AI-based automatic production line scheduling system in industrial internet

Abstract

The invention discloses an AI-based automatic production line scheduling system in an industrial internet, which relates to the technical field of production scheduling and comprises a production plan management module S1, a dynamic scheduling engine module S2, a resource scheduling module S3, a real-time monitoring system module S4, an exception handling center module S5 and a data optimization platform module S6. In the industrial internet, an AI-based automatic production line scheduling system, an X dynamic scheduling engine millisecond response and a multi-agent reinforcement learning engine based on a federated learning architecture realize millisecond response scheduling, each device is used as an autonomous decision-making unit, and dynamic coordination is performed through a distributed Q learning algorithm, so that the vacancy rate of the devices is greatly reduced, and the scheduling efficiency is improved. According to a long-short-term memory network deep analysis model of order delivery cycle compression, emergency order insertion response speed improvement, multi-modal AI quality monitoring, fusion of vibration, thermal imaging and current spectrum, the detection rate is greatly improved compared with a unified sensor, causal reasoning and root cause analysis are performed, a fault causal graph is constructed to position a deep problem, and the average repair time is shortened.

Description

Technical Field

[0001] The present invention relates to the technical field of production scheduling, and specifically to an AI-based automated production line scheduling system in the Industrial Internet. Background Art

[0002] Automated production line scheduling is a core part of modern intelligent manufacturing. It aims to optimize production processes through intelligent means, improve equipment utilization and delivery efficiency, and achieve cost reduction and efficiency improvement by dynamically allocating production tasks, coordinating equipment resources, and shortening production cycles.

[0003] Prior art 1 (Chinese patent with announcement number CN113780747B and announcement date 2024-03-22) A production line task automatic division, grasping and scheduling system and method, including a production line system that can operate independently and a workstation system that can operate independently, the production line system and the workstation system are communicatively connected; and including a process model construction module set in the production line system, for constructing the process required for product production; a production task grasping module set in the workstation system, for obtaining the process information required for product production; a production task division module set in the production line system, for dividing the production task into multiple subtasks according to the process; a scheduling task grasping module set in the workstation system, for grasping subtasks and executing them until all production tasks are completed, etc.; the production line task division and scheduling process is optimized, the production line task process efficiency is improved, flexible production, lean production, and costs are reduced, etc. Prior art 2 (Chinese patent with announcement number CN111898908B and announcement date 2023-06-16) A production line scheduling system and method based on multi-agent. The production scheduling system includes a global management module, a controller agent, and a sub-controller agent. The global management module is the highest-level manager in the production scheduling system and has the highest management authority. The controller agent is the core of the production scheduling system and includes a resource module, a planning and calculation module, and a task management module. The resource module is used to store resource information of the sub-controller agent. The planning and calculation module decomposes tasks and sends them to the task management module. The task management module sorts the decomposed tasks according to priority and then distributes them to the sub-controller agent. The sub-controller agent is used to execute tasks sent by the task management module, ensuring the agent's independent operation capability and improving the work efficiency and management capabilities of the scheduling system.

[0004] Although the scheduling system in the existing technology improves the work efficiency and management capabilities of the scheduling system, it is unable to respond to equipment failures / insertion orders, requires manual rescheduling, relies too much on predictions, schedules production based on historical averages, ignores real-time fluctuations, lacks online optimization capabilities, has low resource coordination efficiency, delays in responding to equipment failures, is seriously lacking in real-time performance, and has weak predictive maintenance capabilities.

[0005] Therefore, we proposed an AI-based automated production line scheduling system in the Industrial Internet to solve the problems raised above. Summary of the Invention

[0006] The purpose of the present invention is to provide an AI-based automated production line scheduling system in the industrial Internet to solve the problems raised in the above-mentioned background technology that the current system on the market cannot respond to equipment failures / insertion orders, requires manual rescheduling, relies too much on predictions, schedules production based on historical averages, ignores real-time fluctuations, lacks online optimization capabilities, has low resource coordination efficiency, delayed response to equipment failures, seriously lacks real-time performance, and has weak predictive maintenance capabilities.

