Industrial product quality tracing and optimizing system based on industrial internet
Through the full-process quality traceability architecture based on the industrial Internet, combined with the Internet of Things, blockchain and deep learning, the problems of imperfect quality traceability and insufficient data analysis have been solved, and accurate traceability and quality optimization of the entire life cycle of the product has been achieved, and the company's product quality and market competitiveness have been improved.
Patent Information
- Application Number
- CN202510550038.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-29
- Publication Date
- 2025-08-12
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The quality traceability of traditional industrial products is incomplete, the data collection and analysis capabilities are insufficient, and the lack of coordination among various departments has led to inefficient quality problem solving and the lack of targeted and effective optimization measures.
Based on the full-process quality traceability architecture of the industrial Internet, combining the Internet of Things and blockchain technology to realize real-time data collection and storage, multi-source heterogeneous data fusion and deep learning to analyze quality data, use reinforcement learning to optimize decision-making, integrate the processes of various departments through collaborative management platforms, and provide a visual monitoring platform.
It realizes accurate traceability of the entire life cycle of the product, improves the quality data analysis capabilities, promotes departmental collaboration, and improves the scientific nature of quality optimization decisions and monitoring efficiency.
Smart Images
Figure CN120471630A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of industrial automation, and in particular relates to an industrial product quality tracing and optimization system based on the Industrial Internet. Background Art
[0002] In today's increasingly competitive global market, product quality has become a core competitive advantage for businesses' survival and development. For industrial companies, effective product quality control and traceability not only enables timely identification and resolution of quality issues, minimizing quality losses, but also strengthens consumer trust in products. However, traditional industrial product quality control methods currently suffer from numerous issues, severely hindering companies' efforts to improve product quality and expand their market share.
[0003] Imperfect quality traceability systems: Traditional product quality traceability relies primarily on manual records and paper documents, resulting in fragmented and incomplete information and difficulty in accurately tracing the entire product lifecycle. Once a quality issue arises, companies are unable to quickly and accurately identify the root cause. The tracing process is cumbersome and time-consuming, resulting in inefficient problem-solving and impacting production and sales. Furthermore, the difficulty in ensuring the accuracy and reliability of traceability information complicates quality improvement efforts.
[0004] Inadequate quality data collection and analysis capabilities: Existing quality data collection methods mostly rely on manual spot checks and simple testing equipment. Data collection is infrequent, coverage is narrow, and the system fails to fully reflect product quality. Furthermore, when analyzing collected data, simple statistical methods are often used, lacking the ability to deeply mine and analyze the data. This makes it difficult to extract valuable information from massive amounts of data and identify underlying patterns and trends in quality issues, resulting in a lack of targeted and effective quality improvement measures.
[0005] Lack of coordination in quality optimization: The industrial production process involves multiple departments and links, such as raw material procurement, production and processing, and quality inspection. However, there is currently a lack of effective coordination mechanisms between departments in quality control, resulting in poor information communication and independent operations. When quality issues arise, departments often blame each other and fail to work together to resolve them. Furthermore, when implementing quality optimization, there is a lack of systematic consideration of the entire production process, making it difficult for optimization measures to achieve the desired results. Summary of the Invention
[0006] This invention provides an industrial product quality traceability and optimization system based on the Industrial Internet. This system aims to achieve quality traceability throughout the entire life cycle of industrial products by leveraging advanced Industrial Internet technologies and innovative modules. It also continuously optimizes product quality in a data-driven manner, thereby improving the product quality and market competitiveness of enterprises. The system includes:
[0007] The full-process quality traceability architecture based on the Industrial Internet deploys IoT devices in all production links to collect quality-related data in real time, transmits it to the cloud-based quality traceability platform via the Industrial Internet, and uses blockchain technology to encrypt and store data, thus achieving quality traceability throughout the product life cycle.
