets to be cleaned;

[0049] Constructing a three-dimensional simulation model of an underwater cavitation-type net washing machine, performing net washing simulation in the three-dimensional simulation model and obtaining the bubble volume simulation mean of each attached pollutant in all the areas to be cleaned, and regulating the underwater cavitation-type net washing machine through difference analysis between the bubble volume simulation mean and the ideal bubble volume value to obtain a first regulation scheme;

[0050] Based on the first control scheme, a secondary screen washing simulation is performed in a three-dimensional simulation model to obtain a number of cavitation effects generated when all attached pollutants are cleaned at each uniform time series node, a Bayesian dynamic time series model is constructed according to the several cavitation effects, a bubble density simulation value is deduced through the Bayesian dynamic time series model, the bubble density simulation value is analyzed, and a second control scheme is generated;

[0051] The water flow flux values ​​and cavitation effect migration trajectories passing through the underwater cavitation type net washing machine at different time nodes are obtained, and a trend regression model is constructed. Based on the water flow flux values ​​and cavitation effect migration trajectories, the final formation area of ​​the water flow migration cavitation high pressure zone is analyzed in the trend regression model. The underwater cavitation type net washing machine is regulated according to the final formation area to obtain a third regulation scheme.

[0052] The present invention solves the technical defects existing in the background technology, and the beneficial technical effects of the present invention are:

[0053] A standard texture feature vector classification decision tree is constructed to identify the texture feature vector to obtain the adsorption characteristic index of each attached pollutant on the net to be cleaned for the manufacturing material; based on the adsorption characteristic index, information gain is calculated for analysis to determine the first type of attached pollutants and the second type of attached pollutants, and calculation planning is performed according to the coordinates of the first type of attached pollutants and the second type of attached pollutants to obtain all the to-be-cleaned ranges of the underwater cavitation type net washing machine for the net to be cleaned; based on the three-dimensional simulation model of the underwater cavitation type net washing machine, the underwater cavitation type net washing machine is regulated by the difference analysis between the bubble volume simulation mean and the ideal bubble volume value of each attached pollutant in all the to-be-cleaned ranges to obtain a first regulation scheme; a Bayesian dynamic time series model is constructed, and the bubble density simulation value is deduced through the Bayesian dynamic time series model, and the bubble density simulation value is analyzed to generate a second regulation scheme; a trend regression model is constructed, and the final formation area of ​​the water flow migration cavitation high pressure zone is analyzed in the trend regression model based on the water flow flux value and the cavitation effect migration trajectory, and the underwater cavitation type net washing machine is regulated according to the final formation area to obtain a third regulation scheme. The present invention can intelligently regulate the bubble cavitation effect of the underwater cavitation type net washing machine, thereby improving the cleaning rate and quality, and making the underwater cavitation type net washing machine more energy-efficient. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in 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 described below are only some embodiments of the present invention. For ordinary technicians in this field, drawings of other embodiments can be obtained based on these drawings without paying creative work.

[0055] Figure 1 A first method flow chart of an intelligent control method for an underwater cavitation type net washing machine is shown;

[0056] Figure 2 A second method flow chart of an intelligent control method for an underwater cavitation type net washing machine is shown;

[0057] Figure 3 A third method flow chart of an intelligent control method for an underwater cavitation type net washing machine is shown;

[0058] Figure 4 The system framework diagram of an intelligent control system of an underwater cavitation type net washing machine is shown. DETAILED DESCRIPTION

[0059] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0060] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0061] The first aspect of the present invention provides an intelligent control method for an underwater cavitation type net washing machine, such as Figure 1 As shown, the following steps are included:

[0062] S102: Acquire actual texture feature vectors of all attached pollutants on the net to be cleaned, and construct a standard texture feature vector classification decision tree, and identify the actual texture feature vectors in the standard texture feature vector classification decision tree to obtain an adsorption characteristic index of each attached pollutant for the manufacturing material;

[0063] S104: Calculating information gain based on the adsorption characteristic index to perform analysis, so as to determine the first type of attached pollutants and the second type of attached pollutants in the pattern distribution coordinate diagram, and performing calculation planning according to the coordinates of the first type of attached pollutants and the second type of attached pollutants to obtain all the to-be-cleaned ranges of the underwater cavitation net washing machine for the nets to be cleaned;

[0064] S106: constructing a three-dimensional simulation model of an underwater cavitation-type net washing machine, performing net washing simulation in the three-dimensional simulation model and obtaining a bubble volume simulation mean value of each attached pollutant in all areas to be cleaned, and regulating the underwater cavitation-type net washing machine through a difference analysis between the bubble volume simulation mean value and an ideal bubble volume value to obtain a first regulation scheme;

[0065] S108: performing secondary screen washing simulation in a three-dimensional simulation model based on the first control scheme to obtain a number of cavitation effects generated when all attached pollutants are cleaned at each uniform time series node, constructing a Bayesian dynamic time series model according to the several cavitation effects, calculating a bubble density simulation value through the Bayesian dynamic time series model, analyzing the bubble density simulation value, and generating a second control scheme;

[0066] S110: Obtain the water flow flux values ​​and cavitation effect migration trajectories passing through the underwater cavitation type net washing machine at different time nodes, and construct a trend regression model. Based on the water flow flux values ​​and cavitation effect migration trajectories, the final formation area of ​​the water flow migration cavitation high pressure zone is analyzed in the trend regression model. The underwater cavitation type net washing machine is controlled according to the final formation area to obtain a third control scheme.

