Hydraulic element machining decision-making system and method based on neural network model

Abstract

The invention relates to the technical field of data processing, in particular to a hydraulic element processing decision-making system and method based on a neural network model, and the system comprises a data loading module which is used for loading a process data table of a database, processing null values of a specific column in the process data table, and obtaining and storing a processed data table; the model training module is used for training a neural network model based on the processed data table; the model loading module is used for loading the trained model if the existence of the trained model is detected; the decision module is used for predicting a processing decision based on the trained model and the obtained numerical processing features; on the basis of the prediction result and a preset mapping relation, determining predicted step content; and screening and displaying target data in the processed data table based on the text decision content and the predicted step content. According to the system, the machining precision of the hydraulic assembly can be improved, the manufacturing quality of hydraulic key parts is ensured, repeated labor can be effectively reduced, and the time cost is greatly reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a hydraulic component processing decision system and method based on a neural network model in the technical field of data processing. Background Art

[0002] In the field of hydraulic component manufacturing, the precise setting of processing parameters is a key link in ensuring product quality and performance. In related technologies, the processing of hydraulic components relies on experienced engineers to adjust processing parameters based on past experience. Although this method can meet basic production needs, relying on manual experience to adjust parameters usually requires a long trial and error cycle, which not only causes repetitive work and increases production time, but may also lead to reduced production efficiency. Moreover, it is difficult to achieve high-precision requirements by manually adjusting parameters, especially when faced with hydraulic components with complex geometries or special material properties, the accuracy of parameter settings will be limited. The trial and error method not only increases material waste, but may also lead to additional rework and repair costs. Summary of the Invention

[0003] The purpose of the present invention is to provide a hydraulic component processing decision system and method based on a neural network model. The technical solutions adopted are as follows:

[0004] In a first aspect, an embodiment of the present invention provides a hydraulic component processing decision system based on a neural network model, the system comprising:

[0005] A data loading module is used to load a process data table from a database, process null values in a specific column of the process data table, and obtain and store a processed data table;

[0006] A model training module, configured to perform back propagation training on a neural network model based on the processed data table to obtain a trained model;

[0007] A model loading module, configured to load the trained model from a specified path if the trained model is detected to exist;

[0008] A decision module is used to obtain the input numerical processing features for the hydraulic component; perform processing decision prediction based on the trained model and the numerical processing features to obtain a prediction result; and determine the predicted work step content corresponding to the processed data table based on the prediction result and a preset mapping relationship; based on the input text decision content for the hydraulic component and the predicted work step content, filter and display target data in the processed data table.

[0009] In some possible implementations, the data loading module is further configured to connect the decision system to a database and load the process data table from the database.

[0010] In some possible implementations, the database is used to store complete process data tables of multiple hydraulic components;

[0011] The data loading module is further used to perform a test connection between the database and the decision system based on a structured query connection statement.

[0012] In some possible implementations, the data loading module is further configured to set the null value of the specific column to be zero, and store each processed data table in a data list library.

[0013] In some possible implementations, the model training module is also used to use the column data belonging to the work step content in the processed data table as the output label, and the data other than the column data belonging to the work step content in the processed data table as the input feature, and normalize the input features and perform one-hot encoding on the output label to build a neural network model including an input layer, a hidden layer and an output layer.

[0014] In some possible implementations, the model loading module is further configured to, if the existence of the trained model is detected, deserialize the trained model, restore the structure and parameters of the trained model, and obtain a restored model; obtain and save the number of neurons in the input layer and output layer of the trained model;

[0015] The decision module is further used to make a processing decision prediction based on the numerical processing features and the number of neurons in the input layer and the output layer to obtain the prediction result.

[0016] In some possible implementations, the model loading module is further configured to return to the model training module if it is detected that the trained model does not exist, so as to retrain the neural network model through the model training module.

[0017] In some possible implementations, the decision module is also used to determine the index of the maximum value in the prediction result; determine the predicted work step content corresponding to the processed data table based on the index of the maximum value and the preset mapping relationship; and filter the target data in the processed data table that matches the predicted work step content and the text decision content.

[0018] In some possible implementations, the decision module is also used to verify the validity of the numerical processing feature; when the numerical processing feature is valid, the format of the numerical processing feature is converted based on the data format matching the trained model to obtain converted data; based on the number of neurons in the input layer and output layer of the trained model, the restored model is used to predict the processing decision plan for the converted data to obtain the prediction result.

