System for evaluating the hydrophobicity grade of energy equipment superhydrophobic surfaces

Abstract

This invention discloses a hydrophobicity rating system for superhydrophobic surfaces of energy equipment, relating to the field of hydrophobicity rating technology. The system includes a data acquisition module, an optimal parameter combination module, a sample preparation module, a hydrophobicity quantification and evaluation module, a test closed-loop calibration module, and a hydrophobicity rating classification and output module. These modules are interconnected, and through their collaborative operation, the system quantifies, performs closed-loop calibration, and classifies the hydrophobicity index of superhydrophobic samples of energy equipment with different microstructures. The closed-loop calibration process fully considers both anti-icing and weather resistance, significantly improving the accuracy and generalization ability of the hydrophobicity rating results. This solves the problems of inaccurate evaluation results and inability to adapt to extreme field conditions caused by traditional evaluation systems relying solely on conventional laboratory environments, single physical tests, and empirical thresholds for rating. This enhances the intelligence level of hydrophobicity rating evaluation for superhydrophobic surfaces of energy equipment.

Description

Technical Field

[0001] This invention relates to the field of hydrophobicity rating assessment technology, and more specifically to a hydrophobicity rating assessment system for superhydrophobic surfaces of energy equipment. Background Technology

[0002] Energy equipment operates in extreme outdoor environments for extended periods. Low temperatures, high humidity, icing, and condensation can easily lead to surface hydrophilicity, causing safety hazards such as insulation failure and decreased mechanical performance. Superhydrophobic surface modification has become a core technology for improving the anti-icing and weather-resistant performance of equipment. Accurate hydrophobicity rating assessment is a crucial prerequisite for the implementation of this technology. Traditional hydrophobicity rating assessment systems have significant limitations. As the energy industry's demands for equipment safety protection become increasingly stringent, power grids and energy companies urgently need a scientific and practical assessment system to guide the optimization of superhydrophobic surface processes and improve the reliability of equipment in extreme environments. Against this backdrop, the research and application of a comprehensive, quantitative, and environmentally friendly superhydrophobic surface hydrophobicity rating assessment system for energy equipment has become an urgent industry need.

[0003] Traditional hydrophobicity assessment techniques for superhydrophobic surfaces of energy equipment focus only on basic hydrophobicity testing and classification. These assessments are primarily conducted in conventional laboratory environments, failing to simulate the field usage scenarios of energy equipment. Furthermore, their methods rely mainly on single physical tests, lacking technical support at the process control and microscopic mechanism levels. Classification is based solely on empirical thresholds and unquantified correlation models between process and performance, leading to inaccurate hydrophobicity assessment results for superhydrophobic surfaces of energy equipment. Therefore, this invention aims to address the problem of how to deeply analyze the correlation mechanism between plasma etching parameters and particle microscopic parameters, adding two core assessment dimensions—anti-icing performance and weather resistance—to optimize the assessment from static classification to dynamic closed-loop evaluation of hydrophobicity levels. Summary of the Invention

[0004] To achieve the above objectives, the present invention is implemented through the following technical solution: a hydrophobicity level evaluation system for superhydrophobic surfaces of energy equipment, comprising a data acquisition module, an optimal parameter combination module, a sample preparation module, a hydrophobicity quantitative evaluation module, a test closed-loop correction module, and a hydrophobicity level classification and output module, wherein the modules are interconnected. The data acquisition module is used to collect core parameter data of plasma etching to ensure the comprehensiveness and accuracy of the data, laying a solid foundation for subsequent analysis. The optimal parameter combination module, combining Langmuir probe and gas discharge theory, acquires particle microscopic data. It then uses a hybrid neural network architecture of convolutional neural network, attention mechanism and bidirectional long short-term memory network, along with plasma etching core parameter data, to establish a correlation model between plasma etching core parameter data and particle microscopic data. This yields the optimal parameter combination for the sample, achieving deep correlation between the data. The optimal parameter combination for the sample in this step provides a reference for the subsequent preparation of superhydrophobic energy equipment samples with different micromorphologies. The sample preparation module, based on the optimal parameter combination of the sample, deploys a microstructure digital twin module. It digitally models the plasma etching process, the evolution of the sample's microstructure, and the final macroscopic performance. By constructing a three-dimensional simulation model of the plasma etching process, users can prepare samples in a virtual space without repeated sample preparation. This reduces the cost of preparing superhydrophobic samples of energy equipment with different micromorphologies, providing technical support for preparing superhydrophobic samples of energy equipment with different micromorphologies. Furthermore, it collects sample surface characteristic parameters, providing a basis for obtaining subsequent hydrophobicity index and weather resistance parameters.

[0005] The hydrophobicity quantification assessment module is divided into a micro-feature deconstruction unit, a cross-domain transfer learning adaptation unit, and a real-time error compensation calibration unit. It is used to construct a hydrophobicity quantification assessment model based on the surface feature parameters of the sample and output the hydrophobicity index. This step is a preliminary assessment of hydrophobicity. Subsequently, a comprehensive assessment and correction will be carried out in combination with anti-icing and weathering factors. The closed-loop calibration module was tested to perform anti-icing and weathering tests on superhydrophobic energy equipment samples with different microstructures. Anti-icing and weathering parameters were obtained. The water transport closed-loop calibration model was constructed by combining incremental random forest algorithm, online learning mechanism and Bayesian optimization algorithm to obtain closed-loop calibration index. The hydrophobicity index was then calibrated. A feedback closed-loop mechanism was set up. The final hydrophobicity index was obtained based on the deviation between the calibrated hydrophobicity index and the original hydrophobicity index. The hydrophobicity classification and output module uses the K-means clustering algorithm to classify the hydrophobicity of superhydrophobic energy equipment samples with different microstructures based on the final hydrophobicity index. The hydrophobicity level is then output through information push, making it convenient for users to obtain relevant performance data in a timely manner.

[0006] A further improvement to the technical solution of this invention lies in the data acquisition module, where the process of acquiring core parameter data for plasma etching includes: Different types of acquisition devices are deployed to collect core parameter data of plasma etching. The acquisition devices include power meters, pressure sensors, flow meters and timers. The core parameter data of plasma etching include the discharge power of the plasma etching equipment, the gas pressure in the plasma etching chamber, the gas flow rate into the etching chamber and the plasma etching time. Specifically, a power meter, pressure sensor, flow meter, and timer are used to collect the discharge power of the plasma etching equipment, the gas pressure in the plasma etching chamber, the gas flow rate into the etching chamber, and the plasma etching time, respectively. The collected data on discharge power of the plasma etching equipment, gas pressure in the plasma etching chamber, gas flow rate into the etching chamber, and plasma etching time were cleaned and standardized.

[0007] A further improvement to the technical solution of this invention lies in the optimal parameter combination module, where the process of acquiring particle microscopic data includes: Particle microscopic data includes particle temperature and particle density; A Langmuir probe was used to collect the probe current and probe voltage characteristic curves in the central region of the etching cavity, and the effective collection area of ​​the Langmuir probe was recorded. The wavelet transform algorithm is used to extract the electron saturation current segment from the probe current and probe voltage characteristic curves. The electron saturation current of this segment is selected, and the electron temperature of the segment is measured. Combined with the effective collection area of ​​the Langmuir probe, the electron current density is calculated. Based on the plasma probe diagnostic theory, the particle temperature and particle density are obtained through fitting calculation. Based on plasma probe diagnostic theory, a fitting model for fused particle energy distribution is established. With particle temperature and particle density as target variables, a genetic algorithm is used to optimize key parameters such as collision cross section and ionization coefficient in the fitting model, minimizing the deviation between theoretical and measured values. Ultimately, high-precision synergistic inversion of particle temperature and density is achieved. Compared with the traditional method of obtaining particle microscopic data in this invention using a single Langmuir probe, the synergistic inversion process here greatly improves the accuracy of particle temperature and particle density.

