A method for constructing a valve flow guide model and a method for applying a valve flow guide model

Abstract

This application relates to a method for constructing a valve flow conduction model and a method for applying the valve flow conduction model. The method for constructing a valve flow conduction model includes: acquiring multi-dimensional valve data; processing the multi-dimensional valve data through feature engineering to obtain target features; the multi-dimensional valve data includes air pressure and valve opening; the target features include air pressure and the air pressure standard deviation within a preset time period; acquiring a pre-defined mapping relationship between air pressure range, air pressure standard deviation range, and operating conditions; determining the operating condition corresponding to the target features based on the mapping relationship, air pressure, and air pressure standard deviation; training a gradient boosting decision tree algorithm using the target feature data under each operating condition as input data and the predicted valve opening as output data to obtain a valve flow conduction model corresponding to the operating condition, with each valve flow conduction model corresponding to a specific operating condition.

Description

Technical Field

[0001] This application relates to the field of flow conductance control, and in particular to a method for constructing a valve flow conductance model and a method for applying the valve flow conductance model. Background Technology

[0002] In fields such as vacuum systems, fluid control, and process industries, valves are key components whose flow characteristics directly affect system performance and control accuracy. Accurately establishing a flow model of a valve is crucial for system simulation, optimization design, and intelligent control.

[0003] Existing technologies face numerous bottlenecks in constructing valve flow conductance models. For instance, due to the complex internal structure and variable flow states of valves, traditional methods struggle to establish universally applicable mathematical models with clear physical mechanisms. Modeling methods relying on human experience or trial-and-error are cumbersome, highly dependent on expert knowledge, and inefficient due to long modeling cycles. Furthermore, when valves operate in vacuum or high-pressure systems, existing models cannot accurately cover all operating conditions as their flow states change, resulting in insufficient model accuracy and an inability to truly reflect actual flow conductance characteristics.

[0004] Therefore, there is an urgent need for a more efficient and accurate valve flow conduction model to provide reliable support for the research and development of high-end equipment and intelligent manufacturing. Summary of the Invention

[0005] This application provides a method for constructing a valve conduction model and a method for applying the valve conduction model, so as to at least solve the problem that the conduction model in related technologies is not accurate enough and cannot truly reflect the actual conduction characteristics.

[0006] In a first aspect, embodiments of this application provide a method for constructing a valve flow conduction model, including:

[0007] Acquire multi-dimensional valve data, and process the multi-dimensional valve data through feature engineering to obtain target features. The multi-dimensional valve data includes air pressure and valve opening. The target features include air pressure and the air pressure standard deviation within a preset time period.

[0008] Obtain the mapping relationship between the preset air pressure range, air pressure standard deviation range and operating conditions, and determine the operating conditions corresponding to the target feature based on the mapping relationship, air pressure and air pressure standard deviation;

[0009] Using the target feature data under each working condition as input data and the predicted valve opening as output data, a gradient boosting decision tree algorithm is trained to obtain the valve flow conduction model for the corresponding working condition. The valve flow conduction model corresponds one-to-one with the working condition.

[0010] In one embodiment, determining the operating condition corresponding to the target feature based on the mapping relationship, air pressure, and air pressure standard deviation includes:

[0011] The pressure range to which it belongs is determined based on the pressure.

[0012] The pressure control stage is determined based on the pressure standard deviation, which includes a pressure regulation stage and a stable stage. The pressure regulation stage and the stable stage include a pressure increase trend and a pressure decrease trend.

[0013] The corresponding operating conditions are determined based on the air pressure range and pressure control stage to which the target characteristic data belongs.

[0014] In one embodiment, the step of processing the multi-dimensional valve data through feature engineering to obtain the target features includes:

[0015] The original features are obtained by extracting features from the multi-dimensional valve data;

[0016] The target features are determined from the original features using forward feature selection and backward feature selection methods.

[0017] In one embodiment, acquiring multi-dimensional valve data includes:

[0018] Determine the valve opening measurement points during the process of the valve going from fully closed to fully open by using equally spaced points or key opening points;

[0019] The valve data at the valve opening measurement point is measured in multiple dimensions, including temperature, pressure, valve opening, and gas flow through the valve. Each parameter in the multiple dimensions of the valve data is time-series data.

[0020] The multi-dimensional valve data is stored in a structured format.

[0021] In one embodiment, before processing the multi-dimensional valve data through feature engineering to obtain the target features, the method further includes preprocessing the multi-dimensional valve data, the preprocessing process including:

[0022] The missing data in the multi-dimensional valve data is completed using the nearest neighbor completion method;

[0023] Noise smoothing is performed on the completed multi-dimensional valve data using a moving average filtering method.

[0024] Anomaly filtering is performed on the noise-smoothed multi-dimensional valve data using the interquartile range method.