[0019] In a second aspect, an embodiment of the present invention provides a hydraulic component processing decision-making method based on a neural network model, the method comprising:

[0020] Obtaining a process data table from a database and an input numerical processing feature for a hydraulic component, processing null values in a specific column of the process data table, and obtaining and storing a processed data table;

[0021] Performing back propagation training on the neural network model based on the processed data table to obtain a trained model;

[0022] If it is detected that the trained model exists, loading the trained model from the specified path;

[0023] Performing a processing decision prediction based on the trained model and the numerical processing feature to obtain a prediction result;

[0024] Based on the prediction result and the preset mapping relationship, determining the predicted work step content corresponding to the processed data table;

[0025] Based on the input text decision content for the hydraulic component and the predicted work step content, target data is filtered and displayed in the processed data table.

[0026] According to a third aspect, a computer program product is provided. The computer program product includes a computer program code, which enables the computer to execute the above method when the computer program code is executed on the computer.

[0027] In a fourth aspect, a computer-readable storage medium is provided, which stores a computer program code. When the computer program code is run on a computer, the computer executes the above method.

[0028] The present invention has the following beneficial effects: In a hydraulic component processing decision system based on a neural network model, a data loading module loads a process data table from a database, processes null values in specific columns in the process data table, and obtains and stores a processed data table. A model training module then performs backpropagation training on the neural network model based on the processed data table to obtain a trained model. This allows full utilization of hydraulic component processing data, enabling intelligent decision-making using the trained neural network model to improve the accuracy of processing decisions. The model loading module then detects whether the trained model exists. If so, it loads the trained model from a specified path, facilitating processing decision prediction in the decision module. Finally, the decision module obtains input numerical processing features for the hydraulic component, obtains prediction results based on the trained model and the numerical processing features, and determines the predicted process step content corresponding to the processed data table based on the prediction results and a preset mapping relationship. Target data is then filtered and displayed in the processed data table based on input textual decision content for the hydraulic component and the predicted process step content. In this way, using the trained model to predict machining decisions can improve the efficiency and accuracy of machining decisions and obtain machining plans that meet specific production needs. This not only significantly improves the machining accuracy of hydraulic components and ensures high manufacturing quality standards for key hydraulic parts, but also effectively reduces repetitive work and significantly reduces time costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0030] Figure 1 This is a schematic diagram of the architecture of a hydraulic component processing decision system based on a neural network model provided by an embodiment of the present invention;

[0031] Figure 2 This is a schematic diagram of the structure of a hydraulic component processing decision system based on a neural network model provided by an embodiment of the present invention;

[0032] Figure 3 Schematic diagram of the training process of the model training module provided in an embodiment of the present invention;

[0033] Figure 4 This is a diagram showing a visualization platform for a decision-making system provided by an embodiment of the present invention;

[0034] Figure 5This is a schematic diagram of some hydraulic components and corresponding characteristic attributes provided by an embodiment of the present invention;

[0035] Figure 6 This is a schematic diagram of the implementation flow of a hydraulic component processing decision-making method based on a neural network model provided by an embodiment of the present invention;

[0036] Figure 7 It is a structural diagram of a computer device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0037] To further illustrate the technical means and effectiveness of the present invention in achieving its intended objectives, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail the specific implementation, structure, features, and effectiveness of a hydraulic component processing decision-making system based on a neural network model. In the following description, references to "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0038] In the description of the embodiments of the present invention, unless otherwise specified, " / " means or, for example, A / B can mean A or B: "and / or" in the text is only a description of the association relationship of associated objects, indicating that there can be three relationships, for example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. In addition, in the description of the embodiments of the present invention, "multiple" refers to two or more than two.

[0039] In the following, the terms "first" and "second" are used for descriptive purposes only and should not be understood to imply or suggest relative importance or implicitly indicate the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features.

[0040] Unless otherwise defined, all technical and scientific terms used in the embodiments of the present invention have the same meanings as commonly understood by one of ordinary skill in the art to which the present invention belongs.

[0041] The embodiment of the present invention provides a hydraulic component processing decision system based on a neural network model, which can intelligently filter out process instance data tables that meet the requirements from a large number of data tables in an SQL database based on the processing characteristics (including numerical values and text) input by the user, and display the number of process instance tables that meet the requirements and some process instance tables, thereby improving the accuracy and efficiency of processing decisions. Figure 1As shown, the data loading module 11 connects to the process database by connecting to the database. It sets null values in the table to "None" and stores the data table in the data table library (allDateTables) list, providing a dataset for subsequent model training and decision making. The model training module 12 preprocesses and encodes the dataset, defines the network structure, and trains the neural network using the backpropagation learning algorithm. It iteratively calls the teacher.RunEpoch() method to complete the training process and uses the Serializer.Save method to save the trained model to a specified file. The model loading module 13 then checks whether the trained model exists. If so, it loads the model using the Serializer.Load method, deserializes the model file, and retrieves the number of neurons in the model's input and output layers. If the trained model does not exist, it returns to the previous module, retrains the model, and saves it. The decision module 14 obtains the numerical processing features entered by the user and converts them into input format. The neural network calculates the maximum index and maps the corresponding content. It then selects the data tables that meet the load neural network prediction and the text decision criteria, compiles the statistics, and displays some of the tables in the data grid view. Among them, text decision is implemented based on container (Contains) statement.