[0067] Furthermore, in a preferred embodiment of the present invention, the actual texture feature vectors of all attached pollutants on the net to be cleaned are obtained, and a standard texture feature vector classification decision tree is constructed, and the adsorption characteristic index of each attached pollutant for the manufacturing material is obtained by identifying the actual texture feature vector in the standard texture feature vector classification decision tree, specifically including the following steps:

[0068] The nets to be cleaned currently fixed on the underwater cavitation net washing machine are photographed by an underwater camera to obtain image information of the nets to be cleaned;

[0069] The GLCM algorithm is introduced to perform grayscale quantization calculation on the image information of the net to be cleaned, and a grayscale co-occurrence matrix is ​​generated. The pixels of the grayscale co-occurrence matrix are traversed based on a preset direction and distance and normalized to obtain the actual texture feature vectors of all attached pollutants on the net to be cleaned;

[0070] Obtaining the actual fishing environment factor set where the net to be cleaned was last fished, searching in a big data network based on the actual fishing environment factor set to obtain the types of attached pollutants associated with the net to be cleaned, and obtaining a number of standard texture feature vectors of each attached pollutant type under the conditions of the actual fishing environment factor set;

[0071] A segmentation threshold is preset based on each type of attached pollutant, and N tree trunk nodes of each type of attached pollutant are generated. Each tree trunk node is segmented by a number of standard texture feature vectors to form M groups of standard texture feature vector branch nodes until the segmentation threshold is reached, thereby obtaining a standard texture feature vector classification decision tree;

[0072] The actual texture feature vectors of all attached pollutants on the net to be cleaned are defined as identification samples, and the identification samples are introduced into a standard texture feature vector classification decision tree for identification. If a leaf node at the end of a branch node can be identified, the identification is stopped, and the type information of each attached pollutant on the net to be cleaned is output;

[0073] The material of the net to be cleaned is obtained, and a comprehensive knowledge graph of fishery is obtained through a big data network. The type information and material information of each attached pollutant are imported into the comprehensive knowledge graph of fishery to obtain the adsorption characteristic index of each attached pollutant for the material of the net.

[0074] It should be noted that underwater cavitation net washing machines can be used in the aquatic field, that is, to clean various nets of different purposes and specifications used in fishery fishing. Due to the complex fishing environment in which fishing nets are located during the fishing process, the nets may adsorb and carry different impurity residues in different fishing environments, such as fish scales, algae debris, body tissues of marine organisms, etc. These impurity residues can be cleaned and removed by controlling the underwater cavitation net washing machine to produce bubbles with a certain cavitation effect. However, different impurity residues have a certain adhesion or adhesion to nets of different materials. Therefore, in order to be able to clean and remove impurities on the nets more accurately and quickly, it is necessary to determine the attached pollutants remaining on the nets to be cleaned in order to further analyze the adsorption characteristics of the attached pollutants on the nets, and control the underwater cavitation net washing machine to produce bubbles with corresponding cavitation effects according to the adsorption characteristics to clean them. This method obtains the standard texture feature vectors of all possible attached pollutant species on the nets to be cleaned through the actual fishing environment factor set where the nets to be cleaned were last caught, and uses the standard texture feature vectors of all possible attached pollutant species to construct a texture feature recognition decision tree. The texture feature recognition decision tree can more quickly and accurately locate the attached pollutant species corresponding to the actual texture feature vector, so as to greatly improve the rate of obtaining and determining the pollutant species and save the traditional complex recognition operation steps; finally, the adsorption characteristic index of each attached pollutant species for the material of the nets to be cleaned is analyzed in the comprehensive knowledge graph of fishery fishing, which improves the targeted control ability of the underwater cavitation net washing machine, ensures that each attached pollutant can be efficiently cleaned, and provides a reliable solution for the subsequent underwater cavitation net washing machine control. Among them, the Chinese name of the GLCM algorithm is the gray-level co-occurrence matrix algorithm, which is an extraction method for describing image texture features, effectively improving the efficiency of image texture description and enhancing the expression accuracy of the image.

[0075] Furthermore, in a preferred embodiment of the present invention, the information gain is calculated based on the adsorption characteristic index for analysis to determine the first type of attached pollutants and the second type of attached pollutants in the pattern distribution coordinate diagram, and the coordinates of the first type of attached pollutants and the second type of attached pollutants are calculated and planned to obtain all the to-be-cleaned ranges of the underwater cavitation type net washing machine for the nets to be cleaned, which specifically includes the following steps:

[0076] Constructing a two-dimensional domain coordinate system, importing the image information of the net to be cleaned into the two-dimensional domain coordinate system for registration, generating a pattern distribution framework, and fitting the texture feature vectors of all the pollutants attached to the net to be cleaned in the pattern distribution framework to obtain a pattern distribution coordinate map of all the pollutants attached to the net to be cleaned;

[0077] Based on the trade-off algorithm, the characteristic entropy and conditional entropy of each adsorption characteristic index are calculated, and the information gain of each adsorption characteristic index is determined according to the characteristic entropy and conditional entropy. If the information gain is greater than a preset information gain, the attached pollutants greater than the preset information gain in the pattern distribution map are marked as Class I attached pollutants, and the remaining attached pollutants are marked as Class II attached pollutants;

[0078] Based on the pattern distribution coordinate map, the coordinate values ​​of two adjacent first-class attached pollutants are obtained, and the Euclidean distance between the coordinate values ​​of two adjacent first-class attached pollutants is calculated. If the Euclidean distance is greater than the preset Euclidean distance, the planning range threshold of each first-class attached pollutant is preset based on the coordinate value;

[0079] Taking each first-class attached pollutant as the center point, the surrounding second-class attached pollutants are planned to be included until the planning range threshold of each first-class attached pollutant is reached, and several to-be-cleaned ranges centered on each first-class attached pollutant are generated, which are defined as the first to-be-cleaned ranges;

[0080] If the Euclidean distance is less than the preset Euclidean distance, the adjacent first-class attached pollutants are used as the combination center point benchmark, and the surrounding second-class attached pollutants are included and planned based on the Euclidean distance until the preset Euclidean distance is met, and several adjacent first-class attached pollutants are used as the combination center to be cleaned ranges, which are defined as the second to-be-cleaned ranges;

[0081] A plurality of first to-be-cleaned areas and a plurality of second to-be-cleaned areas are combined to obtain all to-be-cleaned areas of the underwater cavitation type net washing machine for the nets to be cleaned.