[0007] To achieve the above-mentioned objectives, the present invention provides the following technical solutions: an AI-based automated production line scheduling system in the industrial Internet, comprising S1: a production plan management module, S2: a dynamic scheduling engine module, S3: a resource scheduling module, S4: a real-time monitoring system module, S5: an exception handling center module, and S6: a data optimization platform module. The production plan management module converts high-level production plans into executable work orders to ensure that orders are delivered on time. The dynamic scheduling engine module responds to production disturbances (equipment failures / insertion orders) in real time and dynamically optimizes resource allocation. The resource scheduling module accurately matches equipment, materials, and manpower requirements to eliminate production bottlenecks. The real-time monitoring system module perceives the production line status in seconds to provide a data basis for decision-making. The exception handling center module quickly locates the root cause of the fault to minimize production stoppage losses. The data optimization platform module drives continuous improvement of the system to achieve a "perception-decision-optimization" closed loop.

[0008] Preferably, the production planning management module breaks down the order based on the conversion process from the bill of materials to the process route, and outputs a directed acyclic graph. It prioritizes the work orders through the earliest delivery priority algorithm, outputs the corresponding work order queue, achieves production capacity balance through the linear programming constraint solver, and generates the corresponding equipment load heat map. At the same time, it checks the completeness of materials through the real-time docking warehouse operation management system to realize material shortage warning report.

[0009] Preferably, the dynamic scheduling engine module performs multi-objective optimization scheduling through genetic algorithms to minimize completion time and energy consumption. When equipment fails, rescheduling can be initiated. The real-time rescheduling response speed is fast, and the automatic guided vehicle path cooperates to reduce material handling time.

[0010] Preferably, the equipment resources in the resource scheduling module are preemptively scheduled based on the equipment status, and technical support is provided by cross-platform communication real-time data collection technology. Material resources are bound to on-time delivery and wireless radio frequency identification with the support of automatic guided vehicle scheduling system integration technology. Human resources match the skill matrix through real-time visual management system docking technology and use mobile terminals to push tasks, realizing full-link tracking of materials → equipment → personnel.

[0011] Preferably, the monitoring dimensions in the real-time monitoring system module include three aspects: equipment status monitoring, quality process control and energy consumption management. Among them, equipment status monitoring performs real-time detection of the comprehensive efficiency of the equipment through the cooperation of vibration / temperature sensors and edge computing. Quality process control realizes the control of process capability index through machine design and statistical process control. At the same time, the energy consumption curve of unit product is monitored through data feedback from smart meters and energy analysis models to realize energy consumption management.

[0012] Preferably, the exception handling center module divides the fault level into three levels. The first level is a transient fault. The transient fault PLC program can be automatically triggered to automatically recover the equipment. The second level is a device-level fault. When a device-level fault occurs, the failure mode analysis library is required to analyze the cause of the failure, so as to notify the maintenance personnel to switch the corresponding equipment. The third level is a system-level fault. The system fault defect is verified through digital twin simulation, and the faulty equipment is safely shut down, and the emergency production plan is activated to ensure that production remains normal.

[0013] Preferably, the data optimization platform module has a digital twin simulation function, which pre-verifies the scheduling strategy through 3D production line modeling and physical engine. The data optimization platform module also has a predictive maintenance function, which can provide an early warning of bearing failures 3 days in advance through a long short-term memory network deep analysis model. At the same time, the data optimization platform module also has a KPI deep analysis function, which can discover "voltage fluctuation → yield decline" through association rule mining.

[0014] Preferably, the innovation of the production planning management module is the intelligent order splitting based on the knowledge graph, the innovation of the dynamic scheduling engine module is the distributed scheduling driven by federated learning, the innovation of the resource scheduling module is the virtual-real resource mapping enabled by digital twins, the innovation of the real-time monitoring system module is multimodal AI fusion monitoring, the innovation of the exception handling center module is the root cause analysis driven by causal reasoning, and the innovation of the data optimization platform module is the physical information neural network.

[0015] Preferably, the production planning management module can avoid the problem of relying on manual experience to disassemble the bill of materials, which is error-prone and time-consuming, low manual scheduling efficiency, and inability to cope with emergency orders. The dynamic scheduling engine module can avoid the problem of high centralized scheduling response delay, inability to adapt to multi-workshop collaboration, and high equipment idle rate due to fixed scheduling. The resource scheduling module can avoid the problem of physical resources and system data being out of sync, high scheduling distortion rate, and material waiting occupying a long production cycle. The real-time monitoring system module can avoid the problem of high false alarm rate of a single sensor and delayed detection of quality anomalies leading to batch scrapping. The exception handling center module can avoid the problem of relying on expert rule base, long time-consuming new fault diagnosis, and long average fault repair time. The data optimization platform module can avoid the problem of pure data-driven models violating physical laws and relying on experience to cause optimization lag.

