Electrical control cabinet BOM generation method, computer equipment and medium

Abstract

The invention discloses an electrical control cabinet BOM generation method, and the method comprises the steps: classifying BOM parts of an electrical control cabinet product into three modules: a universal assembly, a standard assembly and an additional assembly, enabling the assembly of each module to be associated with a corresponding keyword, and automatically generating the BOM of the electrical control cabinet product in a mode of constructing a neural network model. In this way, the workload of designers can be greatly reduced, product type selection is more oriented to be unified, and the design cost and the production cost can be greatly reduced; and compared with the traditional BOM, the development of the modularized BOM not only aims to simplify data contents, but also ensures that all related departments can obtain accurate and reliable BOM information by combining product design and optimization of manufacturing technical parameters.

Description

Technical Field

[0001] The present invention relates to the field of electric control cabinet generation, and in particular to a BOM table generation method in the electric control cabinet production process, a system for implementing the method, and a computer device and a medium for running the method. Background Art

[0002] In the current era of high customization and technology intensiveness, the management of bill of materials (BOM) has become increasingly critical, especially in the field of product design and production. With the advancement of technology, BOM has evolved from a simple list of materials to an important tool to support diversified, modular design and efficient production. Especially in modern manufacturing, in order to maximize the reuse of diversified products and quickly adjust the configuration to adapt to market changes, high requirements are placed on BOM (Bill of Materials) management. The accuracy and real-time nature of BOM data is the framework to support product life cycle management and efficient order processing, and these data are crucial to the success of a company's operations.

[0003] Although modular intelligent BOM has been fully applied in many fields, there are still many challenges when applying modular BOM to the field of electrical control cabinets. The challenges faced by electrical control cabinet BOM are mainly manifested in: (1) The demand is increasing, and different industries and application scenarios (such as medical and food storage) have demand for electrical control cabinets; (2) The product types and configurations are complex, including a variety of combinations between different unit types (for example: screw, scroll, piston units) and terminal types; (3) Market and regulatory changes require rapid adjustments to the design of control cabinets to adapt to new market regulations. As the number of product types increases, the number of components and BOM complexity increase dramatically. At present, during production, the BOM of electrical control cabinets is still mainly operated by manual screening. Due to the wide variety of control cabinets and the large number of components required for each type of electrical control cabinet, the manual construction of the BOM of electrical control cabinets is a heavy workload for designers, and designers may have differences in selection for many basic components, which will make the product BOM more complicated and increase the design and production costs of the enterprise. Therefore, it is necessary to further improve the existing method of generating the BOM of electrical control cabinets. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a method for automatically building and generating a BOM of an electrical control cabinet in view of the deficiencies in the prior art.

[0005] A method for generating a BOM for an electrical control cabinet comprises the following steps: S1, sample collection: classifying the BOM parts in a sample list database table of an existing electrical control cabinet product into general components, standard components and additional components, and setting a plurality of keywords to associate the general components, standard components and additional components with the corresponding keywords; S2, sample classification: taking the BOM keywords, general components, standard components and additional components of the same electrical control cabinet product as a data set, and dividing all the collected data sets into a training set; S3, establishing a BOM generation model: constructing a neural network model, training the BOM generation model through the training set, comparing the general components, standard components and additional components generated by the BOM generation model with the actual general components, standard components and additional components, and calculating the model loss. When the model loss is less than a specified threshold, the BOM generation model training is completed; S4, according to the design requirements of a new electrical control cabinet product, inputting the corresponding keywords into the BOM generation model, and generating the corresponding product BOM through the BOM generation model.

[0006] Preferably, when the BOM of the electrical control cabinet product is collected in step S1, the natural language description in the BOM data is first encoded, and the text description of each component is converted into a high-dimensional context-related embedding vector; then the numerical features in the BOM are extracted and normalized; finally, the embedded vector and the digital feature are combined into a unified feature vector.

[0007] Preferably, the neural network model in step S3 is a multi-layer perceptron (MLP) model, whose input layer has a node number equal to the dimension of the joint feature vector; the hidden layer is designed with 3 hidden layers, each containing 64, 128 or 256 neurons; activation function: ReLU (Rectified Linear Unit) is used as the activation function of the hidden layer; output layer: the number of output nodes is equal to the number of classification labels, which are classified into general components, functional components, and additional components, then the output layer has 3 nodes; softmax is used as the activation function of the output layer to achieve multi-classification probability output.