[0082] It should be noted that since the attached pollutants are usually unevenly distributed on the fishing nets, and different attached pollutants also have different adsorption characteristic indexes for the nets, the combination of the two will require the underwater cavitation net washing machine to run to each attached pollutant location in turn and control the corresponding cavitation effect bubble parameters to clean the attached pollutants one by one. This process will increase the cleaning workload of the underwater cavitation net washing machine and the cleaning efficiency is low, which greatly increases the cleaning energy consumption output of the underwater cavitation net washing machine, which is not conducive to energy conservation and environmental protection development. Therefore, this method can calculate and analyze the information gain of the corresponding adsorption characteristic index of each attached pollutant to the net. If the information gain is greater than the preset information gain, it means that the attached pollutant here has a large adsorption viscosity to the net, and needs to be removed by moderate cavitation effect bubbles that can produce greater impact force and can pass through the mesh. It is marked as a type of attached pollutant. Otherwise, it means that the attached pollutant has a small adsorption viscosity to the net and can be removed by dense and tiny cavitation effect bubbles. Compared with the unified bubble specification cleaning control decision of the traditional underwater cavitation net washing machine, this process greatly improves the rationality and standardization of the underwater cavitation net washing machine to clean the net, and different cavitation effects are formulated according to different pollutant objects. The effect bubble cleaning scheme ensures the cleaning intensity and removal rate of different pollutant objects and improves the cleaning effect. In addition, this method takes the first type of attached pollutants as the center to include a certain number of the second type of attached pollutants to form different cleaning ranges, so as to adjust the cleaning mode of the underwater cavitation net washing machine and realize the underwater cavitation net washing machine to simultaneously generate bubbles of uniform size to respectively clean the first type of attached pollutants and the second type of attached pollutants in the cleaning range. Compared with the traditional steps of cleaning the attached pollutants one by one, this method is more efficient and improves the cleaning efficiency. At the same time, it reduces the workload and energy consumption output of the underwater cavitation net washing machine, making the underwater cavitation net washing machine more energy-saving and economically efficient.

[0083] Furthermore, in a preferred embodiment of the present invention, a three-dimensional simulation model of an underwater cavitation type net washing machine is constructed, a net washing simulation is performed in the three-dimensional simulation model and the bubble volume simulation mean of each attached pollutant in all the areas to be cleaned is obtained, and the underwater cavitation type net washing machine is regulated by difference analysis between the bubble volume simulation mean and the ideal bubble volume value to obtain a first regulation scheme, such as Figure 2 As shown, the specific steps include:

[0084] S202: Obtaining the size parameters and corresponding adsorption characteristic indexes of all attached pollutants in each range to be cleaned, and importing them into the preset parameter control decision of the underwater cavitation type net washing machine for analysis, and generating the initial working parameters of the underwater cavitation type net washing machine for cleaning each range to be cleaned; wherein the initial working parameters include the initial vibration frequency, the initial amplitude and the initial water flow rate;

[0085] S204: constructing a three-dimensional simulation model of an underwater cavitation-type net washing machine, inputting initial working parameters of the underwater cavitation-type net washing machine in each range to be cleaned into the three-dimensional simulation model to perform net washing simulation, so as to obtain a bubble volume simulation mean value for cleaning the first type of attached pollutants and a bubble volume simulation mean value for cleaning the second type of attached pollutants;

[0086] S206: introducing the Cardinal curve method, calculating the quadratic polynomial coefficients between the size parameters of all attached pollutants and the corresponding adsorption characteristic indexes in the Cardinal curve method, obtaining a number of quadratic polynomial coefficients, and performing spline interpolation fitting on the several quadratic multinomial coefficients to obtain a number of Cardinal curves;

[0087] S208: Drawing a Cardinal surface based on the plurality of Cardinal curves, and performing volume calculation by aligning the Cardinal surface to the range to be cleaned to obtain an ideal bubble volume value required for cleaning the first type of attached pollutants and an ideal bubble volume value required for cleaning the second type of attached pollutants;

[0088] S210: Calculate the difference between the simulated mean value of the bubble volume for cleaning the first type of attached pollutants and the simulated mean value of the bubble volume for cleaning the second type of attached pollutants to obtain a simulated bubble difference; and simultaneously calculate the difference between the ideal bubble volume value required for cleaning the first type of attached pollutants and the ideal bubble volume value required for cleaning the second type of attached pollutants to obtain an ideal bubble difference;

[0089] S212: introducing a hash algorithm to calculate a hash function between the simulated bubble differences and the ideal bubble differences; if the hash function is greater than a preset hash function, regulating the operating parameters of the underwater cavitation screen washing machine to obtain a first regulation scheme.