[0008] A further improvement to the technical solution of this invention lies in the fact that the optimal parameter combination module, combining a hybrid neural network architecture of convolutional neural network, attention mechanism and bidirectional long short-term memory network, and plasma etching core parameter data, establishes a correlation model to obtain the optimal parameter combination of the sample. The process includes: The optimal parameter combination for the sample includes the optimal discharge power of the plasma etching equipment for superhydrophobic surface samples of energy equipment with different micromorphologies, the optimal gas pressure in the plasma etching chamber, the optimal gas flow rate into the etching chamber, and the optimal plasma etching time. By integrating core parameter data of plasma etching and microscopic data of particles, the high-dimensional data is reduced through principal component analysis. Redundant features in the integrated data are removed, and the core correlation factors in the integrated data are retained. The data is divided into a first training set, a first test set, and a first validation set in a ratio of 8:1:1. The validation set is used to monitor the overfitting risk in the model training process in real time. A hybrid neural network architecture employing convolutional neural networks, attention mechanisms, and bidirectional long short-term memory networks is used. Plasma etching core parameter data and particle microscopic data are taken as input, and the optimal parameter combination of the sample is taken as output. The convolutional neural network extracts local key features from the first training set data, and the attention mechanism strengthens the weights of strongly correlated features between the plasma etching core parameters and particle microscopic data. The bidirectional long short-term memory network captures temporal dependencies in the data sequence, collaboratively learning the correlation between the plasma etching core parameter data, particle microscopic data, and the optimal parameter combination of the sample to obtain an initial correlation model. It should be noted that the reason for using plasma etching core parameter data and particle microscopic data as input is twofold: firstly, the plasma etching core parameter data is a direct control variable of the sample's microstructure, determining the basic effect of the etching process; secondly, the particle microscopic data reflects the intensity and efficiency of particle interaction during the etching process, affecting the formation quality of the microstructure. Using both as input allows for the precise establishment of the correlation between the process, particle interaction, and optimal parameters through the hybrid neural network, providing a scientific basis for preparing samples with the target microstructure, which aligns with the core design logic of process control, microscopic mechanisms, and performance correlation in this invention. The first validation set data is input into the initial association model. An early stopping strategy is used to avoid overfitting of the model. At the same time, an adaptive learning rate adjustment algorithm is used to dynamically adjust the parameters of the association model to obtain an optimized association model. The first test set data is input into the optimized association model. The performance of the current association model is evaluated by the confusion matrix and mean squared error. The model hyperparameters are further fine-tuned using the Bayesian optimization algorithm to obtain the final association model. By combining the current plasma etching core parameter data and particle microscopic data, the optimal parameter combination for the sample is output.

[0009] A further improvement to the technical solution of this invention lies in the sample preparation module, which, based on the optimal parameter combination parameters of the sample, deploys a microstructure digital twin module to construct a three-dimensional simulation model of the plasma etching process, thereby preparing superhydrophobic energy equipment samples with different micromorphologies, and extracting surface feature parameters of the samples. The surface characteristic parameters of the samples include the structural height, spacing, roughness, static contact angle, and dynamic roll-off angle of the superhydrophobic energy equipment samples with different micromorphologies. Based on theories of fluid mechanics, plasma dynamics, and surface chemistry, and according to the plasma etching process, the evolution of the microstructure of the sample, and its macroscopic properties (hydrophobicity), a three-dimensional simulation model of the plasma etching process is constructed. This enables multi-physics coupling modeling for the preparation of superhydrophobic energy equipment samples with different microstructures. Users can adjust the combination of sample parameters on the digital twin platform, quickly explore the possibilities of a large number of sample parameter combinations in a virtual environment, find the theoretically optimal sample parameter combination, and then use it as the final sample parameter combination to prepare superhydrophobic energy equipment samples with different microstructures. The basis of the three-dimensional simulation model of the plasma etching process can be summarized as three core dimensions, revolving around the core logic of plasma etching, microstructure evolution, and multi-physics coupling. Energy equipment samples were selected, and the substrate of the energy equipment samples was subjected to plasma cleaning treatment to ensure that the superhydrophobic surface of the energy equipment samples was free of impurities. Based on the optimal parameter combination of the final samples, the discharge power of the plasma etching equipment, the gas pressure in the plasma etching chamber, the gas flow rate into the etching chamber, and the plasma etching time of the superhydrophobic surface of the energy equipment samples with different micromorphologies were precisely controlled by combining physical deposition and chemical synthesis methods. The superhydrophobic surface of the energy equipment samples with different micromorphologies was then modified with low surface energy. The role of low surface energy modification is to reduce surface free energy and enhance hydrophobicity. The samples after low surface energy modification need to be dried at constant temperature to obtain the final superhydrophobic energy equipment samples with different micromorphologies. By combining scanning electron microscopy, white light interferometer, and optical contact angle measuring instrument, the structural height, spacing, roughness, static contact angle, and dynamic roll-off angle of the surface of superhydrophobic energy equipment samples with different micromorphologies were collected, thereby realizing the extraction of surface characteristic parameters of the samples.

[0010] A further improvement to the technical solution of this invention lies in the fact that the hydrophobicity quantification evaluation module, based on the surface characteristic parameters of the sample, constructs a hydrophobicity quantification evaluation model and outputs a hydrophobicity index, including the following process: R1. For the micro-feature deconstruction unit, based on fractal geometry theory and fluid dynamics model, the surface feature parameters of the sample are decomposed at multiple scales to extract the fractal dimension and interfacial tension contribution coefficient of the microstructure. This overcomes the limitation of traditional models that rely solely on the original parameters. In the hydrophobicity quantification evaluation module, the object of the microstructure is the final superhydrophobic energy equipment sample with different micro-morphologies. Specifically, based on fractal geometry theory, the surface structure parameters of superhydrophobic energy equipment samples with different micro-morphologies at different scales are decomposed to obtain the box side length of the surface of superhydrophobic energy equipment samples with different micro-morphologies and the minimum number of boxes required to cover all surface micro-protrusions. Combined with the fluid dynamics model, the contact state between the surface structure of superhydrophobic energy equipment samples with different micro-morphologies and water is correlated. By statistically analyzing the self-similarity of the surface structure at different scales, the fractal characteristics of the microstructure are quantified to obtain the fractal dimension of the microstructure. Based on the interfacial tension model in fluid dynamics, the interaction strength between the liquid and the surface is analyzed, and the interfacial tension contribution coefficient of the microstructure is derived. R2. For cross-domain transfer learning adaptation units, a meta-learning framework is introduced. The fractal dimension of the microstructure and the contribution coefficient of the interfacial tension are used to train the hydrophobicity quantification evaluation model. Through the task adapter, the model can be quickly adapted to the feature distribution of the current sample to solve the problem of insufficient model training in small sample scenarios. Specifically, using a convolutional neural network as the backbone network, an attention mechanism is introduced to strengthen the weights of the physical features of the microstructure. The fractal dimension and interfacial tension contribution coefficient of the microstructure are used as inputs, and the hydrophobicity index is used as the output to train a hydrophobicity quantification evaluation model. The physical features of the microstructure include the fractal dimension and interfacial tension contribution coefficient of the microstructure. This is the specific implementation of introducing a meta-learning framework to train the hydrophobicity quantification evaluation model. For the different micromorphological types of the superhydrophobic energy equipment samples, a lightweight adapter module is constructed to quickly fine-tune the parameter distribution of the hydrophobicity quantification evaluation model and achieve cross-domain data adaptation. Here, the task adapter is used to quickly adapt to the feature distribution of the current sample and solve the problem of insufficient model training in small sample scenarios. R3. For the real-time error compensation calibration unit, an online Bayesian inference algorithm is embedded to capture the measurement noise of surface characteristic parameters of energy equipment superhydrophobic samples with different micromorphologies in real time, dynamically output the error compensation coefficient, and correct the prediction results of the hydrophobicity quantification evaluation model in real time to ensure the stability of the hydrophobicity index output. Combined with the fractal dimension and interfacial tension contribution coefficient of the current microstructure, the final hydrophobicity index is output.

[0011] A further improvement to the technical solution of this invention lies in the fact that the process of testing the closed-loop correction module to obtain test parameters for superhydrophobic energy equipment samples with different microstructures includes: The tests conducted on superhydrophobic energy equipment samples with different microstructures included anti-icing performance tests and weathering resistance tests, with test parameters including anti-icing index and weathering resistance index. Set up an anti-icing performance test environment, the temperature of which is within the range of... Between, relative humidity Between these steps, superhydrophobic energy equipment samples with different microstructures were placed in an anti-icing performance test environment. Using a thermal imager and a timer, the freezing time of the surface of the superhydrophobic energy equipment samples with different microstructures was recorded under the anti-icing performance test environment. Using an ice adhesion tester, the ice adhesion of the surface of the superhydrophobic energy equipment samples with different microstructures under the anti-icing performance test environment was collected. A weathering resistance testing environment is set up, which includes a natural environment and an aging environment. The natural environment is one where environmental parameters are not altered by human intervention, while the aging environment is one where weathering resistance is tested through natural weathering. High and low temperature cycling and ultraviolet radiation intensity The simulation was conducted by combining scanning electron microscopy, white light interferometer and optical contact angle measuring instrument to collect the surface characteristic parameters of superhydrophobic energy equipment samples with different micromorphologies under natural and aging environments. Combined with the hydrophobicity quantitative evaluation model, the hydrophobicity index of superhydrophobic energy equipment samples with different micromorphologies under natural and aging environments was obtained. The freezing time and ice adhesion of superhydrophobic energy equipment samples with different microstructures under anti-icing performance testing environment are standardized. This standardization process solves the dimension problem in the subsequent calculation of the anti-icing index by weighted summation method. Weights are assigned to the freezing time and ice adhesion of superhydrophobic energy equipment samples with different microstructures under the standardized anti-icing performance testing environment, and the anti-icing index is calculated by weighted summation method. The weather resistance index is obtained by comparing the hydrophobicity index of the surface of superhydrophobic energy equipment samples with different microstructures under aging conditions with the hydrophobicity index of the surface of superhydrophobic energy equipment samples with different microstructures under natural conditions.