[0008] The quality analysis module based on multi-source heterogeneous data fusion adopts the following innovative module formula in the quality analysis process: Among them, QI is the comprehensive evaluation value of quality indicators, which is used to comprehensively measure the product quality status; Di represents the quality-related data collected by the i-th data source, which includes multi-source heterogeneous data such as production data, raw material data, and quality inspection data, and n is the total number of data sources; fi(Di) is the feature extraction and quantification function for the i-th data source data, which is trained based on deep learning and data mining techniques and can convert raw data into quantitative values that contribute to quality indicators. For example, through in-depth analysis of production temperature data, its quantitative impact on product quality stability is obtained. wi is the weight coefficient of the i-th data source data. Its value is determined based on the importance of different data sources in the product quality formation process. It is determined by fitting a large amount of experimental data and actual production experience, and is used to balance the contribution of data from different data sources in the comprehensive evaluation of quality indicators. This module integrates and processes multi-source heterogeneous quality data from different equipment and departments, and uses deep learning and data mining techniques to conduct in-depth analysis of the integrated data, explore potential patterns and laws in the data, establish a quality analysis model, and accurately identify the type, cause, and influencing factors of quality problems.
[0009] The quality optimization decision module based on reinforcement learning abstracts the quality optimization process into a Markov decision process. By defining the state space, action space, and reward function, it learns the optimal quality optimization decision strategy to achieve continuous optimization of product quality.
[0010] The quality control platform based on collaborative management integrates the quality-related business processes of various departments within the enterprise, realizes real-time information sharing and collaborative work, and has functions such as quality task allocation, progress tracking and performance evaluation;
[0011] The visual quality traceability and optimization monitoring platform displays product quality traceability information, quality analysis results, quality optimization decision suggestions, etc. through a graphical interface, and supports remote operation and control.
[0012] Furthermore, in the full-process quality traceability architecture based on the industrial Internet, IoT devices include sensors, RFID tags, etc., which transmit data to the cloud platform through wireless communication technologies such as industrial Ethernet and 5G.
[0013] Furthermore, in the quality optimization decision module based on reinforcement learning, the state space includes information such as product quality indicators, production process parameters, and equipment operating status; the action space includes decisions such as adjustment of production process parameters and equipment maintenance and replacement; and the reward function is set based on the improvement of product quality and reduction of production costs.
[0014] Furthermore, the quality control platform based on collaborative management integrates the quality-related business processes of the company's internal raw material procurement, production and processing, quality inspection and other departments to achieve real-time information sharing and collaborative work.
[0015] Furthermore, the visual quality traceability and optimization monitoring platform is developed using Web technology, and its interface includes modules such as quality traceability information display, quality analysis result display, quality optimization decision suggestion display, and equipment operation status monitoring.
[0016] Furthermore, in the full-process quality traceability architecture based on the industrial Internet, the data blocks stored by blockchain technology contain timestamps, data content and hash values of the previous data blocks, ensuring the authenticity and non-tamperability of the data.
[0017] Furthermore, after data collection, the quality analysis module based on multi-source heterogeneous data fusion performs pre-processing operations such as data cleaning, format conversion, and normalization to eliminate data differences and noise.
[0018] Furthermore, during the training process, the reinforcement learning-based quality optimization decision module selects an action according to the current state, and after executing the action, updates the strategy based on the reward and new state feedback from the environment to maximize the cumulative reward.
[0019] Furthermore, the quality control platform based on collaborative management establishes a quality performance evaluation system to evaluate and assess the quality work of each department and personnel, and implement corresponding incentive measures.
[0020] Furthermore, the visual quality traceability and optimization monitoring platform supports remote monitoring and management of production equipment and quality inspection equipment, improving the flexibility and efficiency of quality control.
[0021] Beneficial effects
[0022] Achieve full-process quality traceability: Through the full-process quality traceability architecture based on the Industrial Internet, we can achieve full life cycle quality traceability of products from raw material procurement to finished product delivery, ensuring the accuracy and reliability of traceability information, and providing strong support for the rapid location and resolution of quality issues.
[0023] Improve quality data analysis capabilities: The quality analysis module based on multi-source heterogeneous data fusion can conduct in-depth analysis of massive quality data, explore potential patterns and trends in the data, accurately identify the causes and influencing factors of quality problems, and provide a scientific basis for quality improvement.
[0024] Optimize quality control decisions: The quality optimization decision module based on reinforcement learning can automatically adjust the quality optimization decision strategy according to the product quality status and changes in the production environment, achieve continuous optimization of product quality, and improve product quality and the economic benefits of the enterprise.
[0025] Enhance departmental collaboration: The quality control platform based on collaborative management integrates the quality-related tasks of various departments within the enterprise.