[0090] It should be noted that in order to generate precise cavitation effect bubbles to efficiently clean attached pollutants, in addition to analyzing the adsorption characteristic index of attached pollutants on the net, it is also necessary to obtain the size of attached pollutants, wherein the size parameters include volume, cross-sectional area and shape; combining the adsorption characteristic index and size parameters, it is possible to know the bubble specifications required to separate the attached pollutants from the net, so that the specific working parameters of the underwater cavitation net washing machine can be controlled according to the required bubble specifications. This method uses a three-dimensional simulation model to simulate the net washing of each range to be cleaned on the net, and analyzes the deviation between the bubble volume generated under the current working parameters of the underwater cavitation net washing machine and the ideal bubble volume, replacing the traditional manual observation and measurement of bubble size changes and manual recording steps, with fast data acquisition and analysis speed, saving time and effort, and avoiding manual measurement errors, and high data reliability. Among them, the Cardinal curve method is a curve drawing method based on spline interpolation, which can generate a smooth curve through given control points and realize the fitting drawing of the smooth curve shape. The Cardinal curve method used in this method can draw a Cardinal surface that reflects the cavitation effect required for attached pollutants. According to the distribution of this Cardinal surface in the range to be cleaned, the ideal bubble volume value can be calculated. The calculation is relatively efficient and there are no redundant calculation steps. It has a certain data analysis smoothness, making the ideal bubble volume value more reliable. Among them, the quadratic polynomial coefficients can reflect the cavitation effect coefficients required between the size parameters of all attached pollutants and the corresponding adsorption characteristic index, providing an analytical basis for the subsequent formulation of the control plan of the underwater cavitation type net washing machine. This method can realize the precise control of the cavitation effect bubble specifications required for the underwater cavitation type net washing machine to clean different attached pollutants, so as to further improve the cleaning performance of the underwater cavitation type net washing machine.

[0091] Furthermore, in a preferred embodiment of the present invention, a secondary screen washing simulation is performed in a three-dimensional simulation model based on the first control scheme to obtain a number of cavitation effects generated when all attached pollutants are cleaned at each uniform timing node, a Bayesian dynamic timing model is constructed according to the several cavitation effects, a bubble density simulation value is calculated by the Bayesian dynamic timing model, the bubble density simulation value is analyzed, and a second control scheme is generated, such as Figure 3 As shown, the specific steps include:

[0092] S302: Preset a number of uniform time sequence nodes, execute the first control scheme on the number of uniform time sequence nodes by controlling the three-dimensional simulation model of the underwater cavitation type net washing machine to perform secondary net washing simulation, and record the cavitation effect generated when cleaning all attached pollutants on each uniform time sequence node to obtain a number of cavitation effect coefficients;

[0093] S304: Introducing the Dynamic Linear Model algorithm to create a state equation for cleaning all attached pollutants at each uniform time sequence node, and constructing an observation equation based on a number of cavitation effect coefficients and the state equation, and predicting the next uniform time sequence node through the state equation to obtain the state equation and cavitation effect coefficient of the next uniform time sequence node;

[0094] S306: rolling update of the corresponding cavitation effect coefficient on each uniform time series node according to the state equation and cavitation effect coefficient of the next uniform time series node and the observation equation, until the states and covariances of several cavitation effect coefficients are updated, and a Bayesian dynamic time series model of secondary screen washing simulation is obtained;

[0095] S308: importing several cavitation effect coefficients into the Bayesian dynamic time series model of the secondary screen washing simulation to perform Bayesian inference, obtaining the bubble rate and bubble flux when cleaning all attached pollutants, and calculating in combination with the bubble rate and bubble flux to obtain the bubble density simulation value when cleaning all attached pollutants;

[0096] S310: Determine whether the bubble density simulation value is in the ideal bubble density interval. If not, obtain the lower limit of the ideal bubble density interval, calculate the Mahalanobis distance between the bubble density simulation value and the lower limit of the interval, and adjust the working parameters of the underwater cavitation screen washing machine based on the Mahalanobis distance to obtain a second control scheme.

[0097] It should be noted that when cleaning different attached pollutants, in addition to ensuring the cavitation effect bubble specifications of the corresponding impact force, it is also necessary to make the bubbles form a certain density circle. The high-density cavitation effect bubbles can increase the frequency of attached pollutants and mesh penetration, so that the attached pollutants can be separated from the net more quickly under the continuous impact of the bubbles, thereby achieving the purpose of improving the cleaning effect of the net. However, after the underwater cavitation net washing machine is adjusted by the first control scheme, although the bubble specifications meet the requirements, the distribution density of the bubbles may not be able to achieve a high-density bubble phenomenon under the parameters of the first control scheme. Therefore, this method constructs a Bayesian dynamic timing model for the underwater cavitation net washing machine after executing the first control scheme to perform a secondary net washing simulation. The Bayesian dynamic timing model can quickly and accurately calculate and analyze the bubble rate and bubble flux corresponding to the cavitation effect coefficient generated at different uniform timing nodes when cleaning attached pollutants, so as to calculate the bubble density distribution of the underwater cavitation net washing machine for all attached pollutants during the secondary net washing simulation according to the bubble rate and bubble flux. If the bubble density simulation value is not in the ideal bubble density range, it means that the bubble density for cleaning attached pollutants cannot achieve high-frequency bubble impact and mesh penetration effect. Therefore, the working parameters of the underwater cavitation net washing machine can be adjusted according to the error of the bubble density simulation value, so that the underwater cavitation net washing machine can create an efficient and accurate bubble density distribution effect to improve the net cleaning quality with high reliability. Among them, the DynamicLinear Model algorithm is a dynamic linear model algorithm. The DynamicLinear Model algorithm can model the system state and observation data as a linear relationship for prediction and analysis, and perform inference and prediction based on the Bayesian framework, allowing a natural combination of prior information and observation data to provide more flexible and reliable prediction results. At the same time, it can effectively deal with noise and uncertainty in observation data, and improve the robustness and reliability of the model.

