CN119626404A - A precise batching system for casting aluminum alloy - Google Patents
A precise batching system for casting aluminum alloy Download PDFInfo
- Publication number
- CN119626404A CN119626404A CN202411688523.XA CN202411688523A CN119626404A CN 119626404 A CN119626404 A CN 119626404A CN 202411688523 A CN202411688523 A CN 202411688523A CN 119626404 A CN119626404 A CN 119626404A
- Authority
- CN
- China
- Prior art keywords
- batching
- aluminum alloy
- historical data
- module
- model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C60/00—Computational materials science, i.e. ICT specially adapted for investigating the physical or chemical properties of materials or phenomena associated with their design, synthesis, processing, characterisation or utilisation
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/10—Pre-processing; Data cleansing
- G06F18/15—Statistical pre-processing, e.g. techniques for normalisation or restoring missing data
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/211—Selection of the most significant subset of features
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/27—Regression, e.g. linear or logistic regression
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/20—Identification of molecular entities, parts thereof or of chemical compositions
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/70—Machine learning, data mining or chemometrics
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Crystallography & Structural Chemistry (AREA)
- Chemical & Material Sciences (AREA)
- Databases & Information Systems (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Probability & Statistics with Applications (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Manufacture And Refinement Of Metals (AREA)
Abstract
本发明公开了一种铸造铝合金精确配料系统，括历史数据收集模块，用于收集铝合金铸造时的历史数据；历史数据处理模块，用于对收集到的历史数据进行清洗和归一化处理；特征选择模块，用于通过相关性分析和特征重要性评估；模型训练模块，用于基于卷积神经网络CNN进行配料比例的预测模型的训练；反馈控制模块，用于将预测结果形成输出指令；执行控制模块，用于根据输出指令，实现配料的称量和自动化控制报警。通过建立模型训练模块预测配料比例，基于均方误差计算回归损失函数，迭代训练使回归损失函数最小化，进而保证预测的配料比例的准确性；并通过反馈控制模块和执行控制模块，实现自动配料和控制。
The invention discloses a precise batching system for casting aluminum alloy, including a historical data collection module for collecting historical data during aluminum alloy casting; a historical data processing module for cleaning and normalizing the collected historical data; a feature selection module for evaluating the importance of features through correlation analysis; a model training module for training a prediction model for batching ratio based on a convolutional neural network (CNN); a feedback control module for forming an output instruction from the prediction result; and an execution control module for realizing batching weighing and automatic control alarm according to the output instruction. The batching ratio is predicted by establishing a model training module, and the regression loss function is calculated based on the mean square error. The regression loss function is minimized by iterative training, thereby ensuring the accuracy of the predicted batching ratio; and automatic batching and control are realized through the feedback control module and the execution control module.
Description
Technical Field
The invention relates to a batching system, in particular to a precise batching system for cast aluminum alloy.
Background
The casting aluminum alloy ingredients refer to materials which are needed to be put in during the casting of a certain specific aluminum alloy, so that the chemical elements of the aluminum alloy can reach the specified standard, and the total weight of the materials is as close to the weight of the casting aluminum alloy as possible; the quality of aluminum alloy casting is affected by various factors, such as the fineness of a casting process, the quality of casting equipment, the selection of raw materials, a batching scheme and the like, wherein the design of the batching scheme is a key point in the field of aluminum alloy casting;
during the batching of aluminum alloys, operators typically perform calculations and review operations manually according to established batching rules. However, due to the complexity of the dosing rules and the multi-condition limitations, coupled with the technical difficulty of recalculating when checking inaccurately, it is often difficult for an operator to calculate the dosing value accurately and without error in a short time. This results in numerical deviations that in turn affect the high quality casting production of aluminum alloys and increase the ageing costs of the industrial production.
Disclosure of Invention
In order to solve the defects of the technology, the invention provides a precise batching system for casting aluminum alloy.