[0042] The following is a detailed description of a specific scheme of a hydraulic component processing decision system based on a neural network model provided by the present invention in conjunction with the accompanying drawings. Figure 2 , which shows a schematic diagram of an implementation flow of a hydraulic component processing decision system based on a neural network model provided by one embodiment of the present invention. The system 200 includes:

[0043] The data loading module 201 is used to load the process data table of the database, process the null values in a specific column of the process data table, and obtain and store a processed data table.

[0044] Here, the data loading module 201 sets the null values of the specific columns to None and stores each processed data table in the data list library. The specific columns in the process data table can be custom columns or columns related to process step content. The data loading module 201 connects the decision system to the database and loads the process data tables from the database. For example, the intelligent decision system connects to the database, loads all process data tables from the database, processes null values in the specific columns, and stores the processed data tables. The hydraulic component processing intelligent decision system is developed based on a Windows Forms application in Visual Studio. The database is a Structured Query Language (SQL) database used to store complete process data tables for multiple hydraulic components. The SQL Server database system has powerful transaction processing capabilities and can store thousands of hydraulic component processing process data tables. Various methods are used to ensure data integrity, providing dataset support for subsequent neural network model training. The data loading module also establishes a test connection between the database and the decision system using a structured query connection statement (i.e., a SqlConnection statement). Before training the model and making decisions, you need to test-connect the SQL Server database system to the intelligent decision-making system developed based on Visual Studio to obtain a processed data set. This connection process is mainly implemented through the SqlConnection statement to ensure stable communication and data exchange between the two. The connection logic is as follows:

[0045] string connectionString="database server name and user account password";

[0046] SqlConnection connection=new SqlConnection(connectionString)

[0047] The connectionString string must include not only the SQL database server name (Data Source) and the database user account (User ID) and password (Password), but also the parameters for trusting the server certificate. That is, Trust Server Certificate must be set to True.

[0048] Based on the above configuration logic, the system will automatically establish a connection with the SQL process database after startup. Subsequently, the system will call the connection.GetSchema("Tables") method to obtain relevant information of all tables in the database, and call the Fill method of the SqlDataAdapter component to execute the SQL query statement to fill it into the DataTable.

[0049] Specifically, for each table name, you can construct the following SQL query statement:

[0050] SELECT * FROM [TableName]

[0051] If a null value appears in the "Work Step Content" column of the populated DataTable, the system will set the null value to the character "None" and then store each DataTable in the allDataTables list to provide a dataset basis for subsequent model training and decision-making.

[0052] The model training module 202 is used to perform back propagation training on the neural network model based on the processed data table to obtain a trained model.

[0053] Here, the model training module 202 uses the data in the column of the work step content in the processed data table as the output label, and uses the data other than the data in the column of the work step content in the processed data table as the input feature, normalizes the input feature, and one-hot encodes the output label, thereby building a neural network model including an input layer, a hidden layer, and an output layer. In some possible implementations, the model training module 202 uses the data in the columns other than the work step content column as the input feature, and the data in the work step content column as the output label, normalizes the input feature, and one-hot encodes the output label, thereby creating a neural network including an input layer, a hidden layer, and an output layer, and uses a back-propagation learning algorithm for training. After the training is completed, the model is saved.

[0054] The model training module is used to train the neural network model to improve the accuracy and efficiency of subsequent decision-making, such as Figure 3As shown, the model training module first extracts data. This involves traversing all data tables loaded from the database, extracting data except for the "Work Step Content" column as input features, and converting these input features to double type. The "Work Step Content" column is then extracted as the output label, and encoded using one-hot encoding. Data other than the "Work Step Content" column is used as input features, and the "Work Step Content" column is used as the output label, as this column will be used to implement decision-making functions. To ensure more stable and efficient neural network training, the input feature data must be converted to double type and normalized to complete data preprocessing. Neural network structure 31 includes: the number of neurons in the input layer equals the number of input features; the number of neurons in the hidden layer equals 10, with the Sigmoid function as the activation function; and the number of neurons in the output layer equals the number of different work step contents.