[0098] Furthermore, in a preferred embodiment of the present invention, the water flux value and the cavitation effect migration trajectory passing through the underwater cavitation type net washing machine at different time nodes are obtained, and a trend regression model is constructed. Based on the water flux value and the cavitation effect migration trajectory, the final formation area of ​​the water flow migration cavitation high pressure zone is analyzed in the trend regression model, and the underwater cavitation type net washing machine is regulated according to the final formation area to obtain a third regulation scheme, which specifically includes the following steps:

[0099] The underwater cavitation type net washing machine is regulated by the first regulation scheme or the second regulation scheme, and after the regulation, the water flow rate value and the cavitation effect migration trajectory passing through the underwater cavitation type net washing machine at different time nodes are obtained;

[0100] Based on the analysis of the water flux value and cavitation effect migration trajectory passing through the underwater cavitation screen washing machine at different time points, the water flux change trend and cavitation effect migration state are obtained;

[0101] The Spearson coefficient is introduced to calculate the correlation coefficient between the water flux change trend and the cavitation effect migration state, and the interference degree between the water flux change trend and the cavitation effect migration state is determined according to the correlation coefficient. If the interference degree is greater than the preset interference degree, a trend regression model is constructed based on the water flux change trend and the cavitation effect migration state.

[0102] The least square method is introduced to estimate the trend regression model to obtain the hysteresis term coefficient of the water flux change trend for the cavitation effect migration state in the trend regression model, and whether the hysteresis term coefficient is significantly 0 is tested to obtain the test result;

[0103] Generate a time series statistic of the water flux change trend for the cavitation effect migration state according to the test result, if the time series statistic is greater than a preset time series statistic, it proves that the lag term coefficient in the test result is significantly close to 0, and output the interference trend of the water flux change trend for the cavitation effect migration state according to the time series statistic;

[0104] Constructing a deep learning prediction model, making predictions in the deep learning prediction model according to the interference trend, obtaining the migration landing point of the cavitation high pressure zone of the underwater cavitation type screen washing machine, and analyzing the migration landing point of the cavitation high pressure zone to obtain the final formation area of ​​the cavitation high pressure zone;

[0105] The current range to be cleaned of the underwater cavitation type net washing machine is obtained, and the overlap rate between the current range to be cleaned and the final formation area of ​​the cavitation high-pressure zone is calculated. If the overlap rate is less than the preset overlap rate, the pressure parameters of the underwater cavitation type net washing machine are regulated based on the overlap rate to obtain a third regulation scheme.

[0106] It should be noted that most underwater cavitation net washing machines can usually generate a high-pressure area to make the bubbles gradually increase and burst to generate impact force. Therefore, if the net is placed in the high-pressure area for cleaning, it can better improve the removal quality of attached pollutants and the cleaning effect; but in the process of the underwater cavitation net washing machine making bubbles to clean the net, the flow of water will affect the movement of the bubbles, thereby causing the state of the cavitation effect to migrate, causing the high-pressure area created by the underwater cavitation net washing machine to migrate synchronously, reducing the formation point and range of the cavitation high-pressure area, which is not conducive to accurately positioning the net in the cavitation high-pressure area. Therefore, this method first constructs a trend regression model based on the water flow change trend and the cavitation effect migration state of the underwater cavitation net washing machine at different time nodes. The trend regression model can perform a calculation test on whether the time series between the water flow change trend and the cavitation effect migration state is stable, and accurately analyze the autocorrelation in the time series data, thereby further expressing the influence relationship, trend and continuity of the water flow change trend on the cavitation effect migration state. Among them, the purpose of checking whether the lag term coefficient is significantly 0 is to analyze whether the time series between the water flow rate change trend and the cavitation effect migration state is smooth. If the time series statistic is greater than the preset time series statistic, it proves that the lag term coefficient in the test result is significantly close to 0, and the time series data at this time is smooth, otherwise the time series is not smooth. This method can analyze the final formation area of ​​the cavitation high pressure zone of the underwater cavitation type net washing machine by analyzing the influence and interference trend between the water flow rate change trend and the cavitation effect migration state, so that the attached pollutants within the range to be cleaned can be always in the cavitation high pressure zone by adjusting the pressure parameters of the underwater cavitation type net washing machine, thereby improving the cleaning quality and effect of the underwater cavitation type net washing machine, avoiding the phenomenon of residual pollutants after cleaning, and reducing the cost output of cleaning rework.

[0107] In addition, the intelligent control method of the underwater cavitation type net washing machine further includes the following steps:

[0108] The cleaning control frequency of the underwater cavitation net washing machine is obtained, and the cleanliness of the net to be cleaned is obtained. When the cleaning control frequency is greater than the preset cleaning control frequency and the cleanliness is less than the preset cleanliness, the data sources corresponding to the cleaning control frequencies and the data source corresponding to the cleanliness are obtained and bound to generate a cleaning abnormality data source;

[0109] Obtain all cleaning mechanisms of the underwater cavitation-type net washing machine, and obtain the maximum control characteristics of each cleaning mechanism, introduce the grey correlation method to calculate the cleaning abnormal data source and the maximum control characteristics of each cleaning mechanism, obtain multiple grey correlations, and determine the abnormal correlation of each cleaning mechanism causing the cleaning abnormal data source according to the grey correlation;

[0110] Extract all cleaning mechanisms corresponding to the maximum control characteristics whose abnormal correlation degree is greater than the preset abnormal correlation degree, and define them as suspected cleaning fault mechanisms;

[0111] Obtain the actual control value of each suspected cleaning control fault mechanism within a preset time period, introduce the Monte Carlo Markov chain to estimate the actual control value of each suspected cleaning fault mechanism within the preset time period, and obtain the probability of abnormal control of each suspected cleaning fault mechanism;

[0112] If the probability of control abnormality in each suspected control fault mechanism is greater than the preset control abnormality probability, the suspected cleaning fault mechanism corresponding to the probability greater than the preset control abnormality probability is marked as a faulty mechanism and uploaded to the underwater cavitation screen washing machine terminal.