[0008] Preferably, when training the model in step 3, the model is initialized using He initialization; the loss function uses the cross entropy loss function; the optimization algorithm uses the Adam optimization algorithm, and the learning rate is set to 0.001; the batch size uses 32 as the mini-batch size; in the training cycle, in each epoch, the training data is used to perform forward propagation and back propagation to update the weights; and the Early Stopping technique is used to prevent overfitting.

[0009] Preferably, the common components in step S1 can be obtained by taking the intersection of the BOMs of all electrical control cabinet products.

[0010] Preferably, in step S1, the sample list database table may be optimized, specifically comprising the following steps: 1) Database table design optimization includes the following steps: Step 1: Select appropriate data types and field lengths to reduce data storage overhead; Step 2: Add BTree indexes and Hash indexes to fields that are frequently used for queries; Step 3: Partition and shard tables that store large amounts of data; 2) Query optimization includes the following steps: Step 1: When querying data in a table, avoid querying all fields in the table and only query the required fields; Step 2: When querying data associated with multiple tables, use left outer joins to associate multiple tables; Step 3: Split complex query statements into multiple simple query statements; 3) Database caching technology improves data processing and includes the following steps: Step 1: The client sends the Redis executed commands and stored data to the Redis server via remote communication; Step 2: The Redis server will parse the sent content and perform the corresponding operation. Redis stores the data in the corresponding data structure in the server memory; Step 3: Redis will parse the sent query command, and then find the corresponding value according to the operation type and the key of the data to be found; Step 5: Return the value to the client via remote communication.

[0011] This patent also discloses a computer device, including a memory and a processor, wherein the memory stores a computer program, and is characterized in that the processor implements the steps of the above method when executing the computer program.

[0012] This patent also discloses a computer storage medium for storing a computer program, which implements the steps of the method described in the claim when the computer program is executed by a processor.

[0013] The above technical solution has the following beneficial effects: the present invention classifies the BOM components of electrical control cabinet products into three modules: general components, standard components and additional components, associates the components of each module with the corresponding keywords, and automatically generates the BOM of electrical control cabinet products by constructing a neural network model, which can greatly reduce the workload of designers, make product selection more oriented to unity, and greatly reduce design costs and production costs; and compared with traditional BOMs, the development of modular BOMs is not only to simplify data content, but also to ensure that all relevant departments can obtain accurate and reliable BOM information by combining product design and optimization of manufacturing technical parameters. This optimization helps to better cope with changes in market demand and challenges in the manufacturing process.

[0014] To further understand the features and technical contents of the present invention, please refer to the following detailed description and drawings of the present invention. However, the drawings provided are only for reference and description and are not intended to limit the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 The figure is a flow chart of a method according to an embodiment of the present invention.

[0016] Figure 2 A schematic diagram of the classification of BOM components according to an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The following is an explanation of the disclosed embodiments of the present invention through specific embodiments. Those skilled in the art can understand the advantages and effects of the present invention from the contents disclosed in this specification. The present invention can be implemented or applied through other different specific embodiments, and the details in this specification can also be modified and changed in various ways based on different viewpoints and applications without departing from the concept of the present invention. In addition, the drawings of the present invention are only simple schematic illustrations and are not depicted according to actual dimensions. It is stated in advance. The following embodiments will further explain the relevant technical contents of the present invention in detail, but the disclosed contents are not intended to limit the scope of protection of the present invention.

[0018] like Figure 1 As shown, the present invention discloses a method for generating a BOM of an electrical control cabinet, which comprises the following steps: S1 Sample collection: Based on the BOM in the existing electrical control cabinet product sample list database table, the existing BOM parts are classified into general components, standard components and additional components. General components mainly refer to the basic components commonly used in all electrical control cabinets, which are not limited by specific application scenarios and are highly versatile. For example, standard power modules, basic wiring terminals, etc., these components can be reused in various types of control cabinets to reduce repeated design. Functional components are due to the different power sizes of refrigeration equipment, and the power of the condenser changes accordingly. The power of the fan and water pump in the condenser is different, and the corresponding contactors, circuit breakers, and thermal relays are also different. The selection of functional components directly affects the functional realization and performance of the control cabinet. Additional components refer to components that need to be added only under specific needs or special application scenarios. They do not affect the basic functions of the control cabinet, but can enhance its additional functions. For example, additional display screens, special communication modules, and backup power supplies. The configuration of additional components can be flexibly selected according to the personalized needs of customers. At the same time, corresponding multiple keywords are set to associate general components, standard components, and additional components with corresponding keywords. Keywords can be common descriptions of products or natural languages ​​in BOM data.