[0025] Secondly, embodiments of this application provide a method for applying a valve flow conduction model, including:

[0026] Acquire target multi-dimensional valve data, and process the target multi-dimensional valve data through feature engineering to obtain target features. The multi-dimensional valve data includes air pressure and valve opening. The target features include air pressure and air pressure standard deviation within a preset time period.

[0027] Obtain the mapping relationship between the preset air pressure range, air pressure standard deviation range and operating conditions, and determine the target operating conditions corresponding to the target characteristics based on the mapping relationship, air pressure and air pressure standard deviation;

[0028] The target valve flow conduction model corresponding to the target operating condition is determined from the valve flow conduction model constructed by the method described in the first aspect, and the target feature is processed by the target valve flow conduction model to obtain the predicted valve opening.

[0029] In one embodiment, when the target operating condition changes, the predicted valve opening is obtained by processing the target features using the target valve flow conductance model, including:

[0030] The first predicted valve opening is obtained by directly processing the target features based on the target valve flow conductance model corresponding to the changed target operating condition; or,

[0031] Based on the target valve flow conduction model corresponding to the target operating conditions before and after the change, the target features are processed to obtain the second predicted valve opening and the third predicted valve opening. The second predicted valve opening and the third predicted valve opening are then weighted and fused to obtain the fourth predicted valve opening.

[0032] Thirdly, embodiments of this application provide a system for constructing a valve flow conduction model, including:

[0033] Acquisition module: used to acquire multi-dimensional valve data, and process the multi-dimensional valve data through feature engineering to obtain target features. The multi-dimensional valve data includes air pressure and valve opening degree, and the target features include air pressure and air pressure standard deviation within a preset time period.

[0034] Mapping module: used to obtain the mapping relationship between the preset air pressure range, air pressure standard deviation range and operating conditions, and to determine the operating conditions corresponding to the target feature based on the mapping relationship, air pressure and air pressure standard deviation;

[0035] Model module: Used to train a gradient boosting decision tree algorithm with target feature data under each working condition as input data and valve prediction opening as output data to obtain the valve flow conduction model for the corresponding working condition. The valve flow conduction model corresponds one-to-one with the working condition.

[0036] Fourthly, embodiments of this application provide an application system for a valve flow conduction model, including:

[0037] Data processing module: used to acquire target multi-dimensional valve data, and process the target multi-dimensional valve data through feature engineering to obtain target features. The multi-dimensional valve data includes air pressure and valve opening. The target features include air pressure and air pressure standard deviation within a preset time period.

[0038] Operating condition mapping module: used to obtain the mapping relationship between the preset air pressure range, air pressure standard deviation range and operating conditions, and determine the target operating condition corresponding to the target feature based on the mapping relationship, air pressure and air pressure standard deviation;

[0039] Prediction module: used to determine the target valve flow conduction model corresponding to the target operating condition from the valve flow conduction model constructed by the method described in the first aspect, and to process the target features through the target valve flow conduction model to obtain the predicted valve opening.

[0040] Fifthly, embodiments of this application provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements a method for constructing a valve flow conduction model as described in the first aspect and a method for applying a valve flow conduction model as described in the second aspect.

[0041] The valve flow conduction model construction method and valve flow conduction model application method provided in this application embodiment have at least the following technical effects.

[0042] This application constructs corresponding valve conductance models for different operating conditions, forming a multimodal conductance model system. This solves the problem of insufficient model accuracy and inability to accurately predict valve opening caused by the complex internal structure and variable flow states of valves, effectively improving the model's prediction accuracy. Through feature engineering methods, the core variables affecting conductance performance are intelligently identified, ensuring that the model calculation only involves these core variables, reducing the number of computational parameters, accelerating the calculation process, and significantly shortening modeling time while maintaining physical consistency.

[0043] Details of one or more embodiments of this application are set forth in the following drawings and description to make other features, objects and advantages of this application more readily apparent. Attached Figure Description

[0044] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0045] Figure 1 This is a flowchart illustrating a method for constructing a valve flow conduction model according to an embodiment of this application;

[0046] Figure 2 This is a schematic diagram illustrating an experimental derivation model and a theoretical model according to an exemplary embodiment;

[0047] Figure 3 This is a flowchart illustrating an application method of a valve flow conduction model according to an embodiment of this application;

[0048] Figure 4 This is a structural block diagram of a valve flow conduction model construction system according to an embodiment of this application;

[0049] Figure 5 This is a structural block diagram of an application system for a valve flow conduction model according to an embodiment of this application;

[0050] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of this application clearer, the application is described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments provided in this application without inventive effort are within the scope of protection of this application.