[0016] Compared with the prior art, the present invention has the following beneficial effects:

[0017] (1) Dynamic scheduling engine with millisecond-level response. The multi-agent reinforcement learning engine based on the federated learning architecture achieves millisecond-level response scheduling. Each device acts as an autonomous decision-making unit and is dynamically coordinated through the distributed Q learning algorithm. The idle rate of the equipment is greatly reduced, the order delivery cycle is compressed, and the response speed of emergency orders is improved.

[0018] (2) Multimodal AI quality monitoring, integrating vibration, thermal imaging, and long-short-term memory network deep analysis models of current spectrum, greatly improves the detection rate compared to unified sensors, conducts causal reasoning root cause analysis, constructs fault cause-effect diagrams to locate deep-seated problems, and reduces the average repair time.

[0019] (3) When equipment fails, the system takes a shorter time to recover. GPU acceleration is supported, and scheduling calculations can be accelerated by 50 times. Through constraint theory, throughput is increased by 22%, and full-link tracking from materials → equipment → personnel is achieved. SPC control charts enable early detection of defects and reduce scrap rates. Data preprocessing reduces bandwidth requirements, the automatic rate of first-level faults is improved, and fault tree analysis is formed into an enterprise knowledge base. The strategy iteration cycle is reduced from weeks to hours, and predictive maintenance reduces unexpected downtime losses.

[0020] (4) The order splitting speed is improved, and the system dynamically adapts to process changes. It supports the coordinated scheduling of 100+ production line units, breaks the "scheduling island", achieves cross-factory production capacity balance, improves the accuracy of resource scheduling, supports "quantum-classical" hybrid computing resource allocation, improves the fault detection rate, reduces 80% of sensor deployment, improves the model generalization capability by 5 times, and reduces the simulation verification requirements by 50%. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 Schematic diagram of the overall framework of the system of the present invention;

[0022] Figure 2 This is a schematic diagram of causal reasoning of the exception handling module of the present invention;

[0023] Figure 3 This is a schematic diagram of the innovative features of the resource scheduling module of the present invention;

[0024] Figure 4 This is a schematic diagram of causal reasoning of the exception handling center module of the present invention. DETAILED DESCRIPTION

[0025] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0026] Example 1: Figure 1-Figure 4 The technical solution shown in the present invention provides the following technical solution: an AI-based automated production line scheduling system in the industrial Internet, disclosed: including S1: production plan management module, S2: dynamic scheduling engine module, S3: resource scheduling module, S4: real-time monitoring system module, S5: exception handling center module, S6: data optimization platform module, characterized in that the production plan management module converts high-level production plans into executable work orders to ensure that orders are delivered on time, the dynamic scheduling engine module responds to production disturbances (equipment failure / insertion orders) in real time, dynamically optimizes resource allocation, and accurately matches equipment, materials, and manpower requirements through the resource scheduling module to eliminate production bottlenecks. The real-time monitoring system module perceives the production line status in seconds, providing a data basis for decision-making, and quickly locates the root cause of the fault through the exception handling center module to minimize production stoppage losses. The data optimization platform module drives continuous improvement of the system to achieve a "perception-decision-optimization" closed loop.

[0027] The innovation of the production planning management module is the intelligent order splitting based on knowledge graph, the innovation of the dynamic scheduling engine module is the distributed scheduling driven by federated learning, the innovation of the resource scheduling module is the virtual-real resource mapping enabled by digital twins, the innovation of the real-time monitoring system module is multimodal AI fusion monitoring, the innovation of the exception handling center module is the root cause analysis driven by causal reasoning, and the innovation of the data optimization platform module is the physical information neural network. The production planning management module can avoid the problem of relying on manual experience to disassemble the bill of materials, which is prone to errors and time-consuming, and manual scheduling is inefficient and cannot cope with the problem of emergency insertion. The dynamic scheduling engine module can avoid To avoid the problems of high response delay in centralized scheduling, inability to adapt to multi-workshop collaboration, and high equipment idle rate due to fixed scheduling, the resource scheduling module can avoid the problems of physical resources and system data being out of sync, high scheduling distortion rate, and long production cycle due to material waiting. The real-time monitoring system module can avoid the problems of high false alarm rate of a single sensor and batch scrapping caused by delayed detection of quality anomalies. The exception handling center module can avoid the problems of reliance on expert rule bases, long time-consuming new fault diagnosis, and long average fault repair time. The data optimization platform module can avoid the problems of pure data-driven models violating physical laws and relying on experience leading to optimization lags.