[0055] Since neural networks can only process numerical data, the output label "work step content" (string type) also needs to be encoded. This method uses one-hot encoding to process the output label.

[0056] Specifically, the method of one-hot encoding the output label of "work step content" is as follows: construct a dictionary outputMapping to store different "work step contents" and their corresponding encoding indexes; traverse all "work step contents" and assign a unique index to each different "work step content"; according to the mapping relationship, convert each "work step content" into a one-hot encoded vector, the vector length is equal to the number of different "work step contents", the value of the corresponding index position is 1, and the other positions are 0.

[0057] After completing the extraction of the data set, the network structure needs to be defined. That is, a three-layer neural network is constructed using the C# programming language, including an input layer (inputSize), an output layer (outputSize), and a hidden layer. The hidden layer contains 10 neurons, and the activation function is the Sigmoid function.

[0058] Specifically, the number of neurons in the input layer is determined by the number of input features, and the number of neurons in the output layer is equal to the number of different "work-step contents". The Sigmoid function can map the input value to the (0,1) interval, introducing nonlinear characteristics to the network, enabling it to learn complex patterns.

[0059] After defining the network structure, the system can start training the model, that is, performing model training 32: the code uses the back propagation learning algorithm to train the neural network. This algorithm is based on the principle of gradient descent. By calculating the gradient of the error, it continuously adjusts the weights and biases of the neural network to make the network output results as close to the true label as possible.

[0060] The learning rate is set to 0.1 to control the step size of each weight update. If the learning rate is too large, the model may not converge; if it is too small, the training speed will be slow. The momentum is set to 0.5. Momentum can help the model skip local optimal solutions during gradient descent and accelerate convergence.

[0061] During the training process, the system calls the teacher.RunEpoch(inputArray, outputArray) method multiple times to complete the training process. Each iteration calculates the output of the network based on the current weights and biases, then calculates the error between the output and the true label, and then updates the weights and biases according to the gradient of the error.

[0062] Specifically, in each iteration, input feature data is passed from the input layer, calculated by the hidden layer, and processed through the weighted summation of neurons and the Sigmoid activation function, ultimately resulting in the output layer's output. The output of the output layer is then compared with the true output label (the one-hot encoded vector) to calculate the error between the two. A common error calculation method is the mean squared error (MSE). Based on the calculated error, the backpropagation algorithm calculates the gradient of the error with respect to the weights and biases of each neuron. The gradient represents the direction and rate at which the error changes with the weights and biases. Based on the calculated gradient, the weights and biases of each neuron in the neural network are updated according to the rules of learning rate and momentum.

[0063] Training continues until the error falls below a set threshold (e.g., 0.01), at which point the model is considered to have converged to a good state and training stops. If the maximum number of iterations (e.g., 1000 iterations) is reached, training stops even if the error still hasn't reached the threshold, to avoid excessive training time.

[0064] For trained models, the system uses the Serializer.Save method to save the model to a specified file (such as "model.bin"). This allows the system to directly load the model for prediction during subsequent decision-making processes without retraining. The code also performs a simple validation of the saved model and attempts to load it from the file. If the load succeeds, the model saving and loading process is normal; if the load fails, an error message is output to facilitate debugging and repair.

[0065] The model loading module 203 is configured to load the trained model from a specified path if it is detected that the trained model exists.

[0066] Here, the model loading module is further configured to, if the trained model is detected to exist, deserialize the trained model, restore the structure and parameters of the trained model, and obtain a restored model; and obtain and save the number of neurons in the input and output layers of the trained model. This allows the decision module to predict the processing decision solution based on the numerical processing features and the number of neurons in the input and output layers, thereby obtaining the predicted result. The model loading module is used to load a trained neural network model. The system first checks whether the model file exists. If so, it loads the model using the Serializer.Load method, deserializes the read model file, and restores the structure and parameters of the neural network model using the same algorithm and configuration used when the model was trained and saved, ensuring the integrity of the model. The module also obtains the number of neurons in the input and output layers of the model and stores them in the inputSize and outputSize variables, respectively, to provide parameter support for subsequent decision making. If the model is detected to be non-existent, the system returns to the model training module to retrain and save the neural network model.

[0067] Decision module 204 is configured to obtain input numerical processing features for hydraulic components; perform processing decision prediction based on the trained model and the numerical processing features to obtain a prediction result; and determine the predicted process step content corresponding to the processed data table based on the prediction result and a preset mapping relationship; and filter and display target data in the processed data table based on the input text decision content for the hydraulic component and the predicted process step content. The numerical processing features match the type of hydraulic component to be processed, including: processing process name, surface roughness, and tool parameters.