[0052] Obviously, the accompanying drawings described below are merely some examples or embodiments of this application. Those skilled in the art can apply this application to other similar scenarios based on these drawings without any inventive effort. Furthermore, it is understood that although the efforts made in this development process may be complex and lengthy, for those skilled in the art related to the content disclosed in this application, any changes to design, manufacturing, or production based on the technical content disclosed in this application are merely conventional technical means and should not be construed as insufficient disclosure of the content of this application.

[0053] In this application, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment that is mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this application may be combined with other embodiments without conflict.

[0054] Unless otherwise defined, the technical or scientific terms used in this application shall have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms “a,” “an,” “an,” “the,” and similar words used in this application do not indicate quantity limitation and may indicate singular or plural. The terms “comprising,” “including,” “having,” and any variations thereof used in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units not listed, or may include other steps or units inherent to these processes, methods, products, or devices. The terms “connected,” “linked,” “coupled,” and similar words used in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. “Multiple” used in this application refers to two or more. “And / or” describes the relationship between related objects, indicating that three relationships may exist; for example, “A and / or B” can represent: A alone, A and B simultaneously, and B alone. The character " / " generally indicates that the preceding and following objects are in an "or" relationship. The terms "first," "second," and "third" used in this application are merely to distinguish similar objects and do not represent a specific ordering of the objects.

[0055] Firstly, embodiments of this application provide a method for constructing a valve flow conduction model. Figure 1 This is a flowchart illustrating a method for constructing a valve flow conduction model according to an embodiment of this application, such as... Figure 1 As shown, the method includes:

[0056] Step S101: Obtain multi-dimensional valve data, and process the multi-dimensional valve data through feature engineering to obtain target features. The multi-dimensional valve data includes air pressure and valve opening degree, and the target features include air pressure and air pressure standard deviation within a preset time period.

[0057] Optionally, it integrates multi-source information such as fluid experimental test data and historical operating data to automatically collect key parameters of valves, such as flow rate, air pressure, temperature, and opening degree. It then utilizes data preprocessing and feature engineering to intelligently identify core variables affecting flow conductance performance. This reduces manual intervention and improves modeling efficiency.

[0058] In one example, step S101 involves acquiring multi-dimensional valve data, including:

[0059] Step S1011: Determine the valve opening measurement points during the process of the valve going from fully closed to fully open by using equally spaced points or key opening points.

[0060] Optional, Figure 2This is illustrated by an exemplary embodiment, which describes a method of gradually adjusting a valve from fully closed (0%) to fully open (100%), and collecting data after maintaining a stable state at each opening position. Measurements are taken at equal intervals or key opening points to ensure that data acquisition covers the entire operating range.

[0061] Step S1012: Measure the multi-dimensional valve data at the valve opening measurement point. The multi-dimensional valve data includes temperature, pressure, valve opening, and gas flow through the valve. Each parameter in the multi-dimensional valve data is time-series data.

[0062] Optional multi-dimensional valve data measurements include: ambient and system temperature (measuring the valve's operating environment temperature); internal cavity pressure (measuring the pressure inside the cavity controlled by the valve); vacuum pump pressure (measuring the working pressure of the vacuum pump); real-time valve opening (precisely measuring the current valve opening (percentage or angle)); and gas flow through the valve (measuring the gas flow through the valve). All measurement data are time-series data, recording the complete process of multi-dimensional valve data changes over time.

[0063] Step S1013: Store the multi-dimensional valve data in a structured format.

[0064] Optionally, the collected multidimensional data can be stored in a structured format to ensure the time synchronization and consistency of the data, thus establishing a complete dataset for subsequent analysis.

[0065] In this way, multi-source information such as fluid experimental test data and historical operation data are integrated, and key parameters such as valve flow rate, air pressure, temperature and opening degree are automatically collected, reducing manual intervention, improving data acquisition efficiency, and helping to improve modeling efficiency.

[0066] In one example, step S101 includes:

[0067] Step S1014: Extract features from the multi-dimensional valve data to obtain the original features.

[0068] Optionally, feature extraction is performed on the multi-dimensional valve data, extracting statistical features from the original multi-dimensional time series data, including: mean, variance, maximum value, minimum value, first gradient (rate of change), second gradient (acceleration of change), Fourier transform coefficients, correlation coefficients between variables, etc., extracting a total of thousands of original features.

[0069] Step S1015: Determine the target features from the original features using forward feature selection and backward feature selection methods.

[0070] Optionally, a forward feature selection method can be used to start from an empty feature set and gradually add the most important features, while a backward feature selection method can be used to start from a complete feature set and gradually remove the least important features. By combining the two feature selection methods, the optimal feature subset can be found. In this way, a target number of the most predictive features can be selected from thousands of features. The target number can be set according to actual needs. Generally, the target number of features is positively correlated with model accuracy. In this application, the target number is set to 100.