[0028] The dynamic scheduling engine responds in milliseconds. The multi-agent reinforcement learning engine based on the federated learning architecture realizes millisecond-level response scheduling. Each device acts as an autonomous decision-making unit and is dynamically coordinated through the distributed Q learning algorithm. The idle rate of the equipment is greatly reduced, the order delivery cycle is compressed, and the response speed of emergency orders is improved. Multimodal AI quality monitoring integrates the long-short-term memory network deep analysis model of vibration, thermal imaging, and current spectrum. The detection rate is greatly improved compared to that of a unified sensor. Causal reasoning root cause analysis, construction of fault cause and effect diagram to locate deep problems, reduce the average repair time, and the system self-recovery time is shorter when the equipment fails. It supports GPU acceleration and the scheduling calculation can be accelerated by 50 times. Through TOC (constraint theory (Theory) Increase throughput by 22%, achieve full-link tracking of materials → equipment → personnel, SPC control charts realize early detection of defects and reduce scrap rate, data preprocessing reduces bandwidth requirements, the automatic rate of first-level faults is improved, and fault tree analysis is formed into an enterprise knowledge base, reducing the strategy iteration cycle from weeks to hours, predictive maintenance reduces unexpected downtime losses, the order splitting speed is improved, dynamic adaptation to process changes, support for 100+ production line units collaborative scheduling, breaking the "scheduling island", achieving cross-factory production capacity balance, improving resource scheduling accuracy, supporting "quantum-classical" hybrid computing resource allocation, improving fault detection rate, reducing 80% of sensor deployment, increasing model generalization capability by 5 times, and reducing 50% of simulation verification requirements.

[0029] Example 2: Figure 1-Figure 4The technical scheme shown in the application provides the following technical scheme: an AI-based automatic production line scheduling system in an industrial internet, which discloses the following: a production plan management module decomposes orders based on conversion processes from a bill of materials to a process route, and outputs a directed acyclic graph; through an earliest due date priority algorithm, the module prioritizes work orders, outputs corresponding work order queues, realizes capacity balancing through a linear programming constraint solver, and generates a corresponding device load thermal map; at the same time, the module checks material sets through a real-time interfaced warehouse operation management system, realizes material shortage early warning reports, and a dynamic scheduling engine module performs multi-objective optimization scheduling through a genetic algorithm to minimize optimization of completion time and energy consumption; when a device fails, the module can start rescheduling, real-time rescheduling has fast response speed, automated guided vehicle path collaborative work reduces material handling time, device resources in a resource scheduling module are based on preemptive scheduling of device states, and cross-platform communication real-time data acquisition technology provides technical support therefor; material resources are bound to just-in-time feeding and wireless radio frequency identification under the support of automated guided vehicle scheduling system integration technology, human resources are matched to skill matrices through real-time visual management system interfacing technology and task pushing through a mobile terminal, realizing full-link tracking of materials, devices, and personnel, monitoring dimensions in a real-time monitoring system module include device state monitoring, quality process control, and energy consumption management, device state monitoring realizes real-time detection of OEE (overall equipment effectiveness) through cooperation of vibration / temperature sensors and edge computing, quality process control realizes control of a process capability index through machine design and SPT (statistical process control), and energy consumption management is realized through monitoring of a unit product energy consumption curve based on data feedback of a smart meter and an energy analysis model, an abnormality processing center module classifies fault levels into three levels, the first level is a transient fault, a transient fault PLC program can be automatically triggered to automatically restore the device, the second level is a device-level fault, when a device-level fault occurs, a failure mode analysis library is needed to analyze the failure cause to notify maintenance personnel to switch the corresponding device, the third level is a system-level fault, a digital twin simulation verification system is used to verify system fault defects, the fault device is safely shut down, and an emergency production plan is started to ensure normal production, a data optimization platform module has a digital twin simulation function, the digital twin simulation function realizes pre-verification of scheduling strategies through 3D production line modeling and a physics engine, the data optimization platform module also has a predictive maintenance function, a long short-term memory network deep analysis model can give a 3-day early warning for bearing failure, and the data optimization platform module also has a KPI deep analysis function, and through association rule mining, it can find that voltage fluctuation leads to yield decline.