[0019] S2 sample classification: Using the keywords, common components, standard components, and additional components of the same electrical control cabinet product BOM as a data set, all BOM collected data sets in the existing electrical control cabinet product sample list database table are divided into training sets for training the neural network model.

[0020] S3 establishes a BOM generation model: constructs a neural network model, trains the BOM generation model through a training set, uses keywords as the input of the BOM generation model, and uses the corresponding common components, standard components, and additional components as the output of the BOM generation model. The model loss is calculated by comparing the common components, standard components, and additional components generated by the BOM generation model with the actual common components, standard components, and additional components. When the model loss is less than the specified threshold, the BOM generation model training is completed.

[0021] S4. After the model training is completed, the corresponding keywords are input into the BOM generation model according to the new electrical control cabinet product design or order requirements, and the corresponding product BOM is generated through the BOM generation model. The product BOM includes three modules: general components, standard components and additional components.

[0022] After the BOM was modularized, the average response time of the system was significantly reduced. The average response time was reduced from 3 seconds to less than 1 second, which improved the user's operating experience and work efficiency. Through modular management, the accuracy of BOM data has been improved by about 30%. This is due to the standardized definition of common components and functional components, which reduces the data inconsistency caused by human errors and repeated entries. The modular BOM solution significantly improves the processing capacity of the system. The number of orders processed during peak periods increased by 50%, from 40 orders per day to 60 orders per day. At the same time, the system can maintain and update BOM data more efficiently, reducing the overall maintenance workload by 20%. According to the test data in actual applications, the various performance indicators after the implementation of the modular BOM solution are better than the traditional BOM management method. These quantitative data provide strong support for the effectiveness of the solution and prove its significant effect in improving production and management efficiency.

[0023] As a preferred solution of this patent, when collecting the BOM of the electrical control cabinet product in step S1, the natural language description in the BOM data is first encoded, and the text description of each component is converted into a high-dimensional context-related embedding vector; then the numerical features in the BOM are extracted and normalized; finally, the embedding vector and the digital features are combined into a unified feature vector. The specific implementation steps are: First, the natural language descriptions in the BOM data are encoded using a pre-trained BERT model. BERT can convert text descriptions of components (such as product names, specifications, etc.) into high-dimensional context-dependent embedding vectors. These embedding vectors can capture the semantic information in component descriptions. The specific steps include: 1. Extract all natural language description fields in the BOM (such as product name, description, purpose, etc.). 2. Use the pre-trained BERT model to encode each description and generate an embedding vector of fixed dimension. 3. The text description of each component will be converted into an embedding vector to represent the text information.

[0024] Next, we process the numerical features, because in addition to the text description, the parts in the BOM usually contain some numerical features and structured data (such as power, size, model, etc.). These features are indispensable information for classification and cannot be directly processed by BERT, so they need to be preprocessed separately. The specific steps are: 1. Extract numerical and structured features (such as power, size, model, etc.); 2. Normalize the numerical features to ensure that their distribution is suitable for the subsequent machine learning model.

[0025] Finally, feature fusion is performed to combine the embedding vector generated by BERT with the numerical features to form a unified feature vector. In this way, text information and numerical information can be used by subsequent classification models at the same time.

[0026] As another preferred embodiment of the present invention, the neural network model in step S3 is a multi-layer perceptron (MLP) model, whose input layer has a node number equal to the dimension of the joint feature vector; the hidden layer is designed with three hidden layers, each containing 64, 128 or 256 neurons; activation function: ReLU (Rectified Linear Unit) is used as the activation function of the hidden layer; output layer: the number of output nodes is equal to the number of classification labels, which are classified into general components, functional components, and additional components, then the output layer has three nodes; softmax is used as the activation function of the output layer to achieve multi-classification probability output.

[0027] When training the model, the model is initialized using He initialization; the loss function uses the cross entropy loss function; the optimization algorithm uses the Adam optimization algorithm, and the learning rate is set to 0.001; the batch size uses 32 as the mini-batch size; in the training cycle, in each epoch, the training data is used to perform forward propagation and back propagation to update the weights; and the Early Stopping technique is used to prevent overfitting.

[0028] The specific combination rules of the neural network model are: 1) Initialize the random forest model and use RandomForestClassifier to train the model.