[0068] The decision module is further configured to determine the index of the maximum value in the prediction result; determine the predicted process step content corresponding to the processed data table based on the index of the maximum value and the preset mapping relationship; and filter the processed data table for target data that matches the predicted process step content and the text decision content. Furthermore, the module is configured to verify the validity of the numerical processing feature; if the numerical processing feature is valid, convert the format of the numerical processing feature based on the data format matched by the trained model to obtain converted data; and predict the processing decision solution for the converted data using the restored model based on the number of neurons in the input and output layers of the trained model to obtain the prediction result.

[0069] In some possible implementations, the decision module is used to implement the final decision function, such as Figure 4As shown, the module can realize visual operation in Windows form application developed based on Visual Studio, that is, users can directly operate on the visual platform (i.e. Figure 4 The decision operation is completed on the XXX hydraulic component decision platform 41 shown in the figure. The decision, the number of decision results, and the exit are also displayed on the visualization platform. It also includes: a decision result display module 42. The specific operation steps are as follows:

[0070] First, the system needs to be initialized before making a decision. That is, it connects to the SQL database and loads all data tables related to hydraulic components. If the model file does not exist, the data is preprocessed, the neural network is trained, and the model is saved. After that, the trained model is loaded.

[0071] After entering the decision platform, users can select the input hydraulic components including pistons, piston rods, flanges, cylinder heads, cylinder bodies, and rod barrels. Then, users can directly set processing features such as processing procedure name, surface roughness, tool parameters, etc. in multiple numerical or text-type TextBox controls in the decision system visualization interface. The specific processing feature selection depends on the type of hydraulic component to be processed. Specific processing features such as Figure 5 As shown, hydraulic component 51 includes a piston, piston rod, flange, cylinder head, cylinder body, and rod barrel. Key machining features of the piston rod include whether fine grinding is required, the chrome-plated outer diameter, and the number of chrome plating layers. Key machining features of the flange include the threaded aperture size. Key machining features of the cylinder head include the milling plate slot width. Key machining features of the cylinder body include the welding material selected after rough turning and whether the cylinder bottom needs fine turning. Key machining features of the rod barrel include whether fine grinding is required, the chrome-plated outer diameter, and the number of chrome plating layers.

[0072] After completing the setting of key processing features, the user can click the decision button. The decision system first checks whether the `allDataTables` list is empty. If so, it prompts the user to check the database. The system then retrieves the numerical processing features entered by the user in the numeric TextBox control and verifies the validity of the numerical features to ensure that the input is a valid number. This is because the neural network model can only process valid numeric input. If the input is not a valid number, such as characters or other non-numeric content, the system prompts the user to re-enter the input to ensure the accuracy of subsequent model calculations. The system also retrieves the processing features entered by the user in the text TextBox control.

[0073] After verification, the system converts the user-entered numerical features for the hydraulic components into the neural network input format. Since the input data was preprocessed during the model training phase, the user-entered numerical features also need to undergo the same preprocessing to ensure they meet the model's input requirements. The loaded neural network model is then used to perform calculations and obtain predictions. Based on the predictions and pre-established mappings, the corresponding "work step content" is determined.

[0074] Specifically, the system converts the input features into `double` type and stores them in an array named `inputFeatures`. The loaded neural network model is then used to calculate the `inputFeatures` array to obtain the output result `output`, which is a vector containing multiple numerical values, each of which corresponds to a neuron in the output layer of the neural network. The system automatically finds the index `predictedIndex` of the maximum value in the `output` array. This index represents the most likely category or result predicted by the neural network. The corresponding "work step content" `predictedStepContent` is obtained through the mapping relationship established when the "work step content" is one-hot encoded during the model training phase. For example, when the index of the maximum value found is 3, and the "work step content" corresponding to index 3 is "fine turning circle", then the result predicted by the model is "fine turning circle".

[0075] Based on the predicted "work step content" and the text decision content for hydraulic components entered by the user, all data tables related to hydraulic components are screened to find data tables that meet the requirements. The text-based processing feature decision is implemented based on the Contain statement.