[0012] A further improvement to the technical solution of this invention lies in the following: A closed-loop correction module is tested. This module, combined with an incremental random forest algorithm, an online learning mechanism, and a Bayesian optimization algorithm, constructs a water-conveying closed-loop correction model to obtain a closed-loop correction index. The hydrophobicity index is then corrected, and a feedback closed-loop mechanism is set up. The process of obtaining the final hydrophobicity index based on the deviation between the corrected and uncorrected hydrophobicity index includes: The test parameters are integrated into a dataset, and the current integrated dataset is divided into a third training set, a third validation set, and a third test set in a ratio of 7:2:1. Using the incremental random forest algorithm, the test parameters are taken as input data and the closed-loop correction index is taken as output data. Through the ensemble learning framework, the nonlinear mapping relationship between the test parameters and the closed-loop correction index is learned to train the hydrophobic closed-loop correction model. At the same time, the online learning mechanism is used to receive the acquired test parameters in real time and dynamically update the parameters of the hydrophobic closed-loop correction model to maintain the real-time performance and adaptability of the hydrophobic closed-loop correction model, thus obtaining a well-trained water-conveying closed-loop correction model. The third validation set data is input into the trained hydrophobic closed-loop calibration model. The performance of the hydrophobic closed-loop calibration model is monitored by an early stopping strategy to avoid overfitting. Combined with the Bayesian optimization algorithm, the hyperparameters of the hydrophobic closed-loop calibration model are adaptively optimized to obtain the optimized hydrophobic closed-loop calibration model.

[0013] The third test set data is input into the optimized hydrophobic closed-loop correction model to further evaluate the prediction accuracy and generalization ability of the hydrophobic closed-loop correction model. The parameters of the hydrophobic closed-loop correction model are adjusted to obtain the final high-precision water transport closed-loop correction model. The hydrophobic closed-loop correction model takes the test parameters as input and the closed-loop correction index as output. The core objective is to accurately correct the hydrophobicity index. It adopts three dimensions: basic coverage algorithm theory, test mechanism and closed-loop optimization logic, all of which are in line with the actual needs of superhydrophobic surface evaluation of energy equipment. Based on the current test parameters, output the corresponding closed-loop correction index; The hydrophobicity index is corrected based on the closed-loop correction index to obtain the corrected hydrophobicity index. A feedback loop mechanism is set up, using the deviation between the corrected and uncorrected hydrophobicity indices as a reference. A Bayesian optimization algorithm is then used to optimize the hydrophobicity closed-loop correction model. Specifically, a hydrophobicity deviation threshold is set through the feedback loop mechanism. When the deviation between the corrected and uncorrected hydrophobicity indices is lower than or equal to the set threshold, the current hydrophobicity closed-loop correction model is not corrected; instead, the corrected hydrophobicity indices are directly used as the final hydrophobicity indices for subsequent application to energy equipment with different microstructures. The hydrophobicity level of hydrophobic samples is classified as follows: When the deviation between the corrected hydrophobicity index and the original hydrophobicity index is higher than the set hydrophobicity deviation threshold, the secondary scheduling Bayesian optimization algorithm performs hyperparameter fine-tuning on the current hydrophobicity closed-loop correction model until the deviation between the corrected hydrophobicity index and the original hydrophobicity index is lower than or equal to the set hydrophobicity deviation threshold. Combined with the current test parameters, the final hydrophobicity index is obtained, which is used for the subsequent classification of the hydrophobicity level of superhydrophobic energy equipment samples with different micromorphologies.

[0014] The hydrophobicity index is corrected based on the closed-loop correction index to obtain the final hydrophobicity index.

[0015] A further improvement to the technical solution of this invention lies in the hydrophobicity level classification and output module, which, by combining the density peak clustering algorithm and the K-means clustering algorithm, classifies the hydrophobicity level of superhydrophobic energy equipment samples with different microstructures based on the final hydrophobicity index, including steps A1-A5: Hydrophobicity ratings include weak hydrophobicity, moderate hydrophobicity, high hydrophobicity, and super hydrophobicity. Step A1: The final hydrophobicity index is standardized. Differential weights are assigned according to the micromorphology type of the sample to obtain the micromorphology feature weights of the sample. The working condition factor is set in combination with extreme field conditions (such as low temperature and high humidity, icing, strong ultraviolet radiation, etc.). Then, a multi-dimensional evaluation vector of hydrophobicity index, sample micromorphology weights and working condition factor is constructed. Combined with the double verification of elbow rule and profile coefficient, the optimal k value is determined to be 4. The optimal k value corresponds to 4 hydrophobicity levels. Step A2 involves analyzing the density distribution of the multi-dimensional evaluation vectors using a density peak clustering algorithm. Four standardized hydrophobicity indices are selected as four cluster centers, and an upper limit for the number of iterations and a convergence threshold for the cluster centers are set. These are numbered as follows: , , and This implements the initialization process for cluster centers. Step A3: Calculate the Euclidean distance from each standardized hydrophobicity index to the four cluster centers, and assign the superhydrophobic energy equipment sample corresponding to each hydrophobicity index to the category corresponding to the nearest cluster center. Step A4: Take the average hydrophobicity index of all superhydrophobic energy equipment samples within the category corresponding to each cluster center, and then update each cluster center. Step A5: Repeat steps A3 and A4, record the actual number of iterations. When the actual number of iterations reaches the set number of iterations and the cluster centers meet the set cluster center convergence threshold, stop the iteration and determine the final cluster centers. Based on the size of the final cluster centers, classify the hydrophobicity level of the superhydrophobic energy equipment samples with different micromorphologies. The specific classification process is as follows: the final cluster center value ranges from 0 to 1. When the final cluster center is greater than or equal to 0 and less than 0.5, it corresponds to a weak hydrophobicity level; when the final cluster center is greater than or equal to 0.5 and less than 0.7, it corresponds to a medium hydrophobicity level; when the final cluster center is greater than or equal to 0.7 and less than 0.85, it corresponds to a high hydrophobicity level; and when the final cluster center is greater than or equal to 0.85 and less than or equal to 1, it corresponds to a super hydrophobicity level.

[0016] A further improvement to the technical solution of this invention lies in that the hydrophobicity level classification and output module, through information push, outputs the final hydrophobicity level in the following process: Each superhydrophobic energy equipment sample with a different microstructure was assigned a number, and the number was correlated with the final hydrophobicity index and hydrophobicity level. The number, hydrophobicity index and hydrophobicity level of the superhydrophobic energy equipment samples with different microstructures were integrated in the form of graphic and textual reports. The system uses information push notifications via PC pop-ups, mobile push notifications, and emails to send the numbers, hydrophobicity indices, and hydrophobicity levels of superhydrophobic energy equipment samples with different microstructures to a preset receiving end, thereby enabling a visual summary output of the final hydrophobicity level.

[0017] The beneficial effects of this invention are as follows: Compared with traditional hydrophobicity evaluation systems, the hydrophobicity rating system for superhydrophobic surfaces of energy equipment in this invention closely integrates plasma probe diagnostic technology, convolutional neural network algorithm, random forest closed-loop correction technology, and modern information technology. This allows for the precise capture of core plasma etching parameter data and particle microscopic data. Based on this data, the optimal etching parameter combination for samples with different micromorphologies is obtained, enabling the preparation of superhydrophobic samples of energy equipment with different micromorphologies. This yields surface characteristic parameters of the samples, providing an important basis for quantifying the hydrophobicity index and its closed-loop correction. Finally, K-means clustering and density peak clustering algorithms are used to complete the evaluation. The precise classification of hydrophobicity levels enables real-time and comprehensive monitoring of the hydrophobicity levels of superhydrophobic surfaces of energy equipment. Furthermore, it relies on information push to accurately output the classification results. This solves the problems of inaccurate evaluation results and inability to adapt to extreme field conditions caused by traditional evaluation systems that rely solely on conventional laboratory environments, single physical tests, and empirical thresholds for classification. This ensures that the method in this invention can refine the dynamic monitoring standards for the hydrophobicity level evaluation system of superhydrophobic surfaces of energy equipment within a more precise range, making the monitored data more accurate indicators under the same conditions. The development and application of this method significantly enhances the intelligence level of the hydrophobicity level evaluation process for superhydrophobic surfaces of energy equipment. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0019] Figure 1 This is a block diagram of the hydrophobicity rating system for superhydrophobic surfaces of energy equipment according to the present invention. Figure 2This is a map showing the distribution of hydrophobicity levels. Figure 3 This is a graph showing the relationship between the static contact angle and the corrected hydrophobicity index. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] like Figure 1 As shown, the present invention provides a hydrophobicity level evaluation system for superhydrophobic surfaces of energy equipment, including a data acquisition module, an optimal parameter combination module, a sample preparation module, a hydrophobicity quantification evaluation module, a test closed-loop correction module, and a hydrophobicity level classification and output module, wherein the modules are interconnected. The data acquisition module is used to collect core parameter data of plasma etching, laying a solid data foundation for the implementation of subsequent modules; The optimal parameter combination module combines Langmuir probe and gas discharge theory to obtain particle microscopic data. It combines a hybrid neural network architecture of convolutional neural network, attention mechanism and bidirectional long short-term memory network with plasma etching core parameter data to establish a correlation model between plasma etching core parameter data and particle microscopic data, obtain the optimal parameter combination of the sample, and accurately anchor the core process parameters for the preparation of superhydrophobic energy equipment samples with different micromorphologies. The sample preparation module, based on the optimal parameter combination of the sample, deploys a microstructure digital twin module to construct a three-dimensional simulation model of the plasma etching process. It digitally models the plasma etching process, the evolution of the sample's microstructure, and the final macroscopic performance. By constructing a high-fidelity digital twin model, users can prepare samples in a virtual space without repeated sample preparation, reducing the cost of preparing superhydrophobic energy equipment samples with different microstructures. This provides technical support for preparing superhydrophobic energy equipment samples with different microstructures and collects sample surface characteristic parameters, providing a characteristic basis for the quantitative evaluation of the hydrophobicity of superhydrophobic energy equipment samples with different microstructures.