[0026] It streamlines business processes, realizes real-time information sharing and collaborative work, effectively promotes collaboration among departments, and improves the efficiency and effectiveness of quality control.
[0027] Convenient quality monitoring and management: The visual quality traceability and optimization monitoring platform provides managers with an intuitive and convenient monitoring and management interface, supports remote operation and control, and improves the flexibility and efficiency of quality control. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] Figure 1 Flowchart of system operation principle. DETAILED DESCRIPTION
[0029] Example 1
[0030] (1) Implementation of a full-process quality traceability architecture based on the Industrial Internet
[0031] IoT device deployment and data collection: Various IoT devices are deployed in key stages of automotive production, including stamping, welding, painting, and final assembly. Pressure sensors are installed in the stamping shop to monitor press pressure in real time with an accuracy of ±0.1 MPa, ensuring the quality of stamped parts. Vision sensors are used in the welding shop to capture 20 frames of weld point images per second to monitor weld quality. RFID tags are used in the raw material warehouse to identify each batch of steel and parts. The devices collect data at a set frequency and transmit it to a cloud-based quality traceability platform via a 5G network.
[0032] Blockchain Data Storage and Management: The cloud-based quality traceability platform uses blockchain technology to store data. Each data block contains a timestamp, production process data, and a hash value of the previous data block. For example, a production data block for a batch of automobile engine cylinder blocks, with timestamps accurate to the second, records data from stamping, welding, machining, and other stages. A blockchain encryption module ensures that this data cannot be tampered with. Enterprises can quickly access quality traceability information for the cylinder blocks throughout their lifecycle by entering the vehicle identification number.
[0033] Traceability Information Display and Application: The visual quality traceability and optimization monitoring platform displays traceability information in a graphical interface. In the automotive after-sales service, if a vehicle experiences an engine failure, maintenance personnel can enter the vehicle identification number (VIN) on the platform to clearly view the source of the engine block's raw materials and quality inspection results for each stage of the production process, quickly locating the root cause of the problem. If a problem with a particular batch of steel is discovered, the company can quickly recall the affected batch of vehicles, mitigating quality risks.
[0034] (2) Implementation of quality analysis module based on multi-source heterogeneous data fusion
[0035] Data collection and preprocessing: Collect heterogeneous data from multiple sources, including production equipment, raw material suppliers, and quality inspection departments. Clean production equipment data to remove outliers caused by equipment failures; convert material composition data from different formats provided by raw material suppliers; and normalize quality inspection data to ensure data consistency. For example, automotive part dimensional data measured by different inspection equipment is uniformly converted to international standard units.
[0036] Data fusion and feature extraction: Data fusion technology is used to integrate preprocessed multi-source data. For example, data from stamping shop pressure sensors, welding shop visual image data, and raw material material data can be integrated. Deep learning modules are used to extract features from the integrated data. For example, convolutional neural networks can be used to analyze welding image data to extract features such as weld shape and size. Data mining techniques can also be used to identify potential correlations between raw material material and production process data.
[0037] Quality analysis and problem diagnosis: based on the innovative module formula
[0038] When analyzing the quality of automobile body painting, after a large number of experiments and actual production experience, it is determined that the weight coefficient of production equipment temperature data (set as D1) is w1 = 0.3, and the weight coefficient of paint composition data (set as D2) is w2 = 0.4. Assuming that the current temperature data D1 = 25a^, the trained quantization function f1(D1) calculates the quantitative impact value of the paint quality stability to be 0.8; the paint composition data D2 is calculated by f2(D2) to be 0.6. The comprehensive evaluation value of the paint quality index is
[0039] (The calculation of data from other data sources is similar).
[0040] When the temperature is lower than the set threshold, the system diagnoses a coating quality problem. After analysis, it is caused by unstable temperature control. The company adjusts the temperature control system of the coating workshop accordingly.
[0041] (3) Implementation of quality optimization decision module based on reinforcement learning
[0042] Reinforcement Learning Environment Construction: Define a state space encompassing automotive component quality indicators (such as dimensional accuracy and hardness), production process parameters (such as stamping pressure and welding current), and equipment operating conditions (such as equipment temperature and vibration). The action space includes decisions such as adjusting production process parameters, scheduling equipment maintenance, and switching raw material suppliers. The reward function is set to provide positive rewards when automotive product quality improves and production costs decrease, such as a 10-point increase in reward value for every 1% increase in product qualification rate; negative rewards are given when quality declines or costs increase.
[0043] Module Training and Optimization: By simulating a large number of production scenarios, the reinforcement learning module continuously interacts with the environment. Initially, the module randomly selects actions, such as attempting to increase the welding current by 5A, observing changes in vehicle welding quality and production costs, and then updates its strategy based on reward feedback. After 1,000 iterations of training, the module gradually learns the optimal decision-making strategy, such as adjusting the stamping pressure to 12 MPa under specific equipment conditions to reduce production costs while ensuring quality.
[0044] Quality-optimized decision execution: In actual production, the system collects status information in real time and makes decisions based on trained strategies. When a deviation in the weld quality of a car seat is detected, the system automatically reduces the welding current by 3A and adjusts the welding time based on the strategy. Subsequent quality inspections confirm that the welding quality meets standards, increasing the product qualification rate by 8%.
[0045] (IV) Implementation of a quality control platform based on collaborative management
[0046] Business process integration and system integration: This integrates the business processes of procurement, production, quality inspection, and other departments within the company. The procurement department enters raw material procurement information, including supplier, purchase batch, and quality standards, into the platform; the production department uploads production progress and equipment operation data; and the quality inspection department enters quality inspection results. Through system integration, real-time information sharing is achieved across departments. For example, if the procurement department learns of quality issues with a batch of steel, it can promptly notify the production department to adjust production plans and avoid using that batch of steel.
[0047] Quality Task Assignment and Progress Tracking: The quality control platform assigns tasks based on the type and severity of quality issues. For example, if a batch of tires is found to have substandard dynamic balance, the platform assigns the issue to the quality inspection department for review, and the production department adjusts the tire assembly process. The platform tracks task progress in real time, allowing managers to view the submission date of the quality inspection department's review report and the completion status of the production department's process adjustments, ensuring timely resolution of issues.
[0048] Performance Evaluation and Incentive Mechanism: We have established a quality performance evaluation system to quantitatively assess the quality work of each department. For example, indicators such as the production department's product qualification rate and the quality inspection department's missed inspection rate will be included in the evaluation. Based on the evaluation results, outstanding departments will be awarded bonuses, honorary titles, and other rewards to motivate employees to enhance their quality awareness and work enthusiasm.
[0049] (V) Implementation of Visual Quality Traceability and Optimized Monitoring Platform
[0050] Interface Design: A visual monitoring platform developed using web technology offers a simple and intuitive interface. The quality traceability information display module presents data from all aspects of automotive production in a timeline format. The quality analysis results display module uses charts to display comprehensive assessments and trends of various quality indicators. The quality optimization decision-making suggestion display module provides optimization strategies for quality issues. The equipment operation status monitoring module displays parameters such as equipment temperature and pressure in real time.
[0051] Remote Operation and Control: Supports remote monitoring and management of production equipment and quality inspection equipment. Quality management personnel on business trips can log in to the platform via their mobile phone to view the operating status of production workshop equipment in real time. If they detect abnormal temperatures in coating equipment, they can remotely issue adjustment instructions to ensure normal production and improve the flexibility and efficiency of quality control.
[0052] 3. Summary of Implementation Effects
[0053] Through implementation at this automobile manufacturer, the Industrial Internet-based industrial product quality traceability and optimization system has achieved remarkable results. Quality traceability is precise and efficient, quality analysis is accurate and reliable, quality optimization decisions are scientific and rational, departmental collaboration is smooth, and visual monitoring is convenient. This has comprehensively improved automobile product quality, enhanced the company's market competitiveness, and provided a successful example of quality control in the automotive manufacturing industry and other industrial sectors.
Claims
1. An industrial product quality tracing and optimization system based on the Industrial Internet, characterized by: include: The full-process quality traceability architecture based on the Industrial Internet deploys IoT devices in all production links to collect quality-related data in real time, transmits it to the cloud-based quality traceability platform via the Industrial Internet, and uses blockchain technology to encrypt and store data, thus achieving quality traceability throughout the product life cycle. The quality analysis module based on multi-source heterogeneous data fusion adopts the following innovative module formula in the quality analysis process: Among them, QI is the comprehensive evaluation value of quality indicators, which is used to comprehensively measure the product quality status; Di represents the quality-related data collected by the i-th data source, which includes multi-source heterogeneous data such as production data, raw material data, and quality inspection data, and n is the total number of data sources; fi(Di) is the feature extraction and quantification function for the i-th data source data, which is trained based on deep learning and data mining techniques and can convert raw data into quantitative values that contribute to quality indicators. For example, through in-depth analysis of production temperature data, its quantitative impact on product quality stability is obtained. wi is the weight coefficient of the i-th data source data. Its value is determined based on the importance of different data sources in the product quality formation process. It is determined by fitting a large amount of experimental data and actual production experience, and is used to balance the contribution of data from different data sources in the comprehensive evaluation of quality indicators. This module integrates and processes multi-source heterogeneous quality data from different equipment and departments, and uses deep learning and data mining techniques to conduct in-depth analysis of the integrated data, explore potential patterns and laws in the data, establish a quality analysis model, and accurately identify the type, cause, and influencing factors of quality problems. The quality optimization decision module based on reinforcement learning abstracts the quality optimization process into a Markov decision process. By defining the state space, action space, and reward function, it learns the optimal quality optimization decision strategy to achieve continuous optimization of product quality. The quality control platform based on collaborative management integrates the quality-related business processes of various departments within the enterprise, realizes real-time information sharing and collaborative work, and has functions such as quality task allocation, progress tracking and performance evaluation; The visual quality traceability and optimization monitoring platform displays product quality traceability information, quality analysis results, quality optimization decision suggestions, etc. through a graphical interface, and supports remote operation and control.
2. The industrial product quality tracing and optimization system based on the Industrial Internet according to claim 1 is characterized in that: In the full-process quality traceability architecture based on the Industrial Internet, IoT devices include sensors, RFID tags, etc., which transmit data to the cloud platform through wireless communication technologies such as Industrial Ethernet and 5G.
3. The industrial product quality tracing and optimization system based on the Industrial Internet according to claim 1 is characterized in that: In the reinforcement learning-based quality optimization decision module, the state space includes information such as product quality indicators, production process parameters, and equipment operating status; the action space includes decisions such as production process parameter adjustments and equipment maintenance and replacement; and the reward function is set based on product quality improvement and production cost reduction.
4. The industrial product quality tracing and optimization system based on the Industrial Internet according to claim 1 is characterized in that: The quality control platform based on collaborative management integrates the quality-related business processes of the company's internal raw material procurement, production and processing, quality inspection and other departments to achieve real-time information sharing and collaborative work.
5. The industrial product quality tracing and optimization system based on the Industrial Internet according to claim 1 is characterized in that: The visual quality traceability and optimization monitoring platform is developed using Web technology, and its interface includes modules such as quality traceability information display, quality analysis result display, quality optimization decision suggestion display, and equipment operation status monitoring.
6. The industrial product quality tracing and optimization system based on the Industrial Internet according to claim 1 is characterized in that: In the full-process quality traceability architecture based on the Industrial Internet, the data blocks stored by blockchain technology contain timestamps, data content and the hash value of the previous data block, ensuring the authenticity and non-tamperability of the data.
7. The industrial product quality tracing and optimization system based on the Industrial Internet according to claim 1 is characterized in that: After data collection, the quality analysis module based on multi-source heterogeneous data fusion performs pre-processing operations such as data cleaning, format conversion, and normalization to eliminate data differences and noise.
8. The industrial product quality tracing and optimization system based on the Industrial Internet according to claim 1 is characterized in that: During the training process, the reinforcement learning-based quality optimization decision module selects actions based on the current state, and after executing the actions, updates the strategy based on the rewards and new states fed back by the environment to maximize the cumulative rewards.
9. The industrial product quality tracing and optimization system based on the Industrial Internet according to claim 1 is characterized in that: The quality control platform based on collaborative management establishes a quality performance evaluation system to evaluate and assess the quality work of each department and personnel, and implement corresponding incentive measures.
10. The industrial product quality tracing and optimization system based on the Industrial Internet according to claim 1 is characterized in that: The visual quality traceability and optimization monitoring platform supports remote monitoring and management of production equipment and quality inspection equipment, improving the flexibility and efficiency of quality control.