[0076] Specifically, the Contains statement is an efficient tool for detecting whether a string contains a specific substring. The main string (Main String) represents the complete text to be searched, and the substring (Substring) represents the specific keyword or phrase to be found. In the context of hydraulic component processing, when this method is integrated into a more complex system, that is, integrated with a neural network model, it can be used to check whether specific keywords exist in data related to processing steps, process names, or other related text features. When the decision system detects that the text-type TextBox control is not empty, the system will use the Contains statement to implement the decision of the text-type processing feature. The Contains statement is implemented through the following logic:

[0077] return IndexOf(value,StringComparison.Ordinal)>=0;

[0078] The IndexOf method is used to search for a specified value in a string. If the value is found, it returns the zero-based index of the first occurrence of the value, that is, an index greater than or equal to zero. In this case, the method returns true. If the value is not found, it returns false. More specifically, the following code example illustrates how the Contains statement works:

[0079]

[0080] For the above code, when traversing all data tables, the system checks the variable stepContent based on the text decisionText entered by the user in the TextBox control, that is, it detects whether the value of the "step content" column in the data table contains the string content in decisionText, filters out the data tables whose "step content" contains the text content, and adds these data tables that meet the conditions and have not yet been added to the list to the matchingTables list, so that these qualified data tables can be subsequently counted and displayed.

[0081] The text-based processing feature decision function based on the Contains statement is combined with the neural network model to strengthen the role of the neural network model and achieve more accurate and efficient decision-making functions. When the text-based decision content is not empty, the "work step content" contains the text decision content, and the neural network model will combine with the Contains algorithm to complete the final decision; when the text decision content is empty, it only needs to meet the neural network prediction results.

[0082] Specifically, to implement the decision-making function, the system first traverses the data tables in the allDataTables list. For each data table, the "Work Step Content" column is checked. If the table's "Work Step Content" is equal to `predictedStepContent` and satisfies the text decision conditions, such as `decisionText` is not empty and the "Work Step Content" contains `decisionText`, or `decisionText` is empty, the system will add the table to the `matchingTables` list. For example, if the model predicts "Work Step Content" is "Fine Turning Round" and the user enters the text decision content "Includes External Thread Turning", then only data tables with "Work Step Content" of "Fine Turning Round" and "Includes External Thread Turning" will be filtered out.

[0083] Finally, the decision-making system will count the number of data tables that meet the requirements and display the result number `matchingCount` on the visualization platform. If `matchingCount` is 0, the user will be prompted that no data tables that meet the requirements are found; if `matchingCount` is not 0, some data tables that meet the requirements will be displayed in the data grid view DataGridView control, and the number of data tables that meet the requirements will be displayed. In this way, users can intuitively see the processing technology examples that meet their input conditions, providing a reference basis for their processing decisions. At the same time, users can quickly select according to the decision results to achieve efficient processing and production.

[0084] In an embodiment of the present invention, a hydraulic component processing decision system based on a neural network model loads a process data table from a database through a data loading module, processes null values in specific columns of the process data table, and obtains and stores a processed data table. A model training module then performs backpropagation training on the neural network model based on the processed data table to obtain a trained model. This allows full utilization of hydraulic component processing data, enabling intelligent decision-making using the trained neural network model to improve the accuracy of processing decisions. The model loading module then detects whether the trained model exists. If so, it loads the trained model from a specified path, facilitating processing decision prediction in the decision module. Finally, the decision module obtains input numerical processing features for the hydraulic component, obtains prediction results based on the trained model and the numerical processing features, and determines the predicted process step content corresponding to the processed data table based on the prediction results and a preset mapping relationship. Target data is then filtered and displayed in the processed data table based on the input textual decision content for the hydraulic component and the predicted process step content. In this manner, processing decision prediction using the trained model improves the efficiency and accuracy of processing decisions and results in a processing plan that meets specific production requirements. This not only significantly improves the machining accuracy of hydraulic components and ensures that the manufacturing quality of key hydraulic parts meets high standards, but also effectively reduces repetitive work and significantly reduces time costs.

[0085] The embodiment of the present invention provides a hydraulic component processing decision method based on a neural network model, such as Figure 6 As shown, this can be achieved by following the steps below:

[0086] 601 , obtaining a process data table from a database and input numerical processing features for a hydraulic component, processing null values in a specific column in the process data table, and obtaining and storing a processed data table.

[0087] Here, the process data table is loaded from the database through the data loading module, and the null values in the specific columns in the process data table are processed to obtain and store the processed data table.

[0088] The user interaction platform corresponding to the decision module receives the numerical processing characteristics input by the user for the hydraulic component, and performs processing decision prediction based on the numerical processing characteristics of the hydraulic component.

[0089] 602. Perform back propagation training on the neural network model based on the processed data table to obtain a trained model.

[0090] Here, in the model training module, the neural network model is back-propagated and trained through the processed data table to obtain and store the trained model for subsequent calls, thereby making intelligent processing decision predictions through the trained model.

[0091] 603. If it is detected that the trained model exists, load the trained model from the specified path.