[0022] The hydrophobicity quantification assessment module is divided into a micro-feature deconstruction unit, a cross-domain transfer learning adaptation unit, and a real-time error compensation calibration unit. It is used to construct a hydrophobicity quantification assessment model based on the sample surface feature parameters, output the hydrophobicity index, and realize the accurate quantitative characterization of hydrophobicity. The closed-loop calibration module was tested to perform anti-icing and weathering tests on superhydrophobic energy equipment samples with different microstructures. Anti-icing and weathering parameters were obtained. By combining incremental random forest algorithm, online learning mechanism and Bayesian optimization algorithm, a water transport closed-loop calibration model was constructed to obtain the closed-loop calibration index. The hydrophobicity index was then calibrated. A feedback closed-loop mechanism was set up. Based on the deviation between the calibrated hydrophobicity index and the original hydrophobicity index, the final hydrophobicity index was obtained, which effectively improved the authenticity and adaptability of the hydrophobicity index. The hydrophobicity classification and output module combines density peak clustering and K-means clustering algorithms. Based on the final hydrophobicity index, it classifies the hydrophobicity of superhydrophobic energy equipment samples with different micromorphologies and outputs the hydrophobicity level through information push, thus achieving objective quantification of hydrophobicity classification and efficient delivery of results.

[0023] In some embodiments, the data acquisition module's process for acquiring core parameter data of plasma etching includes: Different types of acquisition devices are deployed to collect core parameter data of plasma etching. The acquisition devices include power meters, pressure sensors, flow meters and timers. The core parameter data of plasma etching include the discharge power of the plasma etching equipment, the gas pressure in the plasma etching chamber, the gas flow rate into the etching chamber and the plasma etching time. Specifically, a power meter, pressure sensor, flow meter, and timer are used to collect the discharge power of the plasma etching equipment, the gas pressure in the plasma etching chamber, the gas flow rate into the etching chamber, and the plasma etching time, respectively. The collected data on discharge power of the plasma etching equipment, gas pressure in the plasma etching chamber, gas flow rate into the etching chamber, and plasma etching time were cleaned and standardized.

[0024] In some embodiments, the optimal parameter combination module, the process of acquiring particle microscopic data includes: Particle microscopic data includes particle temperature and particle density; A Langmuir probe was used to collect the probe current and probe voltage characteristic curves in the central region of the etching cavity, and the effective collection area of ​​the Langmuir probe was recorded. The wavelet transform algorithm was used to extract the electron saturation current segment from the probe current-probe voltage characteristic curves. The electron saturation current of this segment was selected, and the electron temperature within this segment was measured. Combined with the effective collection area of ​​the Langmuir probe, the electron current density was calculated. Based on plasma probe diagnostic theory, the particle temperature and particle density were obtained through fitting calculations. The relevant formulas are as follows: ; ; ;

[0025] in, The particle temperature is expressed in Kelvin (K). Electron current density, measured in amperes per square meter, abbreviated as . ; This refers to the electron saturation current in the electron saturation current segment of the probe current vs. probe voltage characteristic curve, measured in amperes (A). ; The effective collection area of ​​the Langmuir probe, in square meters, is abbreviated as ; Let be the rest mass of the electron, a standard physical constant, with a value of . ; It is the Boltzmann constant, a standard physical constant with a value of [value missing]. Its unit is joule per kelvin, abbreviated as joule per kelvin. ,in, ; It is the Langmuir probe collection coefficient, which is dimensionless and ranges from 0.8 to 0.95. The elementary charge, also known as the electron charge, has a value of The unit is coulomb, abbreviated as C; Particle density, its unit is abbreviated as . ; This is a correction term for the etching reaction, dimensionless, with a value ranging from 0.8 to 0.95; This refers to the gas pressure within the plasma etching chamber, measured in Pascals (Pa). , Therefore, there exists ; It is the electron temperature in the electron saturation current range, and the unit is Kelvin, abbreviated as K. According to the dimensional derivation, the above formula meets the dimensional requirements. These formulas originate from the Langmuir probe diagnostic theory of plasma, which is the core formula of the probe method proposed by Irving Langmuir in the 1920s. It has been improved by plasma physicists and has become one of the classic methods for plasma parameter diagnosis. Based on plasma probe diagnostic theory, a fitting model for fused particle energy distribution is established. With particle temperature and particle density as target variables, a genetic algorithm is used to optimize key parameters such as collision cross section and ionization coefficient in the fitting model, minimizing the deviation between theoretical and measured values. Ultimately, the synergistic inversion of particle temperature and density is achieved. Compared with the traditional method of obtaining particle microscopic data in this invention using a single Langmuir probe, the synergistic inversion process here greatly improves the accuracy of particle temperature and particle density.

[0026] In some embodiments, the optimal parameter combination module, combining a hybrid neural network architecture of convolutional neural network, attention mechanism and bidirectional long short-term memory network with core parameter data of plasma etching, establishes a correlation model to obtain the optimal parameter combination of the sample, including: The optimal parameter combination for the sample includes the optimal discharge power of the plasma etching equipment for superhydrophobic surface samples of energy equipment with different micromorphologies, the optimal gas pressure in the plasma etching chamber, the optimal gas flow rate into the etching chamber, and the optimal plasma etching time. The core parameter data of plasma etching and particle microscopic data were integrated. Principal component analysis (PCA) was used to reduce the dimensionality of the high-dimensional data, eliminating redundant features and retaining the core correlation factors. The data was then divided into a first training set, a first test set, and a first validation set in an 8:1:1 ratio. The validation set was used to monitor the overfitting risk during model training in real time. The process of obtaining the core correlation factors is as follows: First, the core parameter data of plasma etching and particle microscopic data were integrated to form a high-dimensional original dataset. Second, principal component analysis was used to reduce the dimensionality of the high-dimensional data. By analyzing the eigenvalues ​​and eigenvectors of the high-dimensional data, principal components that can explain the main variations in the data (i.e., the core correlations) were selected. Finally, redundant features that contributed little to the data variation and had a weak correlation with the optimal parameter combination of the sample were eliminated, and key features that were strongly correlated with the etching effect and microscopic morphology were retained. These are the core correlation factors. It should be noted that the dimensionality-reduced data is a feature set that retains the core correlations and simplifies the dimensions. This dimensionality-reduced data mainly focuses on the key information supporting the optimal parameter combination of the sample. A hybrid neural network architecture combining convolutional neural networks, attention mechanisms, and bidirectional long short-term memory networks is adopted. The core parameters of plasma etching and particle microscopic data are used as inputs, and the optimal parameter combination of the sample is used as the output. The convolutional neural network extracts local key features from the first training set data, the attention mechanism strengthens the weights of the strong correlation features between the core parameters of plasma etching and particle microscopic data, and the bidirectional long short-term memory network captures the temporal dependencies in the data sequence. The correlation between the core parameters of plasma etching, particle microscopic data, and the optimal parameter combination of the sample is learned collaboratively to obtain the initial correlation model. The first validation set data is input into the initial association model. An early stopping strategy is used to avoid overfitting of the model. At the same time, an adaptive learning rate adjustment algorithm is used to dynamically adjust the parameters of the association model to obtain an optimized association model. The first test set data is input into the optimized association model. The performance of the current association model is evaluated by the confusion matrix and mean squared error. If the performance does not reach the preset threshold, the Bayesian optimization algorithm is used to further fine-tune the model hyperparameters to obtain the final association model. By combining the current plasma etching core parameter data and particle microscopic data, the optimal parameter combination for the sample is output.