[0029] 2) Model parameter settings: n_estimators=100: Set the number of trees in the random forest to 100 to provide sufficient diversity. max_depth=None: Do not limit the depth of the tree, allowing the tree to grow fully to capture all possible rules. min_samples_split=2: Set the minimum number of samples required for a node to continue splitting to 2. max_features='auto': Use the automatically selected optimal number of features to split the node, which is sqrt(number of features). random_state=42: Set the random seed to ensure the reproducibility of the results.

[0030] 3) Model training process: Use the fit(X_train, y_train) method to train the random forest model on the training dataset.

[0031] During the training process, random forest generates different combination rules through multiple decision trees, and gradually learns the relationship between each technical parameter and component combination in the data.

[0032] 4) Rule output: The trained random forest model generates multiple decision paths through a collection of multiple trees, and the leaf node of each tree represents a decision combination rule.

[0033] Use the export_text() method to export the visual tree structure and obtain specific rules.

[0034] For example, a tree splitting rule might be: "If power > 100 kW, select a contactor of size X and a circuit breaker of size Y.

[0035] 5) Feature Importance Analysis: Use model.feature_importances_ to extract the importance score of each feature in the model and identify which technical parameters have the greatest impact on the decision of component combination.

[0036] For example, through calculation, it is found that "power" has the greatest impact on contactor selection, with a weight of 0.65, followed by "expected price", with a weight of 0.25.

[0037] After the above model is trained, the trained random forest model is saved as a model object that can be called (for example, the model is serialized through joblib). In this way, when a new order is received in real time, the technical parameters of the new order are input into the model as keywords, and the model.predict(new_X) method is used to generate the predicted component combination. For example: when a new order is received (power is 120 kW and price is 12500), the model will output a component combination suggestion: "Select a contactor with specification X and a circuit breaker with specification Y".

[0038] As a preferred implementation, the common components in step S1 can be obtained by taking the intersection of the BOMs of all electrical control cabinet products. Because common components mainly refer to basic components commonly used in all electrical control cabinets, they are not restricted by specific application scenarios and are highly versatile. Therefore, they are used in almost all types of products. The part type of the common components can be quickly determined by taking the intersection method.

[0039] Due to the large number of bills of materials, it is easy to cause system jams. The present invention can also optimize the database table of the bill of materials: 1. Database table design optimization includes the following steps: Step 1: Select appropriate data types and field lengths to reduce data storage overhead Step 2: Add BTree indexes and Hash indexes to fields that are frequently used for queries Step 3: Partition and shard tables that store large amounts of data. 2. Query optimization includes the following steps: Step 1: When querying data in a table, avoid querying all fields in the table and only query the required fields. Step 2: When querying data associated with multiple tables, use a left outer join to associate multiple tables. Step 3: Split complex query statements into multiple simple query statements. 3. Database caching technology to improve data processing includes the following steps Step 1: The client sends the Redis executed commands and stored data to the Redis server via remote communication. Step 2: The Redis server will parse the sent content and perform the corresponding operation. Redis stores the data in the corresponding data structure in the server memory. Step 3: Redis will parse the sent query command, and then find the corresponding value based on the operation type and the key of the data to be found. Step 5: Return the value to the client via remote communication.

[0040] Considering the security of BOM data, this paper proposes how to ensure data security and prevent data leakage and tampering during data transmission and storage. 1) Use HTTPS to transmit data, and encrypt data through SSL / TLS encryption protocol to prevent data from being stolen or tampered during transmission. 2) Use AES algorithm to encrypt data. Use AES algorithm to encrypt sensitive data stored in the database (such as BOM list). 3) Fine-grained access control mechanism to ensure that only authorized users can access and operate BOM data. 4) Operation audit and log recording. Audit and log the addition, deletion, modification and query operations of BOM data to ensure the traceability of operations. 5) Regular data backup. Regularly back up BOM data to ensure that data can be quickly restored in the event of accidental loss or damage.

[0041] In order to achieve the above-mentioned purpose, an embodiment of the present invention further provides a computer device, which can be specifically a personal computer, a server, a network device, etc. The physical device includes a storage medium and a processor; the storage medium is used to store a computer program; the processor is used to execute the computer program to implement the steps and methods described in S1 to S4 above.

[0042] Accordingly, this embodiment also provides a computer storage medium on which a computer program is stored, and when the program is executed by the processor, the steps described in S1 to S4 are implemented. Based on this understanding, the technical solution of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each implementation scenario of the present application.