[0092] Here, the model loading module detects whether the trained model exists. If the trained model exists, the trained model is loaded according to the specified path; wherein the specified path is the storage path of the trained model.

[0093] 604 , performing processing decision prediction based on the trained model and the numerical processing feature to obtain a prediction result.

[0094] 605 : Based on the prediction result and the preset mapping relationship, determine the predicted process step content corresponding to the processed data table.

[0095] 606 , based on the input text decision content for the hydraulic component and the predicted work step content, filter and display target data in the processed data table.

[0096] Here, the above steps 604 to 606 can be implemented through a decision module, in which the user inputs the numerical processing characteristics and text decision content for the hydraulic component, and the target data is filtered in the processed data table through the text decision content and the predicted work step content, and is statistically analyzed and displayed.

[0097] The embodiments of the present invention can more accurately predict the appropriate processing "step content" by learning and analyzing a large amount of hydraulic component processing data through a neural network model, thereby improving the accuracy of processing decisions. The intelligent decision-making system based on neural networks can quickly process input data and give decision results, greatly improving decision-making efficiency; it can adapt to the complex and changing situations in the hydraulic component processing process, and optimize the model by continuously collecting feedback data to improve the generalization ability and adaptability of the model. By making full use of historical processing data, data-driven processing decisions are realized, providing strong support for the intelligent processing of hydraulic components. In this way, the embodiments of the present invention not only solve many problems in traditional hydraulic component processing methods, but also provide effective solutions for improving processing accuracy, improving production efficiency, reducing production costs and promoting intelligent manufacturing.

[0098] Optionally, the transmission medium can be a wired link (for example, but not limited to, coaxial cable, optical fiber and digital subscriber line (DSL)) or a wireless link (for example, but not limited to, wireless Fidelity (WIFI), Bluetooth and mobile device network). It should be noted that the system provided in the above embodiment is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device can be divided into different functional modules to complete all or part of the functions described above. In addition, the method embodiments provided in the above embodiments belong to the same concept. The specific implementation process is detailed in the method embodiments, which will not be repeated here.

[0099] Figure 7 FIG. 1 is a schematic diagram of the structure of a computer device provided by an embodiment of the present invention. For example, Figure 7 As shown, the computer device 700 includes: a memory 701, a processor 702, and a computer program 703 stored in the memory 701 and running on the processor 702, wherein when the processor 702 executes the computer program 703, the computer device can execute any one of the hydraulic component processing decision methods based on the neural network model introduced above.

[0100] In addition, an embodiment of the present invention also protects a system, which may include a memory and a processor, wherein the memory stores an executable program code, and the processor is used to call and execute the executable program code to execute a hydraulic component processing decision system based on a neural network model provided by an embodiment of the present invention. This embodiment can divide the system into functional modules according to the above-mentioned method example. For example, it can correspond to each functional module, or two or more functions can be integrated into one processing module. The above-mentioned integrated module can be implemented in the form of hardware. It should be noted that the division of modules in this embodiment is schematic, which is only a logical function division. There may be other division methods in actual implementation. It should be noted that all relevant contents of each step involved in the above-mentioned method embodiment can be referred to the functional description of the corresponding functional module, and will not be repeated here.

[0101] It should be understood that the system provided in this embodiment is used to execute the above-mentioned hydraulic component processing decision-making method based on a neural network model, and therefore can achieve the same effect as the above-mentioned implementation method. In the case of an integrated unit, the system may include a processing module and a storage module. Among them, when the system is applied to a device, the processing module can be used to control and manage the actions of the device. The storage module can be used to support the device to execute mutual program codes, etc. Among them, the processing module can be a processor or a controller, which can implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of the present invention. The processor can also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of digital signal processing (DSP) and a microprocessor, etc., and the storage module can be a memory.

[0102] In addition, the system provided by an embodiment of the present invention may be specifically a chip, component, or module. The chip may include a connected processor and memory; the memory is used to store instructions. When the processor calls and executes the instructions, the chip can execute the hydraulic component processing decision-making method based on a neural network model provided in the above embodiment. This embodiment also provides a computer-readable storage medium storing computer program code. When the computer program code is executed on a computer, it causes the computer to execute the above-mentioned method steps to implement the hydraulic component processing decision-making method based on a neural network model provided in the above embodiment.

