Search

CN-121981582-A - Intelligent manufacturing technology application production system

CN121981582ACN 121981582 ACN121981582 ACN 121981582ACN-121981582-A

Abstract

The invention relates to the technical field of intelligent manufacturing, in particular to an intelligent manufacturing technology application production system which comprises an equipment state analysis module, a product quality analysis module and a result evaluation module, wherein the equipment state analysis module is connected with an equipment management module in an intelligent manufacturing production execution system, working state data of production equipment are collected, whether the working state data are normal or not is predicted by using a support vector machine algorithm, the product quality analysis module is connected with the product management module in the intelligent manufacturing production execution system, product data in a production process are extracted, whether products are qualified or not is predicted by using a decision tree algorithm, a final result is determined by using the result evaluation module according to prediction results of the equipment state analysis module and the product quality analysis module, and the result is judged again by optimizing the equipment state analysis module and the product quality analysis module, and when the result is still abnormal, the final result is determined.

Inventors

  • YANG GUOAN
  • YANG GUANGSHENG

Assignees

  • 广东巨基创新科技有限公司

Dates

Publication Date
20260505
Application Date
20240203

Claims (8)

  1. 1. The intelligent manufacturing technology application production system is characterized by comprising an equipment state analysis module (100), a product quality analysis module (200) and a result evaluation module (300), wherein: The equipment state analysis module (100) is connected with an equipment management module in the intelligent manufacturing execution system, wherein the equipment management module monitors the working state of production equipment, collects the working state data of the production equipment, the working state data comprise but are not limited to equipment network state, equipment alarm information and equipment sensor data, and predicts whether the equipment state of the working state data is normal or not by using a support vector machine algorithm; The product quality analysis module (200) is connected with a product management module in the intelligent manufacturing and production execution system, wherein the product management module monitors product data in a production process, extracts product data generated in the production process of the product management module, and predicts whether the product is qualified or not by utilizing a decision tree algorithm; The result evaluation module (300) determines a final result according to the prediction results of the equipment state analysis module (100) and the product quality analysis module (200); When the result of the equipment state is abnormal, adjusting the algorithm parameters of the support vector machine in the equipment state analysis module (100), carrying out optimization treatment by using a cross validation and grid search method, and after the optimization treatment, judging the result again, and when the result of the equipment state is still abnormal, determining the abnormality as a final result; When the result of the equipment state is normal and the product is unqualified, converting a decision tree algorithm segmentation strategy in a product quality analysis module (200), converting the information gain segmentation strategy into a chi-square statistic segmentation strategy, carrying out result judgment again after optimizing treatment, and determining that the abnormality is a final result when the product is still unqualified; And when the result of the equipment state is normal and the product is qualified, determining that the result is normal as a final result.
  2. 2. The intelligent manufacturing technology application production system according to claim 1, wherein the equipment state analysis module (100) comprises an equipment data collection unit (101) and an equipment data analysis unit (102), the equipment data collection unit (101) is connected with an equipment management module in the intelligent manufacturing execution system, the equipment management module monitors the working state of the production equipment and collects the working state data of the production equipment, the working state data comprises but is not limited to equipment network state, equipment alarm information and equipment sensor data, and the equipment data analysis unit (102) predicts whether the working state data is normal or not by using a support vector machine algorithm and sends a prediction result to the result evaluation module (300).
  3. 3. The intelligent manufacturing technology application production system according to claim 2, wherein the product quality analysis module (200) comprises a product data collection unit (201) and a product data analysis unit (202), the product data collection unit (201) is connected with a product management module in the intelligent manufacturing execution system, the product management module monitors product data in a production process and extracts product data generated in the production process of the product management module, and the product data analysis unit (202) predicts whether a product is qualified or not by using a decision tree algorithm on the product data and sends a prediction result to the result evaluation module (300).
  4. 4. The intelligent manufacturing technology application production system according to claim 3, wherein the result evaluation module (300) comprises a result determination unit (301) and a method optimization unit (302), the result determination unit (301) receives the device data analysis unit (102) and the product data analysis unit (202) and determines a final result according to the prediction results of the device state analysis module (100) and the product quality analysis module (200), and the method optimization unit (302) adjusts the support vector machine algorithm parameters in the device data analysis unit (102) and converts the decision tree algorithm segmentation strategy in the product data analysis unit (202).