[0030] The production plan management module constructs a process knowledge graph: device capacity, material attributes, and process constraints are converted into a relationship network, and automatic process decomposition: the optimal process path is predicted through a graph neural network.

[0031] The overall algorithm formula of the production planning management module is:

[0032]

[0033] Where $\mathcal{P}^*$: optimal process path, $\mathbb{P}$: set of all feasible paths, $X$: node feature matrix, $A$: adjacency matrix.

[0034] Dynamic scheduling engine module, federated scheduling architecture: each production line edge node locally trains the scheduling model and only shares the model parameters, multi-agent reinforcement learning: each automated guided vehicle / robotic arm makes autonomous decisions as an agent.

[0035] The core scheduling formula of the dynamic scheduling engine module is:

[0036]

[0037] T j =max(0,f j -d j )(Delay time

[0038]

[0039] Variable Description Variable Description:

[0040] xj,k(t): A binary decision variable indicating whether task j is assigned to resource k at time t.

[0041] fj: actual completion time of task j

[0042] :Vector element-wise inequalities

[0043] Resource scheduling module, real-time mirroring: building millisecond-level digital twins through 5G+ time-sensitive networks, resource DNA encoding: assigning a unique quantum encryption ID to each device / material to ensure full-link traceability.

[0044] The function utilization formula in digital twin resource mapping is:

[0045] Assume the rated power is P rated , current power P current =18.7kW.

[0046]

[0047] Here, α is the ideal load ratio (e.g., α=0.8), which penalizes deviation from the optimal load.

[0048] Cross-modal learning of the real-time monitoring system module: integrating vibration, thermal imaging, voiceprint, and current data to train a long short-term memory network deep parsing model; self-supervised anomaly detection: using normal data for training without the need to label fault samples.

[0049] The core algorithm of cross-modal learning is:

[0050]

[0051] in K(t)=[f1(X1(t));…;f M (X M (t))]

[0052] g(·)=LSTM(Concat[CrossAttn(X i (tk),X j (tk))] i≠j )

[0053] σ(·): Sigmoid activation function

[0054] f m (·): Modality-specific encoders (e.g., CNN for images, 1D-CNN for vibrations)

[0055] CrossAttn(·): Cross-modal attention (see below for details)

[0056] Where k: time series sliding window size

[0057] Abnormal processing center module, causal discovery algorithm: build fault causal graph based on PC algorithm, meta-learning: learn diagnostic strategies from historical faults and adapt to unknown faults.

[0058] Data optimization platform module, PINN architecture: embeds equipment physics equation constraints in the loss function, end-to-end optimization: directly outputs executable scheduling improvement instructions (such as "increase injection pressure by 8%").

[0059] Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. The AI-based automated production line scheduling system in the Industrial Internet includes S1: production plan management module, S2: dynamic scheduling engine module, S3: resource scheduling module, S4: real-time monitoring system module, S5: exception handling center module, and S6: data optimization platform module. It is characterized by: The production plan management module converts high-level production plans into executable work orders. The dynamic scheduling engine module responds to production disturbances such as equipment failures / interrupted orders in real time and dynamically optimizes resource allocation. The resource scheduling module accurately matches equipment, materials, and manpower requirements. The real-time monitoring system module perceives the production line status in seconds. The exception handling center module quickly locates the root cause of the fault. The data optimization platform module drives continuous improvement of the system.

2. The AI-based automated production line scheduling system in the Industrial Internet according to claim 1 is characterized by: The production planning management module breaks down orders based on the transformation process from bill of materials to process route, and outputs a directed acyclic graph. It prioritizes work orders through the earliest delivery priority algorithm, outputs the corresponding work order queue, achieves production capacity balance through a linear programming constraint solver, and generates a corresponding equipment load heat map. At the same time, it checks the completeness of materials through the real-time docking warehouse operation management system to realize material shortage warning reports.