[0071] In this way, through feature engineering, the core variables affecting flow conductance performance are intelligently identified, ensuring that the model calculations only involve these core variables, reducing the number of computational parameters and improving computational efficiency. Furthermore, the feature engineering process significantly reduces manual intervention, improving the efficiency and accuracy of data processing.

[0072] In one example, before processing the multi-dimensional valve data through feature engineering to obtain the target features, the method also includes preprocessing the multi-dimensional valve data. The preprocessing process includes:

[0073] Missing data in multi-dimensional valve data is completed using the nearest neighbor completion method.

[0074] The noise smoothing process for the completed multi-dimensional valve data is performed using a moving average filtering method.

[0075] Anomaly filtering is performed on the noise-smoothed multi-dimensional valve data using the interquartile range method.

[0076] Optionally, missing values ​​can be filled with data from adjacent time points using the nearest neighbor completion method; random noise in the data can be smoothed using the moving average filtering technique; and outlier data points can be identified and removed using the interquartile range method. In this way, the acquired multi-dimensional valve data is preprocessed automatically, and outlier data is filtered and completed, which improves data quality and does not rely on manual intervention, thus helping to improve the accuracy of subsequent model training.

[0077] Step S102: Obtain the mapping relationship between the preset air pressure range, air pressure standard deviation range and operating conditions, and determine the operating conditions corresponding to the target characteristics based on the mapping relationship, air pressure and air pressure standard deviation.

[0078] Optionally, considering that the flow state of a valve changes significantly with pressure range and pressure control stage when operating in a vacuum or high-pressure system, a single model cannot accurately cover all operating conditions. Therefore, this application trains corresponding models for different operating conditions. Data sets adapted to the physical mechanisms of each operating condition are divided based on air pressure and air pressure standard deviation, and sub-models adapted to the physical mechanisms of each operating condition are trained based on the datasets for each operating condition. The pressure range is directly divided according to air pressure. The pressure control stage is judged based on the fluctuation of the air pressure standard deviation. If the air pressure standard deviation is less than a certain value, it indicates that a stable stage has been reached; otherwise, it is still in the pressure regulation stage. The division of the pressure range can be set according to actual application requirements.

[0079] In one example, step S102 includes: determining the pressure range to which the target characteristic data belongs based on the air pressure; determining the pressure control stage to which it belongs based on the air pressure standard deviation, wherein the pressure control stage includes a pressure adjustment stage and a stabilization stage, and the pressure adjustment stage and the stabilization stage include a pressure rise trend and a pressure fall trend; and determining the corresponding operating condition based on the air pressure range and pressure control stage to which the target characteristic data belongs.

[0080] Optionally, the pressure range can be divided directly based on atmospheric pressure. For example, atmospheric pressure can be divided into four ranges: 0-1T, 1-10T, 10-100T, and 100-1000T (T: Torr). The pressure control phase is determined by the standard deviation of atmospheric pressure. If the standard deviation is less than a preset value, it indicates that a stable phase has been reached; otherwise, it remains in the pressure regulation phase. For example, the pressure control phase can be divided into four cases: (pressure increase, pressure decrease) * (pressure regulation, stable).

[0081] In this way, considering the significant changes in the flow state of valves when they are working in vacuum or high-pressure systems with varying pressure ranges and pressure control stages, and the difficulty of accurately covering all operating conditions with a single model, different operating conditions are divided according to pressure and pressure standard deviation, and the physical phenomena of different situations are modeled separately, which helps to improve the accuracy of the model.

[0082] Step S103: Using the target feature data under each working condition as input data and the predicted valve opening as output data, train the gradient boosting decision tree algorithm to obtain the valve flow conduction model for the corresponding working condition. The valve flow conduction model corresponds one-to-one with the working condition.

[0083] Optionally, a Gradient Boosting Decision Tree (GBDT) is chosen as the training model. GBDT has strong capabilities in handling nonlinear relationships, good ability to capture feature interactions, and some robustness to outliers, and can provide feature importance ranking. During model training, 100 selected optimal features are used as input to the model for each operating condition, and the predicted valve opening is used as the output. These optimal features include statistical and variable characteristics across multiple dimensions, such as temperature, air pressure, and flow rate. During training, the dataset is divided into training, validation, and test sets. Cross-validation is used to optimize the model's hyperparameters, and mean squared error (MSE), mean absolute error (MAE), and coefficient of determination (R²) are used as evaluation metrics. Furthermore, after model training, the model's generalization ability is validated on an independent test set. Based on the validation results, the feature set and model parameters are adjusted to obtain the valve flow guidance model for the corresponding operating condition.