[0113] It should be noted that as the running time of the underwater cavitation net washing machine increases and passes, the internal cleaning mechanism will rust, age, wear or break, which will greatly reduce the cleaning quality and speed of the underwater cavitation net washing machine when cleaning the nets, resulting in poor cleaning effect. In addition, the maintenance of traditional equipment requires the staff to manually dismantle the machine to locate the fault position, which is slow, time-consuming and labor-intensive, and is not conducive to the high-quality operation and maintenance of the underwater cavitation net washing machine. This method can efficiently and accurately analyze and locate whether there is a fault in the cleaning mechanism of the underwater cavitation net washing machine, thereby providing accurate fault repair location information for the staff, ensuring the maintenance efficiency of the underwater cavitation net washing machine, and further improving the service life of the underwater cavitation net washing machine.

[0114] The second aspect of the present invention provides an intelligent control system for an underwater cavitation type net washing machine, the intelligent control system for an underwater cavitation type net washing machine comprises a memory 41 and a processor 42, the memory 41 stores an intelligent control method program for an underwater cavitation type net washing machine, and when the intelligent control method program for an underwater cavitation type net washing machine is executed by the processor 42, Figure 4 As shown, implement the following steps:

[0115] Acquire the actual texture feature vectors of all attached pollutants on the net to be cleaned, and construct a standard texture feature vector classification decision tree, and identify the actual texture feature vectors in the standard texture feature vector classification decision tree to obtain the adsorption characteristic index of each attached pollutant for the manufacturing material;

[0116] Calculating information gain based on the adsorption characteristic index for analysis to determine the first type of attached pollutants and the second type of attached pollutants in the pattern distribution coordinate diagram, and calculating and planning based on the coordinates of the first type of attached pollutants and the second type of attached pollutants to obtain all the to-be-cleaned ranges of the underwater cavitation type net washing machine for the nets to be cleaned;

[0117] Constructing a three-dimensional simulation model of an underwater cavitation-type net washing machine, performing net washing simulation in the three-dimensional simulation model and obtaining the bubble volume simulation mean of each attached pollutant in all the areas to be cleaned, and regulating the underwater cavitation-type net washing machine through difference analysis between the bubble volume simulation mean and the ideal bubble volume value to obtain a first regulation scheme;

[0118] Based on the first control scheme, a secondary screen washing simulation is performed in a three-dimensional simulation model to obtain a number of cavitation effects generated when all attached pollutants are cleaned at each uniform time series node, a Bayesian dynamic time series model is constructed according to the several cavitation effects, a bubble density simulation value is deduced through the Bayesian dynamic time series model, the bubble density simulation value is analyzed, and a second control scheme is generated;

[0119] The water flow flux values ​​and cavitation effect migration trajectories passing through the underwater cavitation type net washing machine at different time nodes are obtained, and a trend regression model is constructed. Based on the water flow flux values ​​and cavitation effect migration trajectories, the final formation area of ​​the water flow migration cavitation high pressure zone is analyzed in the trend regression model. The underwater cavitation type net washing machine is regulated according to the final formation area to obtain a third regulation scheme.