In order to solve the technical problems, the technical scheme adopted by the invention is that the accurate batching system for casting aluminum alloy comprises a historical data collection module, a processing module and a processing module, wherein the historical data collection module is used for collecting historical data of aluminum alloy during casting, and the historical data comprises raw material component data, batching proportion and production environment parameters;
The historical data processing module is used for cleaning and normalizing the collected historical data to eliminate the influence of abnormal values, and the normalization processing formula is as follows:
Wherein x' is normalized data, x i is the value of the ith historical data, x max is the maximum value in the historical data, and x min is the minimum value in the historical data;
the feature selection module is used for selecting the feature with the largest influence on the quality of the aluminum alloy as the model input through correlation analysis and feature importance evaluation;
The model training module is used for training a prediction model of the batching proportion based on the convolutional neural network CNN;
The feedback control module is used for collecting raw material components and production environment parameters in real time, predicting the proportion of ingredients, forming an output instruction from a prediction result, and sending the output instruction to the execution control module;
and the execution control module is used for controlling the batching scale and the conveying equipment based on the PLC controller according to the output instruction, so as to realize weighing of the batching and automatic control alarm.
Further, the calculation formula of the correlation coefficient r in the correlation analysis is as follows:
Wherein x i is the value of the i-th history data; is the average of this historical data.
Further, the feature importance assessment method is to assess the importance of features based on a tree model, i.e. calculate the number of times each feature is used in the tree model or the importance of the feature used for splitting based on XGBoost model.
Further, evaluating feature importance based on XGBoost model, comprising the steps of:
step 1, constructing a data set according to historical data;
Step 2, dividing the data set into a feature variable and a target variable, and dividing the training set and the testing set;
initializing XGBoost models, and optimizing super parameters based on grid search;
Step 4, based on the five-fold cross verification technology, finding out the optimal super-parameters, and evaluating XGBoost the performance of the model according to the negative mean square error;
Step 5, acquiring contribution degrees of all the features, and sorting the contribution degrees in a descending order;
Step 6, sorting the names and the corresponding contribution of the features according to descending order, and storing;
and 7, evaluating the importance of the features according to the storage result.
Further, a regression loss function is calculated based on a mean square error MSE in a model training module, an Adam optimizer is used for updating model parameters, and the model parameters are adjusted through iterative training, so that the regression loss function is minimized;
The mean square error MSE is calculated as:
Where n is the number of samples, x i is the true ingredient ratio for the ith sample, and y i is the predicted ingredient ratio for the ith sample.
Further, the model training module performs prediction model optimization based on the Dropout layer after training the prediction model for preventing the prediction model from overfitting.
Further, the feedback control module comprises a temperature sensor for monitoring temperature change in the aluminum alloy smelting process;
the component analyzer is used for analyzing chemical components in the aluminum alloy raw materials, and the accuracy of the batching proportion is ensured by accurately measuring the content of various elements.
Further, the execution control module comprises a weight sensor which is used for being installed on the batching scale, collecting weight data of various raw materials in real time and ensuring batching accuracy.
The invention discloses an accurate batching system for casting aluminum alloy, which predicts batching proportion by a model training module, calculates regression loss function based on mean square error, minimizes regression loss function by iterative training, further ensures accuracy of the predicted batching proportion, reduces numerical deviation, improves high-quality casting productivity of aluminum alloy, realizes automatic batching and control by a feedback control module and an execution control module, and saves manpower.
Drawings
FIG. 1 is a system architecture diagram of the present invention.
FIG. 2 is a flow chart of steps of an embodiment of the present invention.
Detailed Description
The invention will be described in further detail with reference to the drawings and the detailed description.