[0084] This approach solves the problem of insufficient model accuracy and inability to accurately predict valve opening caused by the complex internal structure and variable flow states of valves. By constructing sub-models for different pressure ranges and pressure control stages, a multimodal flow conductance model system is formed, which effectively improves the model's prediction accuracy, accurately predicts valve opening, and provides a reference for valve control.

[0085] Figure 2 This is a schematic diagram illustrating the experimental derivation model and the theoretical model according to an exemplary embodiment, such as... Figure 2 As shown, the experimental derivation model is highly consistent with the theoretical model, and the curves almost overlap, indicating that the experimental data has extremely high accuracy and reliability. The mean square error between the two is extremely small, indicating that the experimental process has low noise and the system behavior is close to the ideal state.

[0086] In summary, this application utilizes a valve conductance modeling method based on multi-dimensional data automatic acquisition and fusion to achieve intelligent identification and automated modeling of key conductance parameters. Furthermore, it constructs corresponding sub-models for different pressure ranges and pressure control stages, forming a multi-modal conductance model system. This solves the problem of insufficient model accuracy and inability to accurately predict valve opening caused by the complex internal structure and variable flow states of valves, effectively improving the model's prediction accuracy. By combining simulation and machine learning, and employing feature engineering methods, the core variables affecting conductance performance are intelligently identified, ensuring that model calculations only involve these core variables, reducing the number of computational parameters, accelerating the calculation process, and significantly shortening modeling time while maintaining physical consistency. Moreover, by fusing experimental data and simulation results for closed-loop correction, the model's ability to reproduce actual flow characteristics is effectively improved, ensuring high accuracy in conductance prediction and further enhancing modeling efficiency and accuracy.

[0087] Secondly, embodiments of this application provide a method for applying a valve flow conduction model. Figure 3 This is a flowchart illustrating an application method of a valve flow conduction model according to an embodiment of this application, such as... Figure 3 As shown, the method includes:

[0088] Step S201: Obtain target multi-dimensional valve data, and process the target multi-dimensional valve data through feature engineering to obtain target features. The multi-dimensional valve data includes air pressure and valve opening degree, and the target features include air pressure and air pressure standard deviation within a preset time period.

[0089] Optionally, it integrates multi-source information such as fluid experimental test data and historical operating data to automatically collect key parameters of valves, such as flow rate, air pressure, temperature, and opening degree. It then utilizes data preprocessing and feature engineering to intelligently identify core variables affecting flow conductance performance. This reduces manual intervention and improves modeling efficiency.

[0090] Step S202: Obtain the pre-set mapping relationship between air pressure range, air pressure standard deviation range and operating conditions, and determine the target operating conditions corresponding to the target characteristics based on the mapping relationship, air pressure and air pressure standard deviation.

[0091] Optionally, considering that the flow state of a valve changes significantly with pressure range and pressure control stage when operating in a vacuum or high-pressure system, a single model cannot accurately cover all operating conditions. Therefore, this application trains corresponding models for different operating conditions. Data sets adapted to the physical mechanisms of each operating condition are divided based on air pressure and air pressure standard deviation, and sub-models adapted to the physical mechanisms of each operating condition are trained based on the datasets for each operating condition. The pressure range is directly divided according to air pressure. The pressure control stage is judged based on the fluctuation of the air pressure standard deviation. If the air pressure standard deviation is less than a certain value, it indicates that a stable stage has been reached; otherwise, it is still in the pressure regulation stage. The division of the pressure range can be set according to actual application requirements.

[0092] Step S203: Determine the target valve flow conduction model corresponding to the target operating condition from the valve flow conduction model constructed by the method in the first aspect, and obtain the predicted valve opening by processing the target features through the target valve flow conduction model.

[0093] Optionally, for the acquired multi-dimensional target valve data, the target features are processed according to the target valve flow conductance model corresponding to the data to predict the valve opening. In this way, by adapting a sub-model to the physical mechanism of the target operating condition, high-precision prediction across the entire domain can be achieved.

[0094] In one example, when the target operating condition changes, step S203 includes: directly processing the target features according to the target valve flow conduction model corresponding to the changed target operating condition to obtain the first predicted valve opening.

[0095] In another example, when the target operating condition changes, step S203 includes: processing the target features according to the target valve flow conduction model corresponding to the target operating conditions before and after the change to obtain the second predicted valve opening and the third predicted valve opening, and performing weighted fusion processing on the second predicted valve opening and the third predicted valve opening to obtain the fourth predicted valve opening.

[0096] Optionally, the weighted fusion processing adopted in this application refers to switching the corresponding target valve conductance model when the operating conditions change. During the model switching process, linear interpolation weights are used to smooth the influence of the two models. The closer to the previous model, the greater the weight of the previous model; otherwise, the smaller the weight. The closer to the next model, the greater the weight of the next model; otherwise, the smaller the weight. In this way, the model can cover all operating conditions, improve the model's accuracy, and simultaneously enhance the model's physical interpretability and generalization ability.