  5. 5. The intelligent manufacturing technology application production system according to claim 2, wherein the device data analysis unit (102) predicts whether the device state is normal for the operation state data by using a support vector machine algorithm, and specifically comprises: The support vector machine algorithm comprises a C parameter and a gamma parameter, wherein the C parameter controls the punishment degree of misclassified samples, and the gamma parameter controls the influence range of a single training sample; The method comprises the steps of obtaining a decision boundary, separating samples of different categories, defining an objective function comprising a radial basis function, solving the objective function by using a gradient descent algorithm, finding an optimal solution, selecting a support vector in a training data set, wherein the support vector is a sample point closest to a hyperplane, and determining the position and shape of the hyperplane; The working state data sent by the equipment data collecting unit (101) is input into the trained model, when the output result is 0, one side of the hyperplane is used for indicating that the current working equipment is in a normal state, and when the output result is 1, the other side of the hyperplane is used for indicating that the current working equipment is in an abnormal state.
  6. 6. The intelligent manufacturing technology application production system according to claim 3, wherein the product data analysis unit (202) predicts whether the product data is qualified or not by using a decision tree algorithm, and specifically comprises: Selecting a feature with the maximum information gain in the data as a root node, continuously dividing each sub-node under the root node according to the information gain, and continuously iterating until samples in the nodes belong to the same category, and stopping growing; Entropy represents a measure of data non-purity, and for each class in the data set, the product of the probability of occurrence and the logarithm of this probability is calculated, all of these products are added and take negative values; For each feature in the data set, dividing the data set into a plurality of subsets according to different values of the feature, calculating the entropy of each subset, and carrying out weighted summation by the size of each subset to obtain the total entropy after the data set is divided according to the feature; The information gain is equal to the difference between the original entropy of the data set and the conditional entropy divided according to the feature, and the difference value indicates the degree of data randomness reduction by considering the feature; comparing the information gains of all the features, and selecting the feature with the maximum information gain as a division point; Segmenting the dataset with the selected features, repeating the above process for each subset until all instances belong to the same category; The method comprises the steps of starting from a root node, selecting a corresponding branch according to a judgment condition, entering a next node, selecting the corresponding branch according to the judgment condition of a current node and the value of the characteristic value, entering the next node, recording current passing node data, and continuously selecting the branch from the root node to a leaf node according to the value of the characteristic value until the leaf node is reached, wherein the leaf node corresponds to a result of whether a product is qualified or not, and outputting the result of whether the product is qualified or not and the recorded node data.
  7. 7. The intelligent manufacturing technology application production system according to claim 4, wherein the method optimizing unit (302) adjusts the algorithm parameters of the support vector machine in the device data analyzing unit (102), and specifically comprises: Searching the optimal super-parameter combination by using a cross-validation and grid search method, and defining a continuous value range of C and gamma with the value range of 10 power; Creating a super-parameter combination space, listing all combinations of C and gamma, generating the combination space by a grid search method, and generating a parameter combination for each combination of C value and gamma value; for each parameter combination, cross-validation is used to evaluate the performance of the model, the cross-validation divides the data set into k subsets, performs an iterative loop, each time using k-1 subsets as training sets, the remaining subset as test set; And for each parameter combination, calculating the accuracy of the model on the test set, and selecting the parameter combination with the highest accuracy according to the cross-validation result, so as to perform optimization processing.
  8. 8. The intelligent manufacturing technology application production system according to claim 4, wherein the method optimizing unit (302) converts a decision tree algorithm segmentation strategy in the product data analyzing unit (202), converts an information gain segmentation strategy in the decision tree algorithm into a chi-square statistic segmentation strategy, obtains the result by dividing the sum of the square of the difference between the observed frequency and the expected frequency of each category by the expected frequency, obtains a chi-square value which is larger and indicates that the difference between the actual distribution and the expected distribution is larger, selects the characteristic with the largest chi-square statistic to segment, and repeats the process to construct each branch of the decision tree, so that the optimization processing is performed.