[0120] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art who is familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. An intelligent control method for an underwater cavitation screen washing machine. It is characterized in that The following steps are involved: Acquire the actual texture feature vectors of all attached pollutants on the net to be cleaned, and construct a standard texture feature vector classification decision tree, and identify the actual texture feature vectors in the standard texture feature vector classification decision tree to obtain the adsorption characteristic index of each attached pollutant for the manufacturing material; Calculating information gain based on the adsorption characteristic index for analysis to determine the first type of attached pollutants and the second type of attached pollutants in the pattern distribution coordinate diagram, and calculating and planning based on the coordinates of the first type of attached pollutants and the second type of attached pollutants to obtain all the to-be-cleaned ranges of the underwater cavitation type net washing machine for the nets to be cleaned; Constructing a three-dimensional simulation model of an underwater cavitation-type net washing machine, performing net washing simulation in the three-dimensional simulation model and obtaining the bubble volume simulation mean of each attached pollutant in all the areas to be cleaned, and regulating the underwater cavitation-type net washing machine through difference analysis between the bubble volume simulation mean and the ideal bubble volume value to obtain a first regulation scheme; Based on the first control scheme, a secondary screen washing simulation is performed in a three-dimensional simulation model to obtain a number of cavitation effects generated when all attached pollutants are cleaned at each uniform time series node, a Bayesian dynamic time series model is constructed according to the several cavitation effects, a bubble density simulation value is deduced through the Bayesian dynamic time series model, the bubble density simulation value is analyzed, and a second control scheme is generated; Obtain the water flow flux values ​​and cavitation effect migration trajectories passing through the underwater cavitation type net washing machine at different time nodes, and construct a trend regression model. Based on the water flow flux values ​​and cavitation effect migration trajectories, the final formation area of ​​the water flow migration cavitation high pressure zone is analyzed in the trend regression model. The underwater cavitation type net washing machine is regulated according to the final formation area to obtain a third regulation scheme. The information gain is calculated based on the adsorption characteristic index for analysis to determine the first type of attached pollutants and the second type of attached pollutants in the pattern distribution coordinate diagram, and the coordinates of the first type of attached pollutants and the second type of attached pollutants are calculated and planned to obtain all the cleaning ranges of the underwater cavitation type net washing machine for the nets to be cleaned, which specifically includes the following steps: Constructing a two-dimensional domain coordinate system, importing the image information of the net to be cleaned into the two-dimensional domain coordinate system for registration, generating a pattern distribution framework, and fitting the texture feature vectors of all the pollutants attached to the net to be cleaned in the pattern distribution framework to obtain a pattern distribution coordinate map of all the pollutants attached to the net to be cleaned; Based on the trade-off algorithm, the characteristic entropy and conditional entropy of each adsorption characteristic index are calculated, and the information gain of each adsorption characteristic index is determined according to the characteristic entropy and conditional entropy. If the information gain is greater than a preset information gain, the attached pollutants greater than the preset information gain in the pattern distribution map are marked as Class I attached pollutants, and the remaining attached pollutants are marked as Class II attached pollutants; Based on the pattern distribution coordinate map, the coordinate values ​​of two adjacent first-class attached pollutants are obtained, and the Euclidean distance between the coordinate values ​​of two adjacent first-class attached pollutants is calculated. If the Euclidean distance is greater than the preset Euclidean distance, the planning range threshold of each first-class attached pollutant is preset based on the coordinate value; Taking each first-class attached pollutant as the center point, the surrounding second-class attached pollutants are planned to be included until the planning range threshold of each first-class attached pollutant is reached, and several to-be-cleaned ranges centered on each first-class attached pollutant are generated, which are defined as the first to-be-cleaned ranges; If the Euclidean distance is less than the preset Euclidean distance, the adjacent first-class attached pollutants are used as the combination center point benchmark, and the surrounding second-class attached pollutants are included and planned based on the Euclidean distance until the preset Euclidean distance is met, and several adjacent first-class attached pollutants are used as the combination center to be cleaned ranges, which are defined as the second to-be-cleaned ranges; Combining a plurality of first to-be-cleaned areas and a plurality of second to-be-cleaned areas to obtain all to-be-cleaned areas of the underwater cavitation-type net washing machine for the nets to be cleaned; Among them, the second screen washing simulation is performed in the three-dimensional simulation model based on the first control scheme to obtain a number of cavitation effects generated when all attached pollutants are cleaned at each uniform time series node, a Bayesian dynamic time series model is constructed according to the several cavitation effects, a bubble density simulation value is calculated through the Bayesian dynamic time series model, the bubble density simulation value is analyzed, and a second control scheme is generated, which specifically includes the following steps: A number of uniform timing nodes are preset, and the first control scheme is executed on a number of uniform timing nodes by controlling the three-dimensional simulation model of the underwater cavitation type net washing machine to perform secondary net washing simulation, and the cavitation effect generated when cleaning all attached pollutants on each uniform timing node is recorded to obtain a number of cavitation effect coefficients; The Dynamic Linear Model algorithm is introduced to create a state equation for cleaning all attached pollutants at each uniform timing node, and an observation equation is constructed based on several cavitation effect coefficients and the state equation. The next uniform timing node is predicted through the state equation to obtain the state equation and cavitation effect coefficient of the next uniform timing node. According to the state equation and cavitation effect coefficient of the next uniform time series node and the observation equation, the corresponding cavitation effect coefficient on each uniform time series node is updated in a rolling manner until the states and covariances of several cavitation effect coefficients are updated to obtain a Bayesian dynamic time series model for secondary screen washing simulation; A plurality of cavitation effect coefficients are introduced into the Bayesian dynamic time series model of the secondary screen washing simulation for Bayesian inference to obtain the bubble rate and bubble flux when cleaning all attached pollutants, and the bubble rate and bubble flux are combined for calculation to obtain the bubble density simulation value when cleaning all attached pollutants; Determine whether the bubble density simulation value is in the ideal bubble density interval, if not, obtain the interval lower limit value of the ideal bubble density interval, calculate the Mahalanobis distance between the bubble density simulation value and the interval lower limit value, and adjust the working parameters of the underwater cavitation screen washing machine based on the Mahalanobis distance to obtain a second control scheme; The method of obtaining the water flow flux value and the cavitation effect migration trajectory passing through the underwater cavitation type net washing machine at different time nodes, and constructing a trend regression model, analyzing the final formation area of ​​the water flow migration cavitation high pressure zone in the trend regression model based on the water flow flux value and the cavitation effect migration trajectory, and regulating the underwater cavitation type net washing machine according to the final formation area to obtain the third regulation scheme specifically includes the following steps: The underwater cavitation type net washing machine is regulated by the first regulation scheme or the second regulation scheme, and after the regulation, the water flow rate value and the cavitation effect migration trajectory passing through the underwater cavitation type net washing machine at different time nodes are obtained; Based on the analysis of the water flux value and cavitation effect migration trajectory passing through the underwater cavitation screen washing machine at different time points, the water flux change trend and cavitation effect migration state are obtained; The Spearson coefficient is introduced to calculate the correlation coefficient between the water flux change trend and the cavitation effect migration state, and the interference degree between the water flux change trend and the cavitation effect migration state is determined according to the correlation coefficient. If the interference degree is greater than the preset interference degree, a trend regression model is constructed based on the water flux change trend and the cavitation effect migration state. The least square method is introduced to estimate the trend regression model to obtain the hysteresis term coefficient of the water flux change trend for the cavitation effect migration state in the trend regression model, and whether the hysteresis term coefficient is significantly 0 is tested to obtain the test result; Generate a time series statistic of the water flux change trend for the cavitation effect migration state according to the test result, if the time series statistic is greater than a preset time series statistic, it proves that the lag term coefficient in the test result is significantly close to 0, and output the interference trend of the water flux change trend for the cavitation effect migration state according to the time series statistic; Constructing a deep learning prediction model, making predictions in the deep learning prediction model according to the interference trend, obtaining the migration landing point of the cavitation high pressure zone of the underwater cavitation type screen washing machine, and analyzing the migration landing point of the cavitation high pressure zone to obtain the final formation area of ​​the cavitation high pressure zone; The current range to be cleaned of the underwater cavitation type net washing machine is obtained, and the overlap rate between the current range to be cleaned and the final formation area of ​​the cavitation high-pressure zone is calculated. If the overlap rate is less than the preset overlap rate, the pressure parameters of the underwater cavitation type net washing machine are regulated based on the overlap rate to obtain a third regulation scheme.