Embodiment one:
the accurate batching system for casting aluminum alloy as shown in fig. 1 and 2 comprises a historical data collection module, a batch mixing module and a batch mixing module, wherein the historical data collection module is used for collecting historical data of aluminum alloy casting, and the historical data comprises raw material composition data, a batching proportion and production environment parameters, wherein the production environment parameters comprise a production temperature and a cooling speed;
The historical data processing module is used for cleaning and normalizing the collected historical data to eliminate the influence of abnormal values, and the normalization processing formula is as follows:
Wherein x' is normalized data, x i is the value of the ith historical data, x max is the maximum value in the historical data, and x min is the minimum value in the historical data;
the normalization processing refers to processing that a column of data is changed into a certain fixed interval, and the data is mapped to a range of 0-1 for processing in order to be more convenient for processing the data;
The characteristic selection module is used for selecting the characteristic with the greatest influence on the quality of the aluminum alloy as the model input through correlation analysis and characteristic importance evaluation, wherein the correlation analysis aims at evaluating the relation strength between each characteristic and the quality of the aluminum alloy, and identifying the characteristic highly correlated with the quality index of the aluminum alloy by calculating the correlation coefficient between the characteristic and the quality index, wherein the value of the correlation coefficient is between-1 and 1, and the closer the value is 1 or-1, the stronger the correlation between the characteristic and the quality index of the aluminum alloy is represented;
the calculation formula of the correlation coefficient r in the correlation analysis is as follows:
Wherein x i is the value of the i-th history data; is an average of the historical data;
Feature importance assessment is to evaluate the influence degree of features on a target variable, namely the importance degree of the features in a model, calculate the importance based on the features of a tree model, calculate the number of times each feature is used in the tree model or the importance used for splitting based on XGBoost models, and measure the importance of the features. The more features are used in the XGBoot model to construct a decision tree therein, the higher the feature importance is, the attribute importance is obtained by computing and ordering each attribute in the dataset;
The feature importance assessment method is to assess the importance of the features based on a tree model, namely, calculate the number of times each feature is used in the tree model or the importance of the importance measurement features used for splitting based on XGBoost models;
Evaluating feature importance based on XGBoost model, comprising the steps of:
step 1, constructing a data set according to historical data;
Step 2, dividing the data set into a feature variable and a target variable, and dividing the training set and the testing set;
initializing XGBoost models, and optimizing super parameters based on grid search;
Step 4, based on the five-fold cross verification technology, finding out the optimal super-parameters, and evaluating XGBoost the performance of the model according to the negative mean square error;
Step 5, acquiring contribution degrees of all the features, and sorting the contribution degrees in a descending order;
Step 6, sorting the names and the corresponding contribution of the features according to descending order, and storing;
step 7, evaluating the importance of the features according to the storage result;
The model training module is used for training a prediction model of the batching proportion based on the convolutional neural network CNN;
Calculating a regression loss function based on a mean square error MSE in a model training module, using an Adam optimizer to update model parameters, and adjusting the model parameters through iterative training to minimize the regression loss function;
The mean square error MSE is calculated as:
Wherein n is the number of samples, x i is the true ingredient proportion of the ith sample, y i is the predicted ingredient proportion of the ith sample;
The model training module is used for carrying out prediction model optimization based on the Dropout layer after training the prediction model so as to prevent the prediction model from being fitted excessively;
The feedback control module is used for collecting raw material components and production environment parameters in real time, predicting the proportion of ingredients, forming an output instruction from a prediction result, and sending the output instruction to the execution control module;
The feedback control module comprises a temperature sensor for monitoring the temperature change in the aluminum alloy smelting process;
The component analyzer is used for analyzing chemical components in the aluminum alloy raw materials, and the accuracy of the batching proportion is ensured by accurately measuring the content of various elements;
the execution control module is used for controlling the batching scale and the conveying equipment based on the PLC controller according to the output instruction, so as to realize weighing of the batching and automatic control alarm;
the execution control module comprises a weight sensor and is used for being installed on a batching scale, acquiring weight data of various raw materials in real time and ensuring batching accuracy.
Embodiment two:
The convolutional neural network CNN is a deep learning model or a multi-layer perceptron similar to an artificial neural network, and comprises a data input layer, a convolutional layer, a pooling layer, a full-connection layer and an output layer, wherein a prediction model of the batching proportion is built based on the convolutional neural network CNN, and the model comprises the following structures:
The method comprises the steps of taking preprocessed raw material component data and production environment as input of a prediction model, extracting local features in the input data through a convolution layer through a convolution kernel, capturing spatial relation and mode in the data through a network, carrying out nonlinear transformation by the convolution layer after receiving an activation function, namely a ReLU function, so as to enhance the expression capability of the prediction model, effectively reducing the spatial dimension of the data through an average pooling layer in the pooling layer, carrying out global integration on the local features through a full-connection layer after the extraction of the convolution layer and the dimension reduction treatment of the pooling layer, generating a final batching proportion prediction result, namely multiplying the input feature vector by a weight matrix, adding a bias parameter, and outputting an accurate batching proportion, wherein the formula of the full-connection layer is as follows:
y=σ(ω*x+b);
wherein y is an output batching proportion vector, sigma is a ReLU activation function, x is an input vector, omega is a weight matrix, and b is a bias parameter.