[0043] The present invention classifies the BOM components of electrical control cabinet products into three modules: general components, standard components and additional components, so that the components of each module are associated with corresponding keywords, and the BOM of electrical control cabinet products is automatically generated by constructing a neural network model, which can greatly reduce the workload of designers, make product selection more oriented to unity, and greatly reduce design costs and production costs; and compared with traditional BOMs, the development of modular BOMs is not only to simplify data content, but also to ensure that relevant departments can obtain accurate and reliable BOM information by combining product design and optimization of manufacturing technical parameters. This optimization helps to better cope with changes in market demand and challenges in the manufacturing process.

[0044] The contents disclosed above are only preferred feasible embodiments of the present invention, and are not intended to limit the scope of the present invention. Therefore, all equivalent technical changes made using the contents of the present invention's specification and drawings are included in the scope of the present invention.

Claims

1. A method for generating a BOM of an electrical control cabinet, characterized in that: It includes the following steps: S1. Sample collection: Classify the BOM components in the existing electrical control cabinet product sample list database table into common components, standard components and additional components, and set multiple keywords to associate the common components, standard components and additional components with the corresponding keywords; S2, sample classification: taking the BOM keywords, common components, standard components, and additional components of the same electrical control cabinet product as a data set, all collected data sets are divided into training sets; S3. Establish a BOM generation model: Build a neural network model, train the BOM generation model through the training set, compare the differences between the common components, standard components, and additional components generated by the BOM generation model and the actual common components, standard components, and additional components to calculate the model loss. When the model loss is less than the specified threshold, the BOM generation model training is completed; S4. According to the new electrical control cabinet product order or design requirements, the corresponding keywords are input into the BOM generation model, and the corresponding product BOM is generated through the BOM generation model. The product BOM includes three modules: general components, standard components and additional components.

2. The method for generating the BOM of an electrical control cabinet according to claim 1, characterized in that: When the BOM of the electrical control cabinet product is collected in step S1, the natural language description in the BOM data is first encoded, and the text description of each component is converted into a high-dimensional context-related embedding vector; then the numerical features in the BOM are extracted and normalized; finally, the embedding vector and the digital feature are combined into a unified feature vector.

3. The method for generating the BOM of an electrical control cabinet according to claim 1, characterized in that: The neural network model in step S3 is a multi-layer perceptron (MLP) model, whose input layer has the number of nodes equal to the dimension of the joint feature vector; the hidden layer is designed with three hidden layers, each containing 64, 128 or 256 neurons; activation function: ReLU (Rectified LinearUnit) is used as the activation function of the hidden layer; output layer: the number of output nodes is equal to the number of classification labels, which are classified into common components, functional components, and additional components, then the output layer has three nodes; softmax is used as the activation function of the output layer to achieve multi-classification probability output.

4. The method for generating the BOM of the electrical control cabinet according to claim 3 is characterized in that: In step 3, the model is initialized using He initialization when training the model; the loss function uses the cross entropy loss function; the optimization algorithm uses the Adam optimization algorithm, and the learning rate is set to 0.001; the batch size uses 32 as the mini-batch size; In each epoch of the training cycle, the training data is used to perform forward propagation and back propagation to update the weights. And use EarlyStopping technology to prevent overfitting.

5. The method for generating the BOM of an electrical control cabinet according to claim 1, characterized in that: In step S1, the common components can be obtained by taking the intersection of all the BOMs of electrical control cabinet products.

6. The method for generating the BOM of an electrical control cabinet according to claim 1, characterized in that: In step S1, the sample list database table can be optimized, which specifically includes the following steps: 1) Database table design optimization includes the following steps: Step 1: Select appropriate data types and field lengths to reduce data storage overhead; Step 2: Add BTree indexes and Hash indexes to fields that are frequently used for queries; Step 3: Partition and shard tables that store large amounts of data; 2) Query optimization includes the following steps: Step 1: When querying data in a table, avoid querying all fields in the table and only query the required fields; Step 2: When querying data associated with multiple tables, use a left outer join to associate multiple tables; Step 3: Split complex query statements into multiple simple query statements; 3) Database caching technology to improve data processing includes the following steps Step 1: The client sends the Redis executed commands and stored data to the Redis server via remote communication; Step 2: The Redis server will parse the sent content and perform the corresponding operation; Redis stores the data in the corresponding data structure in the server memory; Step 3: Redis will parse the sent query command, and then find the corresponding value based on the operation type and the key of the data to be found; Step 5: Return the value to the client via remote communication.

7. A computer device comprising a memory and a processor, wherein a computer program is stored in the memory, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

8. A computer storage medium for storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images