[0027] In some embodiments, the sample preparation module, based on the optimal parameter combination parameters of the sample, deploys a microstructure digital twin module to prepare superhydrophobic energy equipment samples with different micromorphologies, and the process of extracting sample surface characteristic parameters includes: The surface characteristic parameters of the samples include the structural height, spacing, roughness, static contact angle, and dynamic roll-off angle of the superhydrophobic energy equipment samples with different micromorphologies. Based on theories of fluid mechanics, plasma dynamics, and surface chemistry, and according to the plasma etching process, the evolution of the microstructure of the sample, and its macroscopic properties (hydrophobicity), a three-dimensional simulation model of the plasma etching process is constructed. This model enables multi-physics coupling modeling for the preparation of superhydrophobic energy equipment samples with different microstructures. Users can adjust the combination of sample parameters on the digital twin platform, quickly explore the possibilities of a large number of sample parameter combinations in a virtual environment, find the theoretically optimal sample parameter combination, and then use it as the final sample parameter combination to prepare superhydrophobic energy equipment samples with different microstructures. The basis of the three-dimensional simulation model of the plasma etching process can be summarized into three core dimensions, revolving around the core logic of plasma etching, microstructure evolution, and multi-physics coupling. Specifically, theoretically, the plasma dynamics equation and Langmuir probe diagnostic theory are the core, supporting the physical rationality of the motion of etching particles, energy transfer, and parameter correlation, clarifying the intrinsic relationship between microstructure and hydrophobicity, providing a theoretical basis for the simulation target, and verifying the feasibility of the technical path of virtual parameter adjustment and optimal combination selection based on the digital twin "physical entity-virtual model-optimization decision" architecture. The aforementioned final sample parameter combination refers to the plasma etching parameter combination that, through multi-physics field coupling simulation verification in a digital twin virtual environment, enables the prepared superhydrophobic sample to simultaneously meet the requirements of microscopic morphology characteristics, synergistic matching with particle microscopic data, and possess anti-icing and weathering potential. It typically includes optimal discharge power, optimal gas pressure, optimal gas flow rate, and optimal etching time, and is the core basis for subsequent physical sample preparation. Energy equipment samples were selected, and the substrate of the energy equipment samples was subjected to plasma cleaning treatment to ensure that the superhydrophobic surface of the energy equipment samples was free of impurities. Based on the optimal parameter combination of the final samples, the discharge power of the plasma etching equipment, the gas pressure in the plasma etching chamber, the gas flow rate into the etching chamber, and the plasma etching time of the superhydrophobic surface of the energy equipment samples with different micromorphologies were precisely controlled by combining physical deposition and chemical synthesis methods. The superhydrophobic surface of the energy equipment samples with different micromorphologies was then modified with low surface energy. The role of low surface energy modification is to reduce surface free energy and enhance hydrophobicity. The samples after low surface energy modification need to be dried at constant temperature to obtain the final superhydrophobic energy equipment samples with different micromorphologies. Using a scanning electron microscope, the structural height and spacing of the surface of the superhydrophobic energy equipment samples with different microstructures were collected. A white light interferometer was used to collect the surface roughness of the superhydrophobic energy equipment samples with different microstructures. An optical contact angle meter was used to collect the static contact angle and dynamic roll-off angle of the surface of the superhydrophobic energy equipment samples with different microstructures, thereby extracting the characteristic parameters of the sample surface.

[0028] In some embodiments, the process by which the hydrophobicity quantification evaluation module constructs a hydrophobicity quantification evaluation model based on sample surface characteristic parameters and outputs a hydrophobicity index includes: For the microscopic feature deconstruction unit, based on fractal geometry theory and fluid dynamics model, the surface feature parameters of the sample are decomposed at multiple scales to extract the fractal dimension and interfacial tension contribution coefficient of the microstructure. This overcomes the limitation of traditional models that rely solely on the original parameters. In the hydrophobicity quantification evaluation module, the object of the microstructure is the final superhydrophobic energy equipment sample with different micromorphologies. The calculation process of the fractal dimension and interfacial tension contribution coefficient of the microstructure includes: ; ;

[0029] in, and These represent the fractal dimension of the microstructure and the contribution coefficient of interfacial tension, respectively. The side length of the box on the surface of the superhydrophobic sample of energy equipment with different micromorphologies. The minimum number of boxes required to cover all surface micro-protrusions The static contact angle of the surface of the superhydrophobic energy equipment samples with different micromorphologies was determined. The static contact angle of the superhydrophobic sample (without microstructure) substrate for energy equipment. Roughness of the surface of superhydrophobic energy equipment samples with different micromorphologies; For cross-domain transfer learning adaptation units, a meta-learning framework is introduced. The fractal dimension of the microstructure and the contribution coefficient of the interfacial tension are used to train the hydrophobicity quantification evaluation model. Through the task adapter, the model can be quickly adapted to the feature distribution of the current sample, thus solving the problem of insufficient model training in small sample scenarios. Specifically, using a convolutional neural network as the backbone network, an attention mechanism is introduced to strengthen the weights of the physical features of the microstructure. The fractal dimension and interfacial tension contribution coefficient of the microstructure are used as inputs, and the hydrophobicity index is used as the output to train a hydrophobicity quantification evaluation model. The physical features of the microstructure include the fractal dimension and interfacial tension contribution coefficient of the microstructure. This is the specific implementation of introducing a meta-learning framework to train the hydrophobicity quantification evaluation model. For the different micromorphological types of the superhydrophobic energy equipment samples, a lightweight adapter module is constructed to quickly fine-tune the parameter distribution of the hydrophobicity quantification evaluation model and achieve cross-domain data adaptation. Here, the task adapter is used to quickly adapt to the feature distribution of the current sample and solve the problem of insufficient model training in small sample scenarios. For the real-time error compensation calibration unit, an online Bayesian inference algorithm is embedded to capture the measurement noise of surface characteristic parameters of superhydrophobic energy equipment samples with different microstructures in real time. The error compensation coefficient is dynamically output to correct the prediction results of the hydrophobicity quantification evaluation model in real time, ensuring the stability of the hydrophobicity index output. Combined with the fractal dimension and interfacial tension contribution coefficient of the current microstructure, the final hydrophobicity index is output. This hydrophobicity index is dimensionless and distributed between 0 and 1. Its correlation with the fractal dimension and interfacial tension contribution coefficient of the microstructure is as follows: [Fractal dimension inverse...] The degree of self-similarity and complexity of the microstructure is positively correlated with the hydrophobicity index. The higher the fractal dimension, the higher the hydrophobicity index, indicating that it is easier to capture air to form a three-phase interface of solid, gas, and liquid, reducing the contact area between the liquid and the surface, thereby improving hydrophobicity. The interfacial tension contribution coefficient quantifies the regulatory effect of the microstructure on the interfacial tension between solid and liquid. It is positively correlated with the hydrophobicity index. The larger the interfacial tension contribution coefficient, the more significant the weakening effect of the microstructure on the interfacial tension, the greater the resistance to liquid spreading on the surface, the larger the static contact angle, the smaller the dynamic roll-off angle, and the higher the hydrophobicity index.

[0030] In some embodiments, the process of testing the closed-loop correction module to test superhydrophobic energy equipment samples with different microstructures and obtaining test parameters includes: The tests conducted on superhydrophobic energy equipment samples with different microstructures included anti-icing performance tests and weathering resistance tests, with test parameters including anti-icing index and weathering resistance index. Set up an anti-icing performance test environment, the temperature of which is within the range of... Between, relative humidity In this process, superhydrophobic energy equipment samples with different microstructures were placed in an anti-icing performance testing environment. Using a thermal imager and timer, the freezing time of the samples under different microstructures was recorded. An ice adhesion tester was used to collect the ice adhesion force on the samples under the same environment. It should be explained that the principle of recording the freezing time of the samples using a thermal imager and timer is as follows: the thermal imager helps observe the temperature non-uniformity and the starting point of freezing on the surface of the samples, while the timer records the specific freezing time. A weathering resistance testing environment is set up, which includes a natural environment and an aging environment. The natural environment is one where environmental parameters are not altered by human intervention, while the aging environment is one where weathering resistance is tested through natural weathering. High and low temperature cycling and ultraviolet radiation intensity The simulation was conducted by combining scanning electron microscopy, white light interferometer and optical contact angle measuring instrument to collect the surface characteristic parameters of superhydrophobic energy equipment samples with different micromorphologies under natural and aging environments. Combined with the hydrophobicity quantitative evaluation model, the hydrophobicity index of superhydrophobic energy equipment samples with different micromorphologies under natural and aging environments was obtained. The freezing time and ice adhesion of superhydrophobic energy equipment samples with different microstructures under anti-icing performance testing conditions were standardized. This standardization process resolved the dimensionality problem in the subsequent weighted summation method for calculating the anti-icing index. Weights were assigned to the freezing time and ice adhesion of the standardized superhydrophobic energy equipment samples with different microstructures under the anti-icing performance testing conditions. The weighted summation method was then used to calculate the anti-icing index, and the formula is as follows: ;

[0031] in, The ice resistance index ranges from 0 to 1. and The weights of freezing time and ice adhesion on the surface of superhydrophobic energy equipment samples with different micromorphologies under the standardized anti-icing performance test environment are respectively: and The ice-freeing time and ice adhesion of superhydrophobic energy equipment samples with different microstructures under standardized anti-icing performance testing conditions are respectively: The weather resistance index is obtained by comparing the hydrophobicity index of the surface of superhydrophobic energy equipment samples with different microstructures under aging conditions with the hydrophobicity index of the surface of superhydrophobic energy equipment samples with different microstructures under natural conditions.