3. The AI-based automated production line scheduling system in the Industrial Internet according to claim 1 is characterized in that: The dynamic scheduling engine module uses genetic algorithms to perform multi-objective optimization scheduling, minimizes completion time and energy consumption, and can start rescheduling when equipment fails. The real-time rescheduling has a fast response speed, and the automatic guided vehicle path cooperates to reduce material handling time.

4. The AI-based automated production line scheduling system in the Industrial Internet according to claim 1, characterized in that: In the resource scheduling module, equipment resources are preemptively scheduled based on equipment status, and technical support is provided by cross-platform communication real-time data collection technology. Material resources are bound to on-time delivery and wireless radio frequency identification with the support of automatic guided vehicle scheduling system integration technology. Human resources match the skill matrix through real-time visual management system docking technology and use mobile terminals to push tasks, realizing full-link tracking of materials → equipment → personnel.

5. The AI-based automated production line scheduling system in the Industrial Internet according to claim 1 is characterized by: The monitoring dimensions in the real-time monitoring system module include three aspects: equipment status monitoring, quality process control and energy consumption management. Among them, equipment status monitoring uses the interaction between vibration / temperature sensors and edge computing to conduct real-time detection of the overall efficiency of the equipment. Quality process control realizes the control of process capability index through machine design and statistical process control. At the same time, the energy consumption curve of unit product is monitored through data feedback from smart meters and energy analysis models to realize energy consumption management.

6. The AI-based automated production line scheduling system in the Industrial Internet according to claim 1, characterized in that: The exception handling center module divides the fault level into three levels. The first level is instantaneous fault. The instantaneous fault PLC program can be automatically triggered to make the equipment automatically recover. The second level is equipment-level fault. When an equipment-level fault occurs, the failure mode analysis library is required to analyze the cause of the failure, so as to notify the maintenance personnel to switch the corresponding equipment. The third level is system-level fault. The system fault defect is verified through digital twin simulation, and the faulty equipment is safely shut down. At the same time, the emergency production plan is activated to ensure that production remains normal.

7. The AI-based automated production line scheduling system in the Industrial Internet according to claim 1, characterized in that: The data optimization platform module has a digital twin simulation function, which pre-verifies the scheduling strategy through 3D production line modeling and a physical engine. The data optimization platform module also has a predictive maintenance function. Through the long short-term memory network deep analysis model, it can provide a three-day advance warning of bearing failures. At the same time, the data optimization platform module also has a KPI in-depth analysis function. Through association rule mining, it can be found that "voltage fluctuation → yield decline".

8. The AI-based automated production line scheduling system in the Industrial Internet according to claim 1, characterized in that: The innovation of the production planning management module is the intelligent order splitting based on knowledge graph, the innovation of the dynamic scheduling engine module is the distributed scheduling driven by federated learning, the innovation of the resource scheduling module is the virtual-real resource mapping enabled by digital twins, the innovation of the real-time monitoring system module is multimodal AI fusion monitoring, the innovation of the exception handling center module is the root cause analysis driven by causal reasoning, and the innovation of the data optimization platform module is the physical information neural network.

9. The AI-based automated production line scheduling system in the Industrial Internet according to claim 8, characterized in that: The production planning management module can avoid the problem of relying on manual experience to disassemble the bill of materials, which is prone to errors and time-consuming, the low efficiency of manual scheduling, and the inability to cope with emergency orders. The dynamic scheduling engine module can avoid the problem of high response delay in centralized scheduling, the inability to adapt to multi-workshop collaboration, and fixed scheduling leading to high equipment idle rate. The resource scheduling module can avoid the problem of physical resources and system data being out of sync, high scheduling distortion rate, and material waiting occupying a long production cycle. The real-time monitoring system module can avoid the problem of high false alarm rate of a single sensor and delayed discovery of quality anomalies leading to batch scrapping. The exception handling center module can avoid the problem of relying on expert rule base, long time-consuming new fault diagnosis, and long average fault repair time. The data optimization platform module can avoid the problem of pure data-driven models violating physical laws and relying on experience to cause optimization lag.

Images