[0103] This embodiment also provides a computer program product. When the computer program product is executed on a computer, it causes the computer to execute the above-mentioned steps to implement the hydraulic component processing decision-making method based on a neural network model provided in the above embodiment. The system, computer-readable storage medium, computer program product, or chip provided in this embodiment are all used to execute the corresponding method provided above. Therefore, the beneficial effects achieved by the system can refer to the beneficial effects of the corresponding method provided above and will not be repeated here. Through the description of the above embodiments, those skilled in the art will understand that for the sake of convenience and brevity, only the division of the above-mentioned functional modules is used as an example. In actual applications, the above-mentioned functions can be distributed among different functional modules as needed, that is, the internal structure of the system can be divided into different functional modules to complete all or part of the functions described above. In the embodiments provided by the present invention, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the division of modules or units is only a logical functional division. In actual implementation, other division methods can be used. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not implemented. On the other hand, the mutual coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interface, system or unit, which may be electrical, mechanical or other forms.

[0104] It should be noted that the above-mentioned order of the embodiments of the present invention is for description only and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or may be advantageous. The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other. Each embodiment focuses on the differences from other embodiments. The above content is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or replacements within the technical scope disclosed by the present invention, which should be covered within the scope of protection of the present invention.

Claims

1. A hydraulic component processing decision system based on a neural network model, characterized in that: The system comprises: A data loading module is used to load a process data table from a database, process null values in a specific column of the process data table, and obtain and store a processed data table; A model training module, configured to perform back propagation training on a neural network model based on the processed data table to obtain a trained model; A model loading module, configured to load the trained model from a specified path if the trained model is detected to exist; A decision module is used to obtain the input numerical processing features for the hydraulic component; perform processing decision prediction based on the trained model and the numerical processing features to obtain a prediction result; and determine the predicted work step content corresponding to the processed data table based on the prediction result and a preset mapping relationship; based on the input text decision content for the hydraulic component and the predicted work step content, filter and display target data in the processed data table.

2. A hydraulic component processing decision system based on a neural network model according to claim 1, characterized in that: The data loading module is further used to connect the decision system with a database and load the process data table from the database.

3. The hydraulic component processing decision system based on a neural network model according to claim 2, characterized in that: The database is used to store process data tables of multiple hydraulic components with integrity; The data loading module is further used to perform a test connection between the database and the decision system based on a structured query connection statement.

4. The hydraulic component processing decision system based on a neural network model according to claim 1, characterized in that: The data loading module is further configured to set the null value of the specific column to be zero, and store each processed data table in a data list library.

5. The hydraulic component processing decision system based on a neural network model according to claim 1, characterized in that: The model training module is also used to use the column data belonging to the work step content in the processed data table as the output label, and the data other than the column data belonging to the work step content in the processed data table as the input feature, and normalize the input features and perform one-hot encoding on the output label to build a neural network model including an input layer, a hidden layer and an output layer.

6. The hydraulic component processing decision system based on a neural network model according to claim 1, characterized in that: The model loading module is further configured to, if the existence of the trained model is detected, deserialize the trained model, restore the structure and parameters of the trained model, and obtain a restored model; and obtain and save the number of neurons in the input layer and output layer of the trained model; The decision module is further used to make a processing decision prediction based on the numerical processing features and the number of neurons in the input layer and the output layer to obtain the prediction result.

7. The hydraulic component processing decision system based on a neural network model according to claim 1 is characterized in that: The model loading module is also used to return to the model training module if it is detected that the trained model does not exist, so as to retrain the neural network model through the model training module.

8. The hydraulic component processing decision system based on a neural network model according to claim 1 is characterized in that: The decision module is further used to determine the index of the maximum value in the prediction result; determine the predicted work step content corresponding to the processed data table based on the index of the maximum value and the preset mapping relationship; and filter the target data that matches the predicted work step content and the text decision content in the processed data table.

9. The hydraulic component processing decision system based on a neural network model according to claim 8, characterized in that: The decision module is also used to verify the validity of the numerical processing feature; when the numerical processing feature is valid, the format of the numerical processing feature is converted based on the data format matching the trained model to obtain converted data; based on the number of neurons in the input layer and output layer of the trained model, the processing decision plan for the converted data is predicted using the restored model to obtain the prediction result.

10. A hydraulic component processing decision-making method based on a neural network model, characterized in that: The method comprises: Obtaining a process data table from a database and an input numerical processing feature for a hydraulic component, processing null values in a specific column of the process data table, and obtaining and storing a processed data table; Performing back propagation training on the neural network model based on the processed data table to obtain a trained model; If it is detected that the trained model exists, loading the trained model from the specified path; Performing a processing decision prediction based on the trained model and the numerical processing feature to obtain a prediction result; Based on the prediction result and the preset mapping relationship, determining the predicted work step content corresponding to the processed data table; Based on the input text decision content for the hydraulic component and the predicted work step content, target data is filtered and displayed in the processed data table.

Images