Digital twin workshop production cooperation adaptive production change reconstruction method

Abstract

The invention provides a digital twin workshop production collaboration adaptive production change reconstruction method, which belongs to the technical field of digital twin production, and comprises the following steps: step 1, constructing a multi-level main body monitor which comprises a node layer, a community layer, a sub-graph layer and a network layer, and realizing monitoring and prediction for a digital twin workshop production collaboration process; 2, comprehensively monitoring the production process by using a monitor of a multi-level main body; step 3, carrying out production fluctuation analysis in combination with a monitor of a multi-level main body by means of an online detector aiming at multi-source interference; and step 4, constructing a production reconstruction joint decision-making device which comprises a community layer decision-making device and a network layer decision-making device, and completing production configuration and task allocation. According to the method, the workshop production cooperation efficiency and the adaptive production switching capability can be effectively improved, the method adapts to complex and changeable industrial production environments, and the method has important practicability and popularization value.

Description

Technical Field

[0001] The present invention belongs to the field of digital twin production technology, and particularly relates to a method for adaptive production change and reconstruction of production collaboration in a digital twin workshop. Background Art

[0002] Traditional workshop production management models are no longer able to cope with the increasingly complex production environment. Traditional assembly lines typically employ fixed workstation layouts and standardized production processes. These upgraded workshops utilize a cell-based system, allowing for flexible adjustments to equipment layout and production processes based on the diverse and dynamic nature of production tasks. Traditional workshop production management methods rely primarily on manual experience, a single data source, and static models, making it difficult to achieve real-time awareness, fluctuation detection, and dynamic optimization of the production process. This is particularly true when faced with multi-dimensional disruptions such as equipment failures, material shortages, and operational deviations. Traditional methods often struggle to respond and adjust quickly, resulting in low production efficiency, wasted resources, and reduced product quality. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to overcome the shortcomings of