[0097] In summary, this application utilizes a valve flow conductance modeling method based on multi-dimensional data automatic acquisition and fusion to achieve intelligent identification and automated modeling of key flow conductance parameters. Furthermore, it constructs corresponding sub-models for different pressure ranges and pressure control stages, forming a multi-modal flow conductance model system. This solves the problem of insufficient model accuracy and inability to accurately predict valve opening caused by the complex internal structure and variable flow states of valves. Through dynamic switching or weighted fusion mechanisms of the multi-modal models, the model can cover all operating conditions, improving model accuracy, physical interpretability, and generalization ability.

[0098] Thirdly, embodiments of this application provide a system for constructing a valve flow conduction model. Figure 4 This is a structural block diagram of a valve flow conduction model construction system according to an embodiment of this application, as shown in the following example. Figure 4 As shown, the system includes:

[0099] Acquisition Module 100: Used to acquire multi-dimensional valve data. The multi-dimensional valve data is processed through feature engineering to obtain target features. The multi-dimensional valve data includes air pressure and valve opening. The target features include air pressure and the standard deviation of air pressure within a preset time period.

[0100] Mapping module 200: used to obtain the mapping relationship between the preset air pressure range, air pressure standard deviation range and operating conditions, and to determine the operating conditions corresponding to the target characteristics based on the mapping relationship, air pressure and air pressure standard deviation.

[0101] Model Module 300: Used to train a gradient boosting decision tree algorithm with target feature data under each working condition as input data and valve prediction opening as output data to obtain the valve flow conduction model for the corresponding working condition. The valve flow conduction model corresponds one-to-one with the working condition.

[0102] In one example, mapping module 200 includes:

[0103] Used to determine the pressure range to which it belongs based on air pressure.

[0104] The pressure control phase is determined by the standard deviation of air pressure. The pressure control phase includes the pressure regulation phase and the stable phase. The pressure regulation phase and the stable phase include the pressure rising trend and the pressure falling trend.

[0105] The corresponding operating conditions are determined based on the air pressure range and pressure control stage to which the target characteristic data belongs.

[0106] In one example, the acquisition module 100 includes: extracting features from multi-dimensional valve data to obtain raw features; and determining target features from the raw features using forward feature selection and backward feature selection methods.

[0107] In one example, obtaining module 100 includes:

[0108] Used to determine the valve opening measurement points during the process of a valve going from fully closed to fully open by using equally spaced points or key opening points.

[0109] The valve data is measured at multiple dimensions, including temperature, pressure, valve opening, and gas flow through the valve. Each parameter in the multiple dimensions valve data is time-series data.

[0110] Store multi-dimensional valve data in a structured format.

[0111] In one example, before processing the multi-dimensional valve data through feature engineering to obtain the target features, the system also includes preprocessing the multi-dimensional valve data. The preprocessing process includes:

[0112] Missing data in multi-dimensional valve data is completed using the nearest neighbor completion method.

[0113] The noise smoothing process for the completed multi-dimensional valve data is performed using a moving average filtering method.

[0114] Anomaly filtering is performed on the noise-smoothed multi-dimensional valve data using the interquartile range method.

[0115] In summary, this application utilizes a valve conductance modeling method based on multi-dimensional data automatic acquisition and fusion to achieve intelligent identification and automated modeling of key conductance parameters. Furthermore, it constructs corresponding sub-models for different pressure ranges and pressure control stages, forming a multi-modal conductance model system. This solves the problem of insufficient model accuracy and inability to accurately predict valve opening caused by the complex internal structure and variable flow states of valves, effectively improving the model's prediction accuracy. By combining simulation and machine learning, and employing feature engineering methods, the core variables affecting conductance performance are intelligently identified, ensuring that model calculations only involve these core variables, reducing the number of computational parameters, accelerating the calculation process, and significantly shortening modeling time while maintaining physical consistency. Moreover, by fusing experimental data and simulation results for closed-loop correction, the model's ability to reproduce actual flow characteristics is effectively improved, ensuring high accuracy in conductance prediction and further enhancing modeling efficiency and accuracy.

[0116] Fourthly, embodiments of this application provide an application system for a valve flow conduction model. Figure 5 This is a structural block diagram of an application system for a valve flow conduction model according to an embodiment of this application, such as... Figure 5 As shown, the system includes:

[0117] Data processing module 400: Used to acquire target multi-dimensional valve data, and process the target multi-dimensional valve data through feature engineering to obtain target features. The multi-dimensional valve data includes air pressure and valve opening degree, and the target features include air pressure and air pressure standard deviation within a preset time period.