[0032] In some embodiments, the closed-loop correction module is tested by combining incremental random forest algorithm, online learning mechanism and Bayesian optimization algorithm to construct water-conveying closed-loop correction model, obtain closed-loop correction index, correct the hydrophobicity index, set up feedback closed-loop mechanism, and obtain the final hydrophobicity index based on the deviation between the corrected hydrophobicity index and the original hydrophobicity index. The process includes: The test parameters are integrated into a dataset, and the current integrated dataset is divided into a third training set, a third validation set, and a third test set in a ratio of 7:2:1. Using the incremental random forest algorithm, test parameters are taken as input data and closed-loop correction index is taken as output data. The nonlinear relationship between ice resistance index, weather resistance index and closed-loop correction index is learned through an ensemble learning framework to train a hydrophobic closed-loop correction model. At the same time, the acquired test parameters are received in real time using an online learning mechanism to dynamically update the parameters of the hydrophobic closed-loop correction model, maintain the real-time performance and adaptability of the hydrophobic closed-loop correction model, and obtain a well-trained water transport closed-loop correction model. The third validation set data is input into the trained hydrophobic closed-loop correction model. The performance of the hydrophobic closed-loop correction model is monitored by the early stopping strategy to avoid overfitting. Combined with the Bayesian optimization algorithm, the hyperparameters of the hydrophobic closed-loop correction model (such as the number of decision trees, the size of the feature subset, the node splitting threshold, etc.) are adaptively optimized to obtain the optimized hydrophobic closed-loop correction model. The third test set data is input into the optimized hydrophobic closed-loop correction model to further evaluate its prediction accuracy and generalization ability. The parameters of the hydrophobic closed-loop correction model are adjusted to obtain the final high-precision water-transporting closed-loop correction model. Combined with the current test parameters, the corresponding closed-loop correction index is output. This hydrophobic closed-loop correction model takes the test parameters as input and the closed-loop correction index as output, with the core objective being accurate correction of the hydrophobicity index. Its underlying principles cover three dimensions: algorithm theory, testing mechanism, and closed-loop optimization logic, all of which align with the actual needs of superhydrophobic surface evaluation for energy equipment. Specifically, the algorithm theory utilizes incremental random forest and Bayesian optimization algorithms to ensure the water-transporting closed-loop correction model has the ability to "handle multi-dimensional nonlinear features, update in real time, and accurately optimize." The incremental random forest algorithm is based on the ensemble learning framework of random forest (reducing overfitting risk through multi-decision tree voting) and incremental learning... The dynamic update characteristic, taking test parameters as input (multi-dimensional nonlinear features), requires real-time integration of new test data to update parameters, which perfectly matches the algorithmic advantages of incremental random forest. Meanwhile, the Bayesian optimization algorithm is based on the hyperparameter adaptive optimization logic of Bayes' theorem (quickly locating the optimal hyperparameters through posterior probability distribution, reducing blind search). The testing mechanism references the online learning mechanism, which is based on the core framework of online learning for updating the parameters of the water-carrying closed-loop correction model on a sample-by-sample basis and maintaining real-time adaptability. It eliminates the need to store all historical data, reducing computational overhead and adapting to sample characteristics under different testing environments, aligning with the application scenarios of online learning. The closed-loop optimization logic draws on the theoretical support of hydrophobicity index quantization correction and feedback closed-loop mechanisms, ensuring that the water-carrying closed-loop correction model can be continuously optimized through dynamic feedback, solving the problems of "poor adaptability and insufficient accuracy" of traditional static models. Ultimately, this achieves the core design of dual-dimensional correction of ice resistance and weather resistance in this invention. The hydrophobicity index is corrected based on the closed-loop correction index to obtain the final hydrophobicity index. The specific calibration process is as follows: Before the calibration process begins, it should be noted that both the closed-loop calibration index and the hydrophobicity index are between 0 and 1. Set the first closed-loop calibration threshold and the second closed-loop calibration threshold. The first closed-loop calibration threshold is less than the second closed-loop calibration threshold. When the closed-loop correction index is less than or equal to the first closed-loop correction threshold, according to The formula is used to correct the hydrophobicity index; when the closed-loop correction index is greater than the first closed-loop correction threshold and less than the second closed-loop correction threshold, no correction is performed on the hydrophobicity index; when the closed-loop correction index is greater than or equal to the second closed-loop correction threshold, correction is performed according to... The formula corrects for the hydrophobicity index, where... This is the closed-loop correction index. The hydrophobicity index before correction. This is the corrected hydrophobicity index, i.e., the corrected hydrophobicity index; A feedback loop mechanism is set up, using the deviation between the corrected and uncorrected hydrophobicity indices as a reference. A Bayesian optimization algorithm is then used to optimize the hydrophobicity closed-loop correction model. Specifically, a hydrophobicity deviation threshold is set through the feedback loop mechanism. When the deviation between the corrected and uncorrected hydrophobicity indices is lower than or equal to the set threshold, the current hydrophobicity closed-loop correction model is not corrected; instead, the corrected hydrophobicity indices are directly used as the final hydrophobicity indices for subsequent application to energy equipment with different microstructures. The hydrophobicity level of hydrophobic samples is classified as follows: When the deviation between the corrected hydrophobicity index and the original hydrophobicity index is higher than the set hydrophobicity deviation threshold, the secondary scheduling Bayesian optimization algorithm performs hyperparameter fine-tuning on the current hydrophobicity closed-loop correction model until the deviation between the corrected hydrophobicity index and the original hydrophobicity index is lower than or equal to the set hydrophobicity deviation threshold. Combined with the current test parameters, the final hydrophobicity index is obtained, which is used for the subsequent classification of the hydrophobicity level of superhydrophobic energy equipment samples with different micromorphologies.

[0033] In some embodiments, the hydrophobicity classification and output module, combining the density peak clustering algorithm and the K-means clustering algorithm, classifies the hydrophobicity levels of superhydrophobic energy equipment samples with different microstructures based on the final hydrophobicity index, including steps A1-A5: Hydrophobicity ratings include weak hydrophobicity, moderate hydrophobicity, high hydrophobicity, and super hydrophobicity. Step A1: The final hydrophobicity index is standardized. Differential weights are assigned according to the micromorphology type of the sample to obtain the micromorphology feature weights of the sample. The working condition factor is set in combination with extreme field conditions (such as low temperature and high humidity, icing, strong ultraviolet radiation, etc.). Then, a multi-dimensional evaluation vector of hydrophobicity index, sample micromorphology weights and working condition factor is constructed. Combined with the double verification of elbow rule and profile coefficient, the optimal k value is determined to be 4. The optimal k value corresponds to 4 hydrophobicity levels. Step A2 involves analyzing the density distribution of the multi-dimensional evaluation vectors using a density peak clustering algorithm. Four standardized hydrophobicity indices are selected as four cluster centers, and an upper limit for the number of iterations and a convergence threshold for the cluster centers are set. These are numbered as follows: , , and This implements the initialization process for cluster centers. Step A3: Calculate the Euclidean distance from each standardized hydrophobicity index to the four cluster centers, and assign the superhydrophobic energy equipment sample corresponding to each hydrophobicity index to the category corresponding to the nearest cluster center. The formula for calculating the Euclidean distance is: , The Euclidean distances from each superhydrophobic energy equipment sample to the four cluster centers are given. Hydrophobicity index of superhydrophobic samples for various energy equipment after standardized treatment. The initial cluster centers are defined as s, where s corresponds to different superhydrophobic energy equipment samples. For different cluster centers, j can take values ​​of 1, 2, 3, and 4; Step A4: Take the average hydrophobicity index of all superhydrophobic energy equipment samples within the category corresponding to each cluster center, and then update each cluster center. The calculation formula involved is as follows: , For the updated cluster centers, The total number of superhydrophobic energy equipment samples in the j-th cluster center. The hydrophobicity index of the s-th superhydrophobic energy equipment sample in the j-th cluster center is the standardized hydrophobicity index. Step A5: Repeat steps A3 and A4, record the actual number of iterations. When the actual number of iterations reaches the set number of iterations and the cluster centers meet the set cluster center convergence threshold, stop the iteration and determine the final cluster centers. Based on the size of the final cluster centers, classify the hydrophobicity level of the superhydrophobic energy equipment samples with different micromorphologies. The specific classification process is as follows: the final cluster center value ranges from 0 to 1. When the final cluster center is greater than or equal to 0 and less than 0.5, it corresponds to a weak hydrophobicity level; when the final cluster center is greater than or equal to 0.5 and less than 0.7, it corresponds to a medium hydrophobicity level; when the final cluster center is greater than or equal to 0.7 and less than 0.85, it corresponds to a high hydrophobicity level; and when the final cluster center is greater than or equal to 0.85 and less than or equal to 1, it corresponds to a super hydrophobicity level.