2. According to claim 1, an intelligent control method for an underwater cavitation type net washing machine, It is characterized in that The method of obtaining the actual texture feature vectors of all attached pollutants on the net to be cleaned, and constructing a standard texture feature vector classification decision tree, identifying the actual texture feature vectors in the standard texture feature vector classification decision tree, and obtaining the adsorption characteristic index of each attached pollutant for the manufacturing material, specifically includes the following steps: The nets to be cleaned currently fixed on the underwater cavitation net washing machine are photographed by an underwater camera to obtain image information of the nets to be cleaned; The GLCM algorithm is introduced to perform grayscale quantization calculation on the image information of the net to be cleaned, and a grayscale co-occurrence matrix is ​​generated. The pixels of the grayscale co-occurrence matrix are traversed based on a preset direction and distance and normalized to obtain the actual texture feature vectors of all attached pollutants on the net to be cleaned; Obtaining the actual fishing environment factor set where the net to be cleaned was last fished, searching in a big data network based on the actual fishing environment factor set to obtain the types of attached pollutants associated with the net to be cleaned, and obtaining a number of standard texture feature vectors of each attached pollutant type under the conditions of the actual fishing environment factor set; A segmentation threshold is preset based on each type of attached pollutant, and N tree trunk nodes of each type of attached pollutant are generated. Each tree trunk node is segmented by a number of standard texture feature vectors to form M groups of standard texture feature vector branch nodes until the segmentation threshold is reached, thereby obtaining a standard texture feature vector classification decision tree; The actual texture feature vectors of all attached pollutants on the net to be cleaned are defined as identification samples, and the identification samples are introduced into a standard texture feature vector classification decision tree for identification. If a leaf node at the end of a branch node can be identified, the identification is stopped, and the type information of each attached pollutant on the net to be cleaned is output; The material of the net to be cleaned is obtained, and a comprehensive knowledge graph of fishery is obtained through a big data network. The type information and material information of each attached pollutant are imported into the comprehensive knowledge graph of fishery to obtain the adsorption characteristic index of each attached pollutant for the material of the net.

3. The intelligent control method of an underwater cavitation type net washing machine according to claim 1, It is characterized in that The three-dimensional simulation model of the underwater cavitation type net washing machine is constructed, the net washing simulation is performed in the three-dimensional simulation model and the bubble volume simulation mean of each attached pollutant in all the areas to be cleaned is obtained, and the underwater cavitation type net washing machine is regulated by the difference analysis between the bubble volume simulation mean and the ideal bubble volume value to obtain the first regulation scheme, which specifically includes the following steps: The size parameters and corresponding adsorption characteristic indexes of all attached pollutants in each range to be cleaned are obtained, and are imported into the preset parameter control decision of the underwater cavitation type net washing machine for analysis, so as to generate the initial working parameters of the underwater cavitation type net washing machine in cleaning each range to be cleaned; wherein the initial working parameters include the initial vibration frequency, the initial amplitude and the initial water flow rate; Constructing a three-dimensional simulation model of an underwater cavitation-type net washing machine, inputting the initial working parameters of the underwater cavitation-type net washing machine in each range to be cleaned into the three-dimensional simulation model to perform net washing simulation, so as to obtain the bubble volume simulation mean value for cleaning the first type of attached pollutants and the bubble volume simulation mean value for cleaning the second type of attached pollutants; The Cardinal curve method is introduced, in which the quadratic polynomial coefficients between the size parameters of all attached pollutants and the corresponding adsorption characteristic index are calculated to obtain several quadratic polynomial coefficients, and several quadratic polynomial coefficients are fitted by spline interpolation to obtain several Cardinal curves; A Cardinal surface is drawn based on the plurality of Cardinal curves, and an ideal bubble volume value required for cleaning the first type of attached pollutants and an ideal bubble volume value required for cleaning the second type of attached pollutants are obtained by aligning the Cardinal surface to the range to be cleaned for volume calculation; Calculate the difference between the simulated mean value of the bubble volume for cleaning the first type of attached pollutants and the simulated mean value of the bubble volume for cleaning the second type of attached pollutants to obtain a simulated bubble difference; and simultaneously calculate the difference between the ideal bubble volume value required for cleaning the first type of attached pollutants and the ideal bubble volume value required for cleaning the second type of attached pollutants to obtain an ideal bubble difference; A hash algorithm is introduced to calculate a hash function between the simulated bubble differences and the ideal bubble differences. If the hash function is greater than a preset hash function, the operating parameters of the underwater cavitation screen washing machine are regulated to obtain a first regulation scheme.

4. An intelligent control system for an underwater cavitation screen washing machine. It is characterized in that The intelligent control system of an underwater cavitation type net washing machine includes a memory and a processor. The memory stores an intelligent control method program of an underwater cavitation type net washing machine. When the intelligent control method program of an underwater cavitation type net washing machine is executed by the processor, the steps of the intelligent control method of an underwater cavitation type net washing machine as described in any one of claims 1 to 3 are implemented.

Citation Information

Patent Citations

  • Air filter and method and device for controlling cleaning thereof

    CN107376525A

  • Ventilation super-cavitation form control method based on machine learning

    CN112363394A

  • Power self-adaptive adjusting system and method for ultrasonic cleaning

    CN118045814A

  • Method and apparatus for controlling robots and the like using a bubble data hierarchy placed along a medial axis

    US6089742A