Description

Intelligent manufacturing technology application production system Technical Field The invention relates to the technical field of intelligent manufacturing, in particular to an intelligent manufacturing technology application production system. Background Conventional production systems often lack intelligent data analysis and real-time monitoring capabilities. In such systems, monitoring of the status of the equipment often relies on periodic manual inspection and predetermined maintenance schedules. This manual-dependent approach is passive in many cases, and may be taken after equipment failure, resulting in longer downtime and higher repair costs. The inspection of the product quality also depends on manual sampling inspection, the data in the manufacturing process are often underutilized, and the product quality analysis can influence the accuracy of the result due to human errors and sampling deviation. In such an environment, predictions of device states lack accurate mathematical models and efficient data processing mechanisms. Even if data is recorded, it is usually only used for post-hoc reporting, not for immediate feedback and adjustment. The product quality problem can be found only when the product is finished or a large-scale problem occurs, and immediate correction cannot be achieved, so that the overall production efficiency and the product quality are affected, and the probability of waste and reworking is increased. Moreover, such conventional systems often lack automated data analysis tools that cannot learn and predict potential failure modes and quality trends from historical data. Thus, optimization of the production flow is slow, and response to innovations is limited, by experience and intuition, rather than data-driven insight. In summary, the traditional system has obvious disadvantages of slow response, low accuracy, high cost, poor efficiency, passivity for equipment prediction, hysteresis for product quality management and low-efficiency utilization of data, and remarkably hinders the improvement of production efficiency and product quality. Lacking timely, data-driven insight and optimization capability, these disadvantages can lead to enterprises facing greater risks and challenges in increasingly competitive market environments. Disclosure of Invention The invention aims to provide an intelligent manufacturing technology application production system for solving the problems in the background technology. In order to achieve the above purpose, the invention provides the following technical scheme that the intelligent manufacturing technology application production system comprises an equipment state analysis module, a product quality analysis module and a result evaluation module, wherein: The equipment state analysis module is connected with an equipment management module in the intelligent manufacturing production execution system, wherein the equipment management module monitors the working state of production equipment, collects the working state data of the production equipment, the working state data comprise but are not limited to equipment network state, equipment alarm information and equipment sensor data, and predicts whether the working state data are normal or not by utilizing a support vector machine algorithm; The product quality analysis module is connected with a product management module in the intelligent manufacturing production execution system, wherein the product management module monitors product data in a production flow, extracts the product data generated in the production flow of the product management module, and predicts whether the product is qualified or not by utilizing a decision tree algorithm; The result evaluation module determines a final result according to the prediction results of the equipment state analysis module and the product quality analysis module; When the result of the equipment state is abnormal, adjusting the algorithm parameters of the support vector machine in the equipment state analysis module, performing optimization treatment by using a cross validation and grid search method, and after the optimization treatment, judging the result again, and when the result of the equipment state is still abnormal, determining the abnormality as a final result; when the result of the equipment state is normal and the product is unqualified, converting a decision tree algorithm segmentation strategy in a product quality analysis module, converting the information gain segmentation strategy into a chi-square statistic segmentation strategy, carrying out result judgment again after optimizing treatment, and determining that the abnormality is a final result when the product is still unqualified; And when the result of the equipment state is normal and the product is qualified, determining that the result is normal as a final result. The device state analysis module comprises a device data collection unit and a device data analysis uni