[0034] In some embodiments, the process by which the hydrophobicity rating and output module outputs the final hydrophobicity rating via information push includes: Each superhydrophobic energy equipment sample with a different microstructure was assigned a number, and the number was correlated with the final hydrophobicity index and hydrophobicity level. The number, hydrophobicity index and hydrophobicity level of the superhydrophobic energy equipment samples with different microstructures were integrated in the form of graphic and textual reports. The system uses information push notifications via PC pop-ups, mobile push notifications, and emails to send the numbers, hydrophobicity indices, and hydrophobicity levels of superhydrophobic energy equipment samples with different microstructures to a preset receiving end, thereby enabling a visual summary output of the final hydrophobicity level.

[0035] like Figure 2 The hydrophobicity level distribution chart, presented in bar chart form, shows the sample quantity distribution of four hydrophobicity levels: weak hydrophobicity, moderate hydrophobicity, high hydrophobicity, and superhydrophobicity. The horizontal axis represents the hydrophobicity level, and the vertical axis represents the sample quantity, which specifically refers to the number of superhydrophobic energy equipment samples with different microstructures. This chart allows for a quick view of the proportion of superhydrophobic energy equipment samples with different hydrophobicity levels, providing an intuitive statistical display for evaluating the output results of this invention and clearly identifying the distribution differences of various hydrophobicity levels.

[0036] like Figure 3 The relationship between static contact angle and corrected hydrophobicity index is shown in the figure. The three sub-figures, from top to bottom, correspond to superhydrophobic energy equipment samples with three different microstructures: micropillar structure, nanowire structure, and hierarchical structure. The horizontal axis represents the static contact angle, and the vertical axis represents the corrected hydrophobicity index. The core of the figure reflects the correlation between static contact angle and hydrophobicity index. The correlation characteristics of samples with different microstructures are different, which provides a visual basis for analyzing the intrinsic relationship between surface morphology, contact angle, and hydrophobicity. Among them, the micropillar structure subgraph reflects the correlation between the static contact angle and the hydrophobicity index of the micropillar structure sample. The average upper limit of the corrected hydrophobicity index is relatively low, indicating that the structure has certain limitations in improving hydrophobicity. The nanowire structure sub-graph reflects the hydrophobicity correlation characteristics of the nanowire structure sample. Its corresponding corrected hydrophobicity index rises rapidly to a high level, indicating that the structure is more likely to achieve high hydrophobicity. The hierarchical sub-graph shows the hydrophobicity of the hierarchical sample. The corresponding corrected hydrophobicity index can be stably maintained at a high level, reflecting that the hydrophobicity of this structure is more stable.

[0037] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A hydrophobicity rating system for superhydrophobic surfaces of energy equipment, comprising a data acquisition module, an optimal parameter combination module, a sample preparation module, a hydrophobicity quantification evaluation module, a test closed-loop calibration module, and a hydrophobicity rating classification and output module. The various modules are connected for communication, characterized in that, The data acquisition module is used to acquire core parameter data of plasma etching; The optimal parameter combination module is used to acquire particle microscopic data, combine a hybrid neural network architecture of convolutional neural network, attention mechanism and bidirectional long short-term memory network with the core parameter data of plasma etching, establish a correlation model, and obtain the optimal parameter combination of the sample. The sample preparation module deploys a microstructure digital twin module to construct a three-dimensional simulation model of the plasma etching process. Combined with the optimal parameter combination of the sample, it prepares superhydrophobic energy equipment samples with different micromorphologies and extracts the surface feature parameters of the sample. The hydrophobicity quantification evaluation module is divided into a micro-feature deconstruction unit, a cross-domain transfer learning adaptation unit, and a real-time error compensation calibration unit, which are used to construct a hydrophobicity quantification evaluation model based on the surface feature parameters of the sample and output the hydrophobicity index. The test closed-loop correction module is used to test the superhydrophobic energy equipment samples with different micromorphologies, obtain test parameters, and construct a water transport closed-loop correction model by combining incremental random forest algorithm, online learning mechanism and Bayesian optimization algorithm to obtain closed-loop correction index. The hydrophobicity index is then corrected, and a feedback closed-loop mechanism is set up to obtain the final hydrophobicity index based on the deviation between the corrected hydrophobicity index and the original hydrophobicity index. The hydrophobicity classification and output module, combining density peak clustering algorithm and K-means clustering algorithm, classifies the hydrophobicity level of the superhydrophobic energy equipment samples with different micromorphologies based on the final hydrophobicity index, and outputs the final hydrophobicity level through information push.

2. The hydrophobicity rating evaluation system for superhydrophobic surfaces of energy equipment according to claim 1, characterized in that, The process of acquiring the core parameter data of plasma etching includes: Different types of acquisition devices are deployed to collect the core parameters of the plasma etching process. The acquisition devices include a power meter, a pressure sensor, a flow meter, and a timer. The core parameters of the plasma etching process include the discharge power of the plasma etching equipment, the gas pressure in the plasma etching chamber, the gas flow rate into the etching chamber, and the plasma etching time. The collected discharge power of the plasma etching equipment, the gas pressure in the plasma etching chamber, the gas flow rate into the etching chamber, and the plasma etching time are subjected to data cleaning and data standardization processing.

3. The hydrophobicity rating evaluation system for superhydrophobic surfaces of energy equipment according to claim 2, characterized in that, The process of acquiring the particle microscopic data includes: The particle microscopic data includes particle temperature and particle density; A Langmuir probe was used to collect the probe current and probe voltage characteristic curves in the central region of the etching cavity, and the effective collection area of ​​the Langmuir probe was recorded. The wavelet transform algorithm is used to extract the electron saturation current segment from the probe current and probe voltage characteristic curves, the electron saturation current of the electron saturation current segment is selected, and the electron temperature of the electron saturation current segment is measured. Combined with the effective collection area of ​​the Langmuir probe, the electron current density is calculated. Based on the plasma probe diagnostic theory, the particle temperature and particle density are obtained through fitting calculation; Based on plasma probe diagnostic theory, a fitting model for fused particle energy distribution is established. The obtained particle microscopic data is used as the target variable. A genetic algorithm is employed to optimize the key parameters of the fitting model, minimize the deviation between theoretical and measured values, and achieve the synergistic inversion of particle temperature and density.

4. The hydrophobicity rating evaluation system for superhydrophobic surfaces of energy equipment according to claim 3, characterized in that, The process of establishing a correlation model based on the hybrid neural network architecture combining convolutional neural networks, attention mechanisms, and bidirectional long short-term memory networks, along with the core parameter data of plasma etching, to obtain the optimal parameter combination for the sample includes: The optimal parameter combination for the sample includes the optimal discharge power of the plasma etching equipment for superhydrophobic surface samples of energy equipment with different micromorphologies, the optimal gas pressure in the plasma etching chamber, the optimal gas flow rate into the etching chamber, and the optimal plasma etching time. The plasma etching core parameter data and the particle micro data are integrated, and the high-dimensional data are reduced in dimensionality by principal component analysis and divided into a first training set, a first test set and a first verification set. Using a hybrid neural network architecture that combines the convolutional neural network, attention mechanism, and bidirectional long short-term memory network, the first training set data is used as input, and the optimal parameter combination of the sample is used as output. The correlation between the core parameter data of plasma etching, the microscopic data of particles, and the optimal parameter combination of the sample is learned collaboratively to obtain an initial correlation model. The first validation set data is input into the initial association model, and the parameters of the initial association model are dynamically adjusted by combining the early stopping strategy and the adaptive learning rate adjustment algorithm to obtain the optimized association model. The first test set data is input into the optimized association model. The performance of the optimized association model is evaluated by the confusion matrix and mean square error. The hyperparameters of the association model are further fine-tuned by the Bayesian optimization algorithm to obtain the final association model. By combining the current plasma etching core parameter data and the particle microscopic data, the optimal parameter combination for the sample is output.