[0118] Operating condition mapping module 500: used to obtain the mapping relationship between the preset air pressure range, air pressure standard deviation range and operating conditions, and to determine the target operating condition corresponding to the target feature based on the mapping relationship, air pressure and air pressure standard deviation.

[0119] Prediction module 600: used to determine the target valve flow conduction model corresponding to the target operating condition from the valve flow conduction model constructed by the first aspect method, and to obtain the predicted valve opening by processing the target features through the target valve flow conduction model.

[0120] In one embodiment, when the target operating condition changes, the prediction module 600 includes: processing the target features directly based on the target valve flow conduction model corresponding to the changed target operating condition to obtain a first predicted valve opening.

[0121] In another embodiment, when the target operating condition changes, the prediction module 600 includes: processing the target features according to the target valve flow conduction model corresponding to the target operating conditions before and after the change to obtain a second predicted valve opening and a third predicted valve opening, and performing weighted fusion processing on the second predicted valve opening and the third predicted valve opening to obtain a fourth predicted valve opening.

[0122] In summary, this application utilizes a valve flow conductance modeling method based on multi-dimensional data automatic acquisition and fusion to achieve intelligent identification and automated modeling of key flow conductance parameters. Furthermore, it constructs corresponding sub-models for different pressure ranges and pressure control stages, forming a multi-modal flow conductance model system. This solves the problem of insufficient model accuracy and inability to accurately predict valve opening caused by the complex internal structure and variable flow states of valves. Through dynamic switching or weighted fusion mechanisms of the multi-modal models, the model can cover all operating conditions, improving model accuracy, physical interpretability, and generalization ability.

[0123] Fifthly, embodiments of this application provide an electronic device, Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements a method for constructing a valve flow conduction model according to the first aspect and a method for applying the valve flow conduction model according to the second aspect. Figure 6 The electronic device 60 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0124] Electronic device 60 may be in the form of a general-purpose computing device, such as a server device. Components of electronic device 60 may include, but are not limited to: at least one processor 61, at least one memory 62, and a bus 63 connecting different system components (including memory 62 and processor 61).

[0125] Bus 63 includes a data bus, an address bus, and a control bus.

[0126] The memory 62 may include volatile memory, such as random access memory (RAM) 621 and / or cache memory 622, and may further include read-only memory (ROM) 623.

[0127] The memory 62 may also include a program / utility 625 having a set (at least one) of program modules 624, including but not limited to: an operating system, one or more application programs, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment.

[0128] The processor 61 executes various functional applications and data processing by running computer programs stored in the memory 62, such as a method for constructing a valve flow conduction model according to the first aspect of this application and a method for applying a valve flow conduction model according to the second aspect.

[0129] Electronic device 60 can also communicate with one or more external devices 64 (e.g., keyboard, pointing device, etc.). This communication can be performed via input / output (I / O) interface 65. Furthermore, the model-generated electronic device 60 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 66. Figure 6 As shown, network adapter 66 communicates with other modules of the model-generated electronic device 60 via bus 63. It should be understood that, although... Figure 6 As not shown, the electronic device 60 generated in conjunction with the model may use other hardware and / or software modules, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID (disk array) systems, tape drives, and data backup storage systems.

[0130] It should be noted that although several units / modules or sub-units / modules of the electronic device have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of the present invention, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided and embodied by multiple units / modules.

[0131] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0132] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for constructing a valve flow conduction model, characterized in that, include: Acquire multi-dimensional valve data, and process the multi-dimensional valve data through feature engineering to obtain target features, including: extracting features from the multi-dimensional valve data to obtain original features; determining target features from the original features using forward feature selection and backward feature selection methods; the multi-dimensional valve data includes air pressure and valve opening, and the target features include air pressure and the air pressure standard deviation within a preset time period; Obtain the mapping relationship between a pre-set air pressure range, an air pressure standard deviation range, and operating conditions; determine the operating conditions corresponding to the target characteristic based on the mapping relationship, air pressure, and air pressure standard deviation, including: The pressure range to which it belongs is determined based on the pressure. The pressure control stage is determined based on the pressure standard deviation. The pressure control stage includes a pressure adjustment stage and a stable stage. The pressure adjustment stage and the stable stage include a pressure increase trend and a pressure decrease trend. If the pressure standard deviation is less than a preset value, it is considered that the stable stage has been reached; otherwise, it is still in the pressure adjustment stage. The corresponding operating conditions are determined based on the air pressure range and pressure control stage to which the target characteristic data belongs; Using the target feature data under each working condition as input data and the predicted valve opening as output data, a gradient boosting decision tree algorithm is trained to obtain a valve flow conduction model for the corresponding working condition, which uses the predicted valve opening to characterize the fluid conduction capacity. The valve flow conduction model corresponds one-to-one with the working condition.