According to the invention, the model training module is constructed, so that the proportioning proportion can be accurately predicted. This process involves a mean square error based calculation to derive a regression loss function. By constantly iterating the training, we strive to minimize this regression loss function to ensure that the predicted ingredient ratios are highly accurate. The method obviously reduces the numerical deviation, thereby improving the efficiency of high-quality casting production of the aluminum alloy.
In addition, the invention introduces a feedback control module and an execution control module, and the two modules work cooperatively, thereby realizing the automation of the batching process. By the mode, the proportion of ingredients can be accurately controlled, the production process can be monitored and adjusted in real time, and the stability and consistency of the production process are ensured. The automatic control system greatly saves human resources, reduces errors and uncertainty of manual operation, and further improves production efficiency and product quality.
The above embodiments are not intended to limit the present invention, and the present invention is not limited to the above examples, but is also intended to be limited to the following claims.
Claims (8)
1. The accurate batching system for casting the aluminum alloy is characterized by comprising a historical data collection module, a batching module and a batching module, wherein the historical data collection module is used for collecting historical data of the aluminum alloy during casting, and the historical data comprises raw material component data, batching proportions and production environment parameters;
The historical data processing module is used for cleaning and normalizing the collected historical data to eliminate the influence of abnormal values, and the normalization processing formula is as follows:
Wherein x' is normalized data, x i is the value of the ith historical data, x max is the maximum value in the historical data, and x min is the minimum value in the historical data;
the feature selection module is used for selecting the feature with the largest influence on the quality of the aluminum alloy as the model input through correlation analysis and feature importance evaluation;
The model training module is used for training a prediction model of the batching proportion based on the convolutional neural network CNN;
The feedback control module is used for collecting raw material components and production environment parameters in real time, predicting the proportion of ingredients, forming an output instruction from a prediction result, and sending the output instruction to the execution control module;
and the execution control module is used for controlling the batching scale and the conveying equipment based on the PLC controller according to the output instruction, so as to realize weighing of the batching and automatic control alarm.
2. The precise batching system for cast aluminum alloy according to claim 1, wherein the calculation formula of the correlation coefficient r in the correlation analysis is:
Wherein x i is the value of the i-th history data; is the average of this historical data.
3. The precision batching system for cast aluminum alloy according to claim 1, wherein the feature importance assessment method is to assess the importance of features based on a tree model, i.e. calculate the number of times each feature is used in the tree model or the importance of the feature used for splitting based on XGBoost model.
4. The precision batching system for cast aluminum alloy according to claim 3, wherein the feature importance is evaluated based on XGBoost model, comprising the steps of:
step 1, constructing a data set according to historical data;
Step 2, dividing the data set into a feature variable and a target variable, and dividing the training set and the testing set;
initializing XGBoost models, and optimizing super parameters based on grid search;
Step 4, based on the five-fold cross verification technology, finding out the optimal super-parameters, and evaluating XGBoost the performance of the model according to the negative mean square error;
Step 5, acquiring contribution degrees of all the features, and sorting the contribution degrees in a descending order;
Step 6, sorting the names and the corresponding contribution of the features according to descending order, and storing;
and 7, evaluating the importance of the features according to the storage result.
5. The precise batching system for cast aluminum alloy according to claim 1, wherein the model training module calculates a regression loss function based on a mean square error MSE, and uses an Adam optimizer to update model parameters, and adjusts the model parameters through iterative training to minimize the regression loss function;
The mean square error MSE is calculated as:
Where n is the number of samples, x i is the true ingredient ratio for the ith sample, and y i is the predicted ingredient ratio for the ith sample.
6. The precision batching system for cast aluminum alloy according to claim 1, wherein the model training module performs prediction model optimization based on Dropout layer after training the prediction model for preventing the prediction model from overfitting.
7. The precision batching system for casting aluminum alloy according to claim 1, wherein the feedback control module comprises a temperature sensor for monitoring temperature variation in the melting process of aluminum alloy;
the component analyzer is used for analyzing chemical components in the aluminum alloy raw materials, and the accuracy of the batching proportion is ensured by accurately measuring the content of various elements.
8. The precise batching system for casting aluminum alloy according to claim 1, wherein the execution control module comprises a weight sensor which is used for being installed on a batching scale, collecting weight data of various raw materials in real time, and ensuring batching accuracy.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411688523.XA CN119626404A (en) | 2024-11-25 | 2024-11-25 | A precise batching system for casting aluminum alloy |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411688523.XA CN119626404A (en) | 2024-11-25 | 2024-11-25 | A precise batching system for casting aluminum alloy |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN119626404A true CN119626404A (en) | 2025-03-14 |
Family
ID=94907954
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202411688523.XA Pending CN119626404A (en) | 2024-11-25 | 2024-11-25 | A precise batching system for casting aluminum alloy |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN119626404A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN120382570A (en) * | 2025-06-27 | 2025-07-29 | 甘肃展通管业有限责任公司 | A method and system for automatic batching control of PVC pipe production |
-
2024
- 2024-11-25 CN CN202411688523.XA patent/CN119626404A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN120382570A (en) * | 2025-06-27 | 2025-07-29 | 甘肃展通管业有限责任公司 | A method and system for automatic batching control of PVC pipe production |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113486457A (en) | Die casting defect prediction and diagnosis system | |
| CN119626404A (en) | A precise batching system for casting aluminum alloy | |
| CN108460213A (en) | Based on the T-S models of multi-cluster prototype to the prediction technique and program of gas flowrate in bosh | |
| CN113591374B (en) | Cement finished product specific surface area prediction method based on convolution simple circulation unit network | |
| CN119479857A (en) | Method, system and equipment for optimizing batching parameters in industrial silicon smelting process | |
| CN113281229A (en) | Multi-model self-adaptive atmosphere PM based on small samples2.5Concentration prediction method | |
| CN112541296A (en) | SO2 prediction method based on PSO-LSSVM | |
| CN118761522A (en) | Production quality prediction method based on kernel principal component analysis and multivariate linear regression | |
| CN113297542B (en) | A method and device for predicting raw ball ratio of a ball making machine | |
| CN116306220A (en) | Rolling force prediction method based on quantum particle swarm algorithm and BP neural network | |
| CN113609772A (en) | Cement product specific surface area prediction method based on convolution quasi-cyclic neural network | |
| CN116258087B (en) | Matte grade soft measurement method and device, electronic equipment and storage medium | |
| JP5962290B2 (en) | Steel heat transfer coefficient prediction apparatus and cooling control method | |
| CN114528764A (en) | Soft measurement modeling method and device based on integral optimization and instant learning | |
| CN114004044A (en) | Quick identification method for thermal error of machine tool spindle based on temperature sensitive point | |
| CN114066135A (en) | Automatic optimization method and system for water quality model parameters | |
| CN110033175A (en) | A kind of flexible measurement method based on integrated multicore Partial Least-Squares Regression Model | |
| CN119442858A (en) | A method for detecting and predicting equipment rule anomalies in digital twins | |
| CN119109772A (en) | A method for identifying communication anomalies of a numerically controlled power supply | |
| CN118798421A (en) | A method for predicting water ecological environment monitoring data based on adaptive model and dynamic data fusion | |
| CN118483968A (en) | Concrete accelerator production quality control system and method | |
| JPH06332506A (en) | Non-linear controller | |
| CN120403036B (en) | Method for calibrating sub-term energy consumption reference value based on heating ventilation air conditioning equipment | |
| CN119517213B (en) | Machine learning-based steel pipe post-production mechanical property prediction method | |
| CN119359154B (en) | Intelligent lubricating oil production control method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |