Equipment foresight maintenance matrix system and method based on MES and AI fusion

Abstract

The invention relates to the technical field of industrial equipment maintenance management, in particular to an equipment prospective maintenance matrix system and method based on MES and AI fusion, and the system comprises a data collection and integration module, a data preprocessing and feature extraction module, an AI prediction model construction and training module, and an equipment state evaluation and early warning module. The data acquisition and integration module is in deep butt joint with the MES system, acquires various parameter data of equipment operation through an IIOT platform in real time, and integrates the various parameter data and management data of the equipment to form a unified equipment data resource pool; and a data preprocessing and feature extraction module. According to the invention, the operation state of the equipment can be monitored in real time, and when the operation parameters of the equipment are abnormally changed, the system can immediately evaluate the fault probability and the maintenance requirement of the equipment again; and advanced prediction, accurate positioning and efficient maintenance decision making of equipment faults can be realized.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial equipment maintenance management, and in particular to an equipment forward-looking maintenance matrix system and method based on the integration of MES and AI. Background Art

[0002] In modern manufacturing, the stable operation of equipment is crucial to production efficiency and product quality. Traditional equipment maintenance strategies are mostly based on post-maintenance or regular preventive maintenance. Post-maintenance often leads to production interruptions, resulting in large economic losses; while regular preventive maintenance may cause waste of resources or equipment failure due to excessive maintenance or untimely maintenance. With the advancement of Industry 4.0, manufacturing execution systems (MES) have been widely used in production process management, and they have accumulated a large amount of equipment operation data. At the same time, artificial intelligence technology has demonstrated strong capabilities in data mining, predictive analysis, and other aspects. Therefore, the effective integration of MES and AI to build a forward-looking maintenance matrix system has become a key requirement for improving the level of equipment maintenance management. Summary of the invention

[0003] The purpose of the present invention is to provide an equipment proactive maintenance matrix system and method based on the fusion of MES and AI, which can monitor the operating status of the equipment in real time.

[0004] The technical solution of the present invention:

[0005] An equipment proactive maintenance matrix system based on the integration of MES and AI, including data collection and integration module, data preprocessing and feature extraction module, AI prediction model construction and training module, and equipment status assessment and early warning module;

[0006] Data collection and integration module: deeply connected with the MES system, collecting various parameter data of equipment operation through the IIOT platform in real time, and integrating various parameter data with equipment management data to form a unified equipment data resource pool;

[0007] Data preprocessing and feature extraction module: cleans the data in the equipment data resource pool, and then extracts and analyzes key feature data related to equipment failure;

[0008] AI prediction model construction and training module: builds an AI prediction model based on the extracted key feature data and trains the AI ​​prediction model;

[0009] Equipment status assessment and early warning module: Input various parameter data into the trained AI prediction model to obtain the equipment health status assessment results; according to the set early warning threshold, when the equipment failure probability output by the AI ​​prediction model exceeds the early warning threshold, the system automatically issues an early warning message.

[0010] Furthermore, in the data acquisition and integration module, various parameter data include temperature, pressure, vibration frequency, and rotation speed; equipment management data include historical maintenance records, process parameters, and production plans;

[0011] In the equipment status assessment and early warning module, the early warning information includes the possible fault type, expected fault time, and fault severity.

[0012] Furthermore, the data preprocessing and feature extraction module is further specifically to clean the data in the device data resource pool to remove noise data and abnormal values; firstly, the data in the device data resource pool is checked for integrity, and then missing, duplicated, and erroneous data are screened and eliminated;

[0013] Then, the MES system presets the equipment attribute dimensions and combines them with the cleaned labeled data to extract and analyze the familiar indicators of the dimensions that affect the equipment failure, and refines the features of the data dimensions with more labels. This feature is the key feature data.

[0014] Furthermore, the AI ​​prediction model construction and training module is further specified as follows: based on the extracted key feature data, an AI prediction model is constructed; and the equipment failure data recorded by the equipment management module in the MES system and the real-time data of the equipment are collected through the IIOT platform, and then the AI ​​prediction model is trained and the feature proportion is optimized in combination with the extracted key feature data, so that the AI ​​prediction model training predicts the equipment failure result data and the real-time operation data of the benchmark equipment. In this process, it also includes learning and adjusting process parameters and parameter proportions to optimize the AI ​​prediction model.

[0015] Furthermore, it also includes a maintenance strategy generation and optimization module and a maintenance effect evaluation and feedback module;

[0016] Maintenance strategy generation and optimization module: Combine the equipment's early warning information, current production task schedule, and maintenance resource status to generate the optimal equipment maintenance strategy; reasonably arrange maintenance time, determine the human and material resources and maintenance steps required for maintenance without affecting production progress; at the same time, continuously optimize the maintenance strategy based on the actual maintenance execution status and subsequent equipment operation data;

[0017] Maintenance effect evaluation and feedback module: After the equipment maintenance is completed, the maintenance effect is evaluated. By comparing the operating parameters and failure rate of the equipment before and after maintenance, it is determined whether the maintenance measures have achieved the expected goals; the evaluation results are fed back to the MES system for adjustment and improvement.

[0018] A proactive equipment maintenance matrix method based on the integration of MES and AI is as follows:

[0019] Step 1: Deeply connect with the MES system, collect various parameter data of equipment operation through the IIOT platform in real time, and integrate various parameter data with equipment management data to form a unified equipment data resource pool;

[0020] Step 2: Clean the data in the equipment data resource pool, and then extract and analyze the key feature data related to equipment failure;

[0021] Step 3: Based on the extracted key feature data, build an AI prediction model and train the AI ​​prediction model;

[0022] Step 4: Input various parameter data into the trained AI prediction model to obtain the health status assessment results of the equipment; according to the set warning threshold, when the equipment failure probability output by the AI ​​prediction model exceeds the warning threshold, the system automatically issues a warning message.

[0023] Furthermore, in step 1, the various parameter data include temperature, pressure, vibration frequency, and rotation speed; the equipment management data includes historical maintenance records, process parameters, and production plans;

[0024] In step 4, the warning information includes the possible fault type, expected fault time, and fault severity.

[0025] Furthermore, the step 2 is further specifically to clean the data in the device data resource pool to remove noise data and outliers; firstly, the data in the device data resource pool is checked for integrity, and then missing, duplicated, and erroneous data are screened and removed;

[0026] Then, the MES system presets the equipment attribute dimensions and combines them with the cleaned labeled data to extract and analyze the familiar indicators of the dimensions that affect the equipment failure, and refines the features of the data dimensions with more labels. This feature is the key feature data.

[0027] Furthermore, the step 3 is further specified as follows: building an AI prediction model based on the extracted key feature data; and collecting equipment real-time data through the equipment failure data recorded by the equipment management module in the MES system and the IIOT platform, and then combining the extracted key feature data to perform AI prediction model training and feature weight optimization, so that the AI ​​prediction model training predicts the equipment failure result data and matches the real-time operation data of the benchmark equipment. In this process, it also includes learning and adjusting process parameters and parameter weights to optimize the AI ​​prediction model.

[0028] Furthermore, it also includes step 5 and step 6;

[0029] Step 5: Generate the optimal equipment maintenance strategy based on the equipment’s early warning information, current production task schedule, and maintenance resource status; reasonably arrange maintenance time, determine the human and material resources required for maintenance, and maintenance steps without affecting production progress; at the same time, continuously optimize the maintenance strategy based on actual maintenance execution and subsequent equipment operation data;

[0030] Step 6: After the equipment maintenance is completed, the maintenance effect is evaluated. By comparing the operating parameters and failure rate of the equipment before and after maintenance, it is determined whether the maintenance measures have achieved the expected goals; the evaluation results are fed back to the MES system for adjustment and improvement.

[0031] Beneficial effects of the present invention:

[0032] The present invention can monitor the operating status of the equipment in real time. When the operating parameters of the equipment change abnormally, the system will immediately re-evaluate the failure probability and maintenance requirements of the equipment; it can achieve early prediction, accurate positioning and efficient maintenance decision-making of equipment failures. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a system principle block diagram of the present invention.

[0034] Figure 2 It is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0035] The present invention will be further described below in conjunction with the accompanying drawings.

[0036] like Figure 1 As shown, the present invention provides an equipment proactive maintenance matrix system based on the integration of MES and AI, including a data acquisition and integration module, a data preprocessing and feature extraction module, an AI prediction model construction and training module, and an equipment status evaluation and early warning module;

[0037] Data collection and integration module: deeply connected with the MES system, collecting various parameter data of equipment operation through the IIOT platform in real time, and integrating various parameter data with equipment management data to form a unified equipment data resource pool;

[0038] Data preprocessing and feature extraction module: cleans the data in the equipment data resource pool, and then extracts and analyzes key feature data related to equipment failure;

[0039] AI prediction model construction and training module: builds an AI prediction model based on the extracted key feature data and trains the AI ​​prediction model;

[0040] Equipment status assessment and early warning module: Input various parameter data into the trained AI prediction model to obtain the equipment health status assessment results; according to the set early warning threshold, when the equipment failure probability output by the AI ​​prediction model exceeds the early warning threshold, the system automatically issues an early warning message.

[0041] In the data acquisition and integration module, various parameter data include temperature, pressure, vibration frequency, and rotation speed; equipment management data includes historical maintenance records, process parameters, and production plans;

[0042] In the equipment status assessment and early warning module, the early warning information includes the possible fault type, expected fault time, and fault severity.

[0043] The data preprocessing and feature extraction module is further specifically to clean the data in the device data resource pool to remove noise data and abnormal values; firstly, the data in the device data resource pool is checked for integrity, and then missing, repeated, and erroneous data are screened and eliminated;

[0044] Then, the MES system presets the equipment attribute dimensions and combines them with the cleaned labeled data to extract and analyze the familiar indicators of the dimensions that affect the equipment failure, and refines the features of the data dimensions with more labels. This feature is the key feature data.

[0045] The AI ​​prediction model construction and training module is further specified as follows: based on the extracted key feature data, an AI prediction model is constructed; and through the equipment failure data recorded by the equipment management module in the MES system and the real-time data of the equipment collected by the IIOT platform, the AI ​​prediction model is trained and the feature weight is optimized in combination with the extracted key feature data, so that the AI ​​prediction model training predicts the equipment failure result data and the real-time operation data of the equipment are matched. In this process, it also includes learning and adjusting process parameters and parameter weights to optimize the AI ​​prediction model.

[0046] It also includes a maintenance strategy generation and optimization module and a maintenance effect evaluation and feedback module;

[0047] Maintenance strategy generation and optimization module: Combine the equipment's early warning information, current production task schedule, and maintenance resource status to generate the optimal equipment maintenance strategy; reasonably arrange maintenance time, determine the human and material resources and maintenance steps required for maintenance without affecting production progress; at the same time, continuously optimize the maintenance strategy based on the actual maintenance execution status and subsequent equipment operation data;

[0048] Maintenance effect evaluation and feedback module: After the equipment maintenance is completed, the maintenance effect is evaluated. By comparing the operating parameters and failure rate of the equipment before and after maintenance, it is determined whether the maintenance measures have achieved the expected goals; the evaluation results are fed back to the MES system for adjustment and improvement.

[0049] The present invention is further described below in conjunction with a specific embodiment:

[0050] A proactive equipment maintenance matrix system based on the integration of MES and AI.

[0051] Data collection and integration module:

[0052] Deeply connect with the MES system, and collect various parameter data of equipment operation in real time through the IIOT platform in the attached figure, such as temperature, pressure, vibration frequency, speed, etc. At the same time, the historical maintenance records, process parameters, production plans and other related data of the equipment are formed into a unified equipment data resource pool to provide a comprehensive data basis for subsequent analysis.

[0053] Data preprocessing and feature extraction module:

[0054] Clean the massive amount of data collected to remove noise data, outliers and other interfering information. The main steps of data cleaning are data integrity check, and then filter out missing, duplicated and erroneous data. Then, use data mining algorithms to extract and analyze key features related to equipment failures. The extraction and analysis method is to extract the dimensions that affect equipment failures through the preset equipment attribute dimensions of the system itself combined with the collected data analysis to understand the indicators, and refine the features of the data dimensions with more markers.

[0055] AI prediction model building and training module:

[0056] Based on the extracted feature data, a variety of artificial intelligence prediction models are constructed, such as neural network models, support vector machine models, etc. These models are trained and optimized using historical equipment failure data and normal operation data, and the optimal parameter configuration of the model is determined through cross-validation and other technologies, so that the model can accurately learn the inherent relationship between equipment failure and feature data, thereby having the ability to predict the future status of the equipment. The specific steps are as follows:

[0057] The equipment failure data recorded by the equipment management module in the MES system and the real-time equipment data collected by the IIOT platform are combined with the extracted equipment feature data for model training and feature weight optimization. This allows the model training to predict the equipment failure result data and benchmark the real-time equipment operation data. This process will also learn and adjust process parameters and parameter weights to optimize the model.

[0058] Equipment status assessment and early warning module:

[0059] Input all kinds of new parameter data into the trained AI prediction model in real time to obtain the health status assessment results of the equipment. According to the set warning threshold, when the equipment failure probability output by the prediction model exceeds the threshold, the system automatically issues a warning message, which includes the possible type of failure, expected failure time, and failure severity, so that maintenance personnel can prepare in advance.

[0060] Maintenance strategy generation and optimization module:

[0061] Combined with the equipment's early warning information, current production task schedule, and maintenance resource status, intelligent algorithms are used to generate the optimal equipment maintenance strategy. For example, without affecting the production schedule, the maintenance time is reasonably arranged, and the human and material resources and maintenance steps required for maintenance are determined. At the same time, based on the actual maintenance execution status and subsequent equipment operation data, the maintenance strategy is continuously optimized to improve the effectiveness and adaptability of the maintenance strategy.

[0062] Maintenance effect evaluation and feedback module:

[0063] After the equipment maintenance is completed, the maintenance effect is evaluated. By comparing the operating parameters, failure rate and other indicators of the equipment before and after maintenance, it is judged whether the maintenance measures have achieved the expected goals. The evaluation results are fed back to the system to adjust and improve the data collection, model training and other links, and continuously improve the performance of the entire equipment forward-looking maintenance matrix system.

[0064] To sum up, the overall process is to collect data through MES + IIOT gateway to collect device data and synchronize it to the intelligent AI data processing engine, perform data extraction and model training according to data modeling and feature engineering standards, and then perform model reasoning and learning. Finally, the results are output to the data center, and the dashboard reports and corresponding equipment predictive maintenance analysis and adjustment of maintenance strategies are synchronized through the data center.

[0065] In another embodiment of the present invention,

[0066] 1. Data processing and management

[0067] 1. Multi-source data fusion and intelligent cleaning technology

[0068] By combining attribute data of different dimensions in the equipment with weighted average, Bayesian estimation method, and labeling algorithm for data fusion calculation and analysis, it can seamlessly connect the MES system, equipment sensor data, historical maintenance records, and external environmental data (such as workshop temperature and humidity, etc.), and use intelligent cleaning technology in the fusion process to identify and correct outliers, duplicate data, and missing data.

[0069] Advantages: By integrating multi-source data, it provides a more comprehensive and accurate information basis for equipment status assessment. Intelligent cleaning technology ensures data quality, reduces incorrect predictions and decisions caused by data errors, and improves the reliability and accuracy of the entire maintenance system. For example, fusion of environmental data can help analyze the performance changes of equipment under different environmental conditions and discover potential failure risks caused by environmental factors in advance.

[0070] 2. Dynamic data update and adaptive mechanism

[0071] A data dynamic update module is designed, which can monitor the changing trends and distribution characteristics of data in real time. When a significant change in the data pattern is detected (such as changes in data characteristics caused by equipment upgrades, process adjustments, etc.), the module automatically triggers the adaptive algorithm to adjust the data collection strategy and subsequent data processing procedures, including the data screening and feature extraction rules required to retrain the AI ​​model.

[0072] Advantages: The system can quickly adapt to various changes in the enterprise's production process and always maintain effective monitoring and prediction capabilities for equipment status. It avoids the failure of maintenance strategies due to the disconnection between data and actual equipment operation, ensures the long-term effectiveness and practicality of the system, and reduces the cost and time required to redevelop the system due to equipment changes or process adjustments.

[0073] 2. AI model construction and application

[0074] 1. Hybrid AI model architecture

[0075] A hybrid AI model architecture is used to combine deep learning models (such as convolutional neural networks for processing sensor image data or time series data) with traditional machine learning models (such as decision trees for generating interpretable rules). During the model training process, the powerful feature learning capabilities of the deep learning model are used to extract complex features from the equipment operation data, and then these features are input into the traditional machine learning model for fault classification and prediction.

[0076] Advantages: It not only takes advantage of the deep learning model in processing large-scale, high-dimensional data, but also uses the interpretability of traditional machine learning models to solve the problem of the difficulty in balancing the interpretability and prediction accuracy of a single AI model. Maintenance personnel can understand the prediction basis of the model based on the rules generated by the decision tree, and use the high-precision prediction results of the deep learning model to improve the scientificity and credibility of maintenance decisions, and promote the smooth promotion and application of the system in the actual production of enterprises.

[0077] 2. Model online learning and optimization algorithm

[0078] A model online learning and optimization algorithm is proposed, which allows the AI ​​model to continuously learn new data during the operation of the device. The algorithm combines small batch gradient descent with dynamic weight adjustment to dynamically adjust the model weight according to the importance of new data and the degree of difference from historical data. At the same time, the model is compressed and pruned regularly to prevent overfitting and increased complexity of the model.

[0079] Advantages: Enables AI models to track changes in the operating status of equipment in a timely manner and continuously improves prediction performance. Compared with traditional offline training models, online learning algorithms can adapt to equipment performance drift and the emergence of new failure modes more quickly, reducing prediction errors caused by model aging. Model compression and pruning technology reduces the model's computing resource requirements and storage costs while ensuring model accuracy, and improves the system's operating efficiency and response speed.

[0080] 3. System Function and Integration

[0081] 1. Visual interaction and intelligent decision-making support platform

[0082] A visual interaction and intelligent decision-making support platform has been built, which uses an intuitive graphical interface to display the real-time status of the equipment, historical operating data, prediction results, maintenance recommendations and other information. At the same time, the platform integrates intelligent decision-making support functions, which can automatically generate a variety of maintenance plans based on the equipment status and production plan, and provide decision support for maintenance personnel by simulating and analyzing the impact of different plans on production.

[0083] Advantages: It greatly improves the efficiency and friendliness of the interaction between maintenance personnel and the system. The visual interface enables maintenance personnel to quickly and comprehensively understand the equipment status, facilitating timely problem discovery and decision-making. The intelligent decision-making assistance function reduces the subjectivity and uncertainty of human decision-making, improves the scientificity and rationality of maintenance strategies, helps optimize production arrangements, and reduces maintenance costs and production losses.

[0084] 2. Cross-platform integration and distributed deployment architecture

[0085] A cross-platform integration and distributed deployment architecture was designed to achieve seamless integration of the forward-looking maintenance matrix system with various operating systems, database platforms, and network architectures within the enterprise. Using a distributed deployment approach, functional modules such as data collection, processing, model training, and application services are distributed on different computing nodes, and collaborative work is achieved through efficient communication protocols and task scheduling mechanisms.

[0086] Advantages: Improves system compatibility and scalability, can adapt to the complex IT environment of different enterprises, and facilitates system deployment and promotion. Distributed deployment architecture makes full use of the internal computing resources of the enterprise, improves the system's processing power and response speed, can meet the needs of large-scale equipment data processing and real-time maintenance decision-making, reduces dependence on a single high-performance server, and improves system reliability and stability.

[0087] In another embodiment of the present invention,

[0088] Expansion plan:

[0089] 1. Combined with IoT edge computing

[0090] Solution description: Deploy edge computing devices on the device side to complete some data preprocessing and simple AI analysis on the edge side. For example, perform preliminary spectrum analysis on the real-time vibration data of the equipment to determine whether there are abnormal frequency components. If so, upload the relevant data and preliminary analysis results to the MES and central AI system for further in-depth analysis. This can reduce the amount of data transmission and the impact of network latency on maintenance decisions with high real-time requirements. It can also reduce the computing burden of the central system and improve the response speed of the overall system.

[0091] Advantages: Enhances the real-time and robustness of the system, especially for scenarios with high requirements for network stability or where devices are distributed in remote areas. Edge computing devices can continue to perform some basic maintenance monitoring and early warning work locally in the event of network interruption, and synchronize data for more comprehensive analysis after the network is restored.

[0092] 2. Introducing digital twin technology

[0093] Solution description: A digital twin model is built for each important equipment, which is updated and simulated in real time based on the equipment's design parameters, historical operation data, and real-time monitoring data. Through the digital twin model, various maintenance strategies for the equipment can be rehearsed in a virtual environment, such as simulating the operation effect of the equipment after replacing a certain component, or predicting the overall impact of different maintenance time points on the production process. At the same time, the digital twin model can also help train maintenance personnel, allowing them to become familiar with the internal structure of the equipment and the maintenance operation process in a virtual environment.

[0094] Advantages: Improve the visualization and accuracy of maintenance decisions. The optimal maintenance plan can be selected through rehearsal to avoid production losses caused by improper maintenance operations or inappropriate maintenance timing. The digital twin model can also be used as a carrier of knowledge inheritance. New employees can quickly master equipment maintenance knowledge and skills through interaction with the digital twin model.

[0095] 3. Integration with supply chain management systems

[0096] Solution description: Connect the equipment forward-looking maintenance matrix system with the company's supply chain management system. When the system predicts that the equipment needs maintenance and parts need to be replaced, it automatically sends parts demand information to the supply chain management system, and the supply chain management system arranges the purchase and distribution of parts based on inventory conditions, supplier information, etc. At the same time, it can also optimize inventory management strategies based on equipment maintenance plans and parts supply cycles to avoid parts backlogs or shortages.

[0097] Advantages: It realizes seamless connection between equipment maintenance and supply chain, and improves the overall operational efficiency of the enterprise. It ensures that the required parts can be obtained in time when the equipment needs maintenance, reduces equipment downtime caused by parts supply problems, and optimizes inventory management to reduce the inventory costs of the enterprise.

[0098] Alternative solution:

[0099] 1. Use federated learning to replace traditional centralized AI training

[0100] Solution description: In enterprises with multiple factories or branches and high data privacy requirements, a federated learning algorithm is used. The equipment data of each factory or branch is trained locally for AI models, and only the parameters or intermediate results of the model (such as gradient information) are encrypted and transmitted and aggregated and updated on the central server. The updated model parameters are then transmitted back to each local node. In this way, AI model training based on multi-source data can be achieved without leaking the original data, thereby improving the generalization ability and prediction accuracy of the model.

[0101] Advantages: It effectively protects the data privacy of the enterprise and avoids the risk of data leakage caused by centralized storage and transmission of data. At the same time, it can make full use of the equipment data resources scattered in various regions of the enterprise and improve the performance of the equipment maintenance system within the entire enterprise.

[0102] 2. Optimizing maintenance strategy generation with reinforcement learning

[0103] Solution description: Apply reinforcement learning algorithm to the maintenance strategy generation module. Take the operating state of the equipment as the environmental state and the maintenance operation as the action. By designing a suitable reward function (such as comprehensive consideration of indicators such as increased equipment uptime and reduced maintenance costs), the reinforcement learning agent can learn the optimal maintenance strategy in the process of continuous trial and error. For example, the agent can decide whether to perform maintenance immediately, delay maintenance, or take some temporary adjustment measures (such as reducing the operating power of the equipment) based on the current health status of the equipment and the production task schedule.

[0104] Advantages: It can automatically adapt to the complex and changing equipment operating environment and production needs, and generate more flexible and optimized maintenance strategies. Compared with the traditional rule-based or fixed model-based maintenance strategy generation method, reinforcement learning can better handle uncertainties and dynamic changes, and improve the intelligence level of enterprise equipment maintenance management.

[0105] like Figure 2 As shown, the present invention also provides a device proactive maintenance matrix method based on the integration of MES and AI, the method is as follows:

[0106] Step 1: Deeply connect with the MES system, collect various parameter data of equipment operation through the IIOT platform in real time, and integrate various parameter data with equipment management data to form a unified equipment data resource pool;

[0107] Step 2: Clean the data in the equipment data resource pool, and then extract and analyze the key feature data related to equipment failure;

[0108] Step 3: Based on the extracted key feature data, build an AI prediction model and train the AI ​​prediction model;

[0109] Step 4: Input various parameter data into the trained AI prediction model to obtain the health status assessment results of the equipment; according to the set warning threshold, when the equipment failure probability output by the AI ​​prediction model exceeds the warning threshold, the system automatically issues a warning message.

[0110] In step 1, various parameter data include temperature, pressure, vibration frequency, and rotation speed; equipment management data include historical maintenance records, process parameters, and production plans;

[0111] In step 4, the warning information includes the possible fault type, expected fault time, and fault severity.

[0112] The step 2 is further specifically to clean the data in the device data resource pool to remove noise data and outliers; firstly, perform an integrity check on the data in the device data resource pool, and then filter out missing, duplicated, and erroneous data;

[0113] Then, the MES system presets the equipment attribute dimensions and combines them with the cleaned labeled data to extract and analyze the familiar indicators of the dimensions that affect the equipment failure, and refines the features of the data dimensions with more labels. This feature is the key feature data.

[0114] The step 3 is further specified as follows: based on the extracted key feature data, an AI prediction model is constructed; and the equipment failure data recorded by the equipment management module in the MES system and the real-time data of the equipment are collected through the IIOT platform, and then the AI ​​prediction model is trained and the feature weight is optimized in combination with the extracted key feature data, so that the AI ​​prediction model training predicts the equipment failure result data and the real-time operation data of the equipment are matched. In this process, it also includes learning and adjusting process parameters and parameter weights to optimize the AI ​​prediction model.

[0115] Also includes step 5 and step 6;

[0116] Step 5: Generate the optimal equipment maintenance strategy based on the equipment’s early warning information, current production task schedule, and maintenance resource status; reasonably arrange maintenance time, determine the human and material resources required for maintenance, and maintenance steps without affecting production progress; at the same time, continuously optimize the maintenance strategy based on actual maintenance execution and subsequent equipment operation data;

[0117] Step 6: After the equipment maintenance is completed, the maintenance effect is evaluated. By comparing the operating parameters and failure rate of the equipment before and after maintenance, it is determined whether the maintenance measures have achieved the expected goals; the evaluation results are fed back to the MES system for adjustment and improvement.

[0118] The above description is only a preferred embodiment of the present invention and should not be construed as limiting the present application. All equivalent changes and modifications made within the scope of the patent application of the present invention should fall within the scope of the present invention.

Claims

1. A proactive equipment maintenance matrix system based on the integration of MES and AI, characterized in that: It includes data collection and integration module, data preprocessing and feature extraction module, AI prediction model construction and training module, and equipment status assessment and early warning module; Data collection and integration module: deeply connected with the MES system, collecting various parameter data of equipment operation through the IIOT platform in real time, and integrating various parameter data with equipment management data to form a unified equipment data resource pool; Data preprocessing and feature extraction module: cleans the data in the equipment data resource pool, and then extracts and analyzes key feature data related to equipment failure; AI prediction model construction and training module: builds an AI prediction model based on the extracted key feature data and trains the AI ​​prediction model; Equipment status assessment and early warning module: Input various parameter data into the trained AI prediction model to obtain the equipment health status assessment results; according to the set early warning threshold, when the equipment failure probability output by the AI ​​prediction model exceeds the early warning threshold, the system automatically issues an early warning message.

2. According to claim 1, the equipment proactive maintenance matrix system based on the integration of MES and AI is characterized in that: In the data acquisition and integration module, various parameter data include temperature, pressure, vibration frequency, and rotation speed; equipment management data includes historical maintenance records, process parameters, and production plans; In the equipment status assessment and early warning module, the early warning information includes the possible fault type, expected fault time, and fault severity.

3. According to claim 1, the equipment proactive maintenance matrix system based on the integration of MES and AI is characterized in that: The data preprocessing and feature extraction module is further specifically to clean the data in the device data resource pool to remove noise data and abnormal values; firstly, the data in the device data resource pool is checked for integrity, and then missing, repeated, and erroneous data are screened and eliminated; Then, the MES system presets the equipment attribute dimensions and combines them with the cleaned labeled data to extract and analyze the familiar indicators of the dimensions that affect the equipment failure, and refines the features of the data dimensions with more labels. This feature is the key feature data.

4. According to claim 1, the equipment proactive maintenance matrix system based on the integration of MES and AI is characterized in that: The AI ​​prediction model construction and training module is further specified as follows: based on the extracted key feature data, an AI prediction model is constructed; and through the equipment failure data recorded by the equipment management module in the MES system and the real-time data of the equipment collected by the IIOT platform, the AI ​​prediction model is trained and the feature weight is optimized in combination with the extracted key feature data, so that the AI ​​prediction model training predicts the equipment failure result data and the real-time operation data of the equipment are matched. In this process, it also includes learning and adjusting process parameters and parameter weights to optimize the AI ​​prediction model.

5. According to claim 1, the equipment proactive maintenance matrix system based on the integration of MES and AI is characterized in that: It also includes a maintenance strategy generation and optimization module and a maintenance effect evaluation and feedback module; Maintenance strategy generation and optimization module: Combine the equipment's early warning information, current production task schedule, and maintenance resource status to generate the optimal equipment maintenance strategy; reasonably arrange maintenance time, determine the human and material resources and maintenance steps required for maintenance without affecting production progress; at the same time, continuously optimize the maintenance strategy based on the actual maintenance execution status and subsequent equipment operation data; Maintenance effect evaluation and feedback module: After the equipment maintenance is completed, the maintenance effect is evaluated. By comparing the operating parameters and failure rate of the equipment before and after maintenance, it is determined whether the maintenance measures have achieved the expected goals; the evaluation results are fed back to the MES system for adjustment and improvement.

6. A proactive equipment maintenance matrix method based on the integration of MES and AI, characterized in that: Here’s how: Step 1: Deeply connect with the MES system, collect various parameter data of equipment operation through the IIOT platform in real time, and integrate various parameter data with equipment management data to form a unified equipment data resource pool; Step 2: Clean the data in the equipment data resource pool, and then extract and analyze the key feature data related to equipment failure; Step 3: Based on the extracted key feature data, build an AI prediction model and train the AI ​​prediction model; Step 4: Input various parameter data into the trained AI prediction model to obtain the health status assessment results of the equipment; according to the set warning threshold, when the equipment failure probability output by the AI ​​prediction model exceeds the warning threshold, the system automatically issues a warning message.

7. According to claim 6, a proactive equipment maintenance matrix method based on the integration of MES and AI is characterized in that: In step 1, various parameter data include temperature, pressure, vibration frequency, and rotation speed; equipment management data include historical maintenance records, process parameters, and production plans; In step 4, the warning information includes the possible fault type, expected fault time, and fault severity.

8. According to claim 6, a proactive equipment maintenance matrix method based on the integration of MES and AI is characterized in that: The step 2 is further specifically to clean the data in the device data resource pool to remove noise data and outliers; firstly, perform an integrity check on the data in the device data resource pool, and then filter out missing, duplicated, and erroneous data; Then, the MES system presets the equipment attribute dimensions and combines them with the cleaned labeled data to extract and analyze the familiar indicators of the dimensions that affect the equipment failure, and refines the features of the data dimensions with more labels. This feature is the key feature data.

9. According to claim 6, a proactive equipment maintenance matrix method based on the integration of MES and AI is characterized in that: The step 3 is further specified as follows: based on the extracted key feature data, an AI prediction model is constructed; and the equipment failure data recorded by the equipment management module in the MES system and the real-time data of the equipment are collected through the IIOT platform, and then the AI ​​prediction model is trained and the feature weight is optimized in combination with the extracted key feature data, so that the AI ​​prediction model training predicts the equipment failure result data and the real-time operation data of the equipment are matched. In this process, it also includes learning and adjusting process parameters and parameter weights to optimize the AI ​​prediction model.

10. The equipment proactive maintenance matrix method based on the integration of MES and AI according to claim 6 is characterized in that: Also includes step 5 and step 6; Step 5: Generate the optimal equipment maintenance strategy based on the equipment’s early warning information, current production task schedule, and maintenance resource status; reasonably arrange maintenance time, determine the human and material resources required for maintenance, and maintenance steps without affecting production progress; at the same time, continuously optimize the maintenance strategy based on actual maintenance execution and subsequent equipment operation data; Step 6: After the equipment maintenance is completed, the maintenance effect is evaluated. By comparing the operating parameters and failure rate of the equipment before and after maintenance, it is determined whether the maintenance measures have achieved the expected goals; the evaluation results are fed back to the MES system for adjustment and improvement.

Images