2. The method for constructing a valve flow conduction model according to claim 1, characterized in that, The acquisition of multi-dimensional valve data includes: Determine the valve opening measurement points during the process of the valve going from fully closed to fully open by using equally spaced points or key opening points; The valve data at the valve opening measurement point is measured in multiple dimensions, including temperature, pressure, valve opening, and gas flow through the valve. Each parameter in the multiple dimensions of the valve data is time-series data. The multi-dimensional valve data is stored in a structured format.

3. The method for constructing a valve flow conduction model according to claim 1, characterized in that, Before processing the multi-dimensional valve data through feature engineering to obtain the target features, the method further includes preprocessing the multi-dimensional valve data, the preprocessing process including: The missing data in the multi-dimensional valve data is completed using the nearest neighbor completion method; Noise smoothing is performed on the completed multi-dimensional valve data using a moving average filtering method. Anomaly filtering is performed on the noise-smoothed multi-dimensional valve data using the interquartile range method.

4. A method for applying a valve flow conduction model, characterized in that, include: Acquire target multi-dimensional valve data, and process the target multi-dimensional valve data through feature engineering to obtain target features. The multi-dimensional valve data includes air pressure and valve opening. The target features include air pressure and air pressure standard deviation within a preset time period. Obtain the mapping relationship between the preset air pressure range, air pressure standard deviation range and operating conditions, and determine the target operating conditions corresponding to the target characteristics based on the mapping relationship, air pressure and air pressure standard deviation; The target valve flow conduction model corresponding to the target operating condition is determined from the valve flow conduction model constructed by the method of any one of claims 1 to 3, and the target feature is processed by the target valve flow conduction model to obtain the predicted valve opening.

5. The application method of the valve flow conduction model according to claim 4, characterized in that, When the target operating condition changes, the predicted valve opening is obtained by processing the target features using the target valve flow conductance model, including: The first predicted valve opening is obtained by directly processing the target features based on the target valve flow conductance model corresponding to the changed target operating condition; or, Based on the target valve flow conduction model corresponding to the target operating conditions before and after the change, the target features are processed to obtain the second predicted valve opening and the third predicted valve opening. The second predicted valve opening and the third predicted valve opening are then weighted and fused to obtain the fourth predicted valve opening.

6. A system for constructing a valve flow conduction model, characterized in that, include: Acquisition Module: Used to acquire multi-dimensional valve data, and process the multi-dimensional valve data through feature engineering to obtain target features, including: extracting features from the multi-dimensional valve data to obtain original features; determining target features from the original features through forward feature selection and backward feature selection methods; the multi-dimensional valve data includes air pressure and valve opening, and the target features include air pressure and the air pressure standard deviation within a preset time period; Mapping module: Used to acquire the mapping relationship between a pre-set pressure range, a pressure standard deviation range, and operating conditions; and to determine the operating condition corresponding to the target characteristic based on the mapping relationship, pressure, and pressure standard deviation, including: The pressure range to which it belongs is determined based on the pressure. The pressure control stage is determined based on the pressure standard deviation. The pressure control stage includes a pressure adjustment stage and a stable stage. The pressure adjustment stage and the stable stage include a pressure increase trend and a pressure decrease trend. If the pressure standard deviation is less than a preset value, it is considered that the stable stage has been reached; otherwise, it is still in the pressure adjustment stage. The corresponding operating conditions are determined based on the air pressure range and pressure control stage to which the target characteristic data belongs; Model module: Used to train gradient boosting decision tree algorithm with target feature data under each working condition as input data and valve prediction opening as output data, to obtain valve flow conduction model for the corresponding working condition, which represents the fluid conduction capacity with valve prediction opening. The valve flow conduction model corresponds one-to-one with the working condition.

7. An application system for a valve flow conduction model, characterized in that, include: Data processing module: used to acquire target multi-dimensional valve data, and process the target multi-dimensional valve data through feature engineering to obtain target features. The multi-dimensional valve data includes air pressure and valve opening. The target features include air pressure and air pressure standard deviation within a preset time period. Operating condition mapping module: used to obtain the mapping relationship between the preset air pressure range, air pressure standard deviation range and operating conditions, and determine the target operating condition corresponding to the target feature based on the mapping relationship, air pressure and air pressure standard deviation; Prediction module: used to determine the target valve flow conduction model corresponding to the target operating condition from the valve flow conduction model constructed by the method of any one of claims 1 to 3, and to process the target features through the target valve flow conduction model to obtain the predicted valve opening.

8. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement a method for constructing a valve flow conduction model as described in any one of claims 1 to 3 and a method for applying a valve flow conduction model as described in any one of claims 4 to 5.

Images