5. The hydrophobicity rating evaluation system for superhydrophobic surfaces of energy equipment according to claim 4, characterized in that, The process of deploying a microstructure digital twin module on the sample preparation module, constructing a three-dimensional simulation model of the plasma etching process, and combining the optimal parameter combination of the sample to prepare superhydrophobic energy equipment samples with different microstructures, and extracting surface feature parameters of the samples includes: The surface characteristic parameters of the samples include the structural height, spacing, roughness, static contact angle, and dynamic roll-off angle of the superhydrophobic energy equipment samples with different micromorphologies. A microstructure digital twin module is deployed on the sample preparation module to construct a three-dimensional simulation model of the plasma etching process. This enables multi-physics coupling modeling of superhydrophobic energy equipment samples with different micromorphologies. By adjusting the optimal parameter combination of the sample, the three-dimensional simulation model is optimized to find the theoretically optimal parameter combination of the sample, which is then used as the final optimal parameter combination of the sample. Based on the optimal parameter combination of the final sample, superhydrophobic energy equipment samples with different micromorphologies were prepared. The superhydrophobic surfaces of the prepared energy equipment samples with different micromorphologies were successively subjected to low surface energy modification and isothermal drying treatment to obtain the final superhydrophobic energy equipment samples with different micromorphologies. By combining scanning electron microscopy, white light interferometer, and optical contact angle measuring instrument, the structural height, spacing, roughness, static contact angle, and dynamic roll-off angle of the surface of the energy equipment superhydrophobic samples with different micromorphologies are collected, thereby realizing the extraction of the surface characteristic parameters of the samples.

6. The hydrophobicity rating evaluation system for superhydrophobic surfaces of energy equipment according to claim 5, characterized in that, The process of constructing a hydrophobicity quantitative evaluation model based on the surface characteristic parameters of the sample and outputting a hydrophobicity index includes: The micro-feature deconstruction unit, based on fractal geometry theory and fluid dynamics model, performs multi-scale decomposition of the surface feature parameters of the sample to extract the physical features of the microstructure. The cross-domain transfer learning adaptation unit introduces a meta-learning framework and a task adapter, and uses the physical features of the microstructure to train the hydrophobicity quantification evaluation model. The prediction result of the hydrophobicity quantification evaluation model is the hydrophobicity index. The real-time error compensation calibration unit, embedded with an online Bayesian inference algorithm, captures the measurement noise of surface characteristic parameters of superhydrophobic energy equipment samples with different micromorphologies in real time, dynamically outputs the error compensation coefficient, and corrects the prediction results of the hydrophobicity quantification evaluation model in real time to obtain the hydrophobicity index.

7. The hydrophobicity rating evaluation system for superhydrophobic surfaces of energy equipment according to claim 6, characterized in that, The process of testing the superhydrophobic energy equipment samples with different microstructures to obtain test parameters includes: The tests conducted on the superhydrophobic energy equipment samples with different microstructures include anti-icing performance tests and weathering resistance tests, and the test parameters include anti-icing index and weathering resistance index. An anti-icing performance test environment was set up, and the superhydrophobic energy equipment samples with different microstructures were placed in the anti-icing performance test environment. The freezing time of the surface of the superhydrophobic energy equipment samples with different microstructures under the anti-icing performance test environment was recorded by combining a thermal imager and a timer. The ice adhesion force of the surface of the superhydrophobic energy equipment samples with different microstructures under the anti-icing performance test environment was collected by an ice adhesion force tester. A weathering resistance test environment is set up, which includes a natural environment and an aging environment. Using a scanning electron microscope, a white light interferometer, and an optical contact angle meter, the surface characteristic parameters of the superhydrophobic energy equipment samples with different microstructures under the natural environment and the aging environment are collected. Combined with the hydrophobicity quantitative evaluation model, the hydrophobicity index of the superhydrophobic energy equipment samples with different microstructures under the natural environment and the aging environment is obtained. The freezing time and ice adhesion of the superhydrophobic energy equipment samples with different microstructures under the anti-icing performance test environment are standardized. Weights are assigned to the freezing time and ice adhesion of the superhydrophobic energy equipment samples with different microstructures under the standardized anti-icing performance test environment. The anti-icing index is calculated by weighted summation method. The weather resistance index is obtained by taking the proportion of the hydrophobicity index of the surface of the superhydrophobic energy equipment sample with different microstructures under the aging environment into the hydrophobicity index of the surface of the superhydrophobic energy equipment sample with different microstructures under the natural environment.

8. The hydrophobicity rating evaluation system for superhydrophobic surfaces of energy equipment according to claim 7, characterized in that, The process of constructing a water-conveying closed-loop correction model by combining incremental random forest algorithm, online learning mechanism and Bayesian optimization algorithm, obtaining closed-loop correction index, correcting the hydrophobicity index, setting up feedback closed-loop mechanism, and obtaining the final hydrophobicity index based on the deviation between the corrected hydrophobicity index and the original hydrophobicity index includes: The test parameters are integrated into a dataset, and the currently integrated dataset is divided into a third training set, a third validation set, and a third test set. Using the incremental random forest algorithm, the third training set is used as input data and the closed-loop correction index is used as output data. The nonlinear mapping relationship between the input data and the output data is learned through the ensemble learning framework to train the hydrophobic closed-loop correction model. At the same time, the test parameters are received in real time using the online learning mechanism to dynamically update the parameters of the hydrophobic closed-loop correction model, so as to obtain the trained water-conveying closed-loop correction model. The third validation set data is input into the trained hydrophobic closed-loop correction model. The performance of the hydrophobic closed-loop correction model is monitored by an early stopping strategy. The hyperparameters of the hydrophobic closed-loop correction model are adaptively optimized by combining the Bayesian optimization algorithm to obtain the optimized hydrophobic closed-loop correction model. The third test set data is input into the optimized hydrophobic closed-loop correction model to further evaluate the prediction accuracy and generalization ability of the hydrophobic closed-loop correction model. The parameters of the hydrophobic closed-loop correction model are adjusted to obtain the final high-precision water transport closed-loop correction model. Combined with the current test parameters, the corresponding closed-loop correction index is output. Based on the closed-loop correction index, the hydrophobicity index is corrected to obtain the corrected hydrophobicity index. A feedback closed-loop mechanism is set up, using the deviation between the corrected hydrophobicity index and the original hydrophobicity index as a reference, and scheduling the Bayesian optimization algorithm to perform closed-loop optimization on the hydrophobicity closed-loop correction model to obtain the final hydrophobicity index.

9. The hydrophobicity rating evaluation system for superhydrophobic surfaces of energy equipment according to claim 8, characterized in that, The process of classifying the hydrophobicity levels of the superhydrophobic energy equipment samples with different microstructures based on the final hydrophobicity index, using the combination of density peak clustering and K-means clustering algorithms, includes steps A1-A5: The hydrophobicity grades include weak hydrophobicity, medium hydrophobicity, high hydrophobicity, and super hydrophobicity. Step A1: The final hydrophobicity index is standardized, and the sample micromorphology feature weight and extreme working condition demand factor are introduced simultaneously to construct a multi-dimensional evaluation vector of hydrophobicity index, sample micromorphology weight and working condition factor. Combined with the elbow rule and profile coefficient for dual verification, the optimal k value is determined to be 4. The optimal k value corresponds to 4 hydrophobicity levels. Step A2: Analyze the density distribution of the multi-dimensional evaluation vector using the density peak clustering algorithm, select four standardized hydrophobicity indices as four cluster centers, and set an upper limit for the number of iterations and a cluster center convergence threshold. Step A3: Calculate the Euclidean distance from each standardized hydrophobicity index to the four cluster centers, and assign the superhydrophobic energy equipment sample corresponding to each hydrophobicity index to the category corresponding to the nearest cluster center; Step A4: Update each cluster center by taking the average of the hydrophobicity indices of all superhydrophobic energy equipment samples within the category corresponding to each cluster center. Step A5: Repeat steps A3 and A4, record the actual number of iterations. When the actual number of iterations reaches the set number of iterations and the cluster center meets the set cluster center convergence threshold, stop the iteration and determine the final cluster center. Based on the size of the final cluster center, classify the hydrophobicity level of the energy equipment superhydrophobic samples with different micromorphologies.

10. The hydrophobicity rating evaluation system for superhydrophobic surfaces of energy equipment according to claim 9, characterized in that, The hydrophobicity rating and output module outputs the final hydrophobicity rating via information push, including the following process: Each of the superhydrophobic energy equipment samples with different microstructures is assigned a number, and the number is associated with the final hydrophobicity index and hydrophobicity level. The number, hydrophobicity index and hydrophobicity level of the superhydrophobic energy equipment samples with different microstructures are integrated in the form of graphic and textual reports. The number, hydrophobicity index, and hydrophobicity level of the superhydrophobic energy equipment samples with different micromorphologies are sent to a preset receiving terminal via information push, PC pop-up window, mobile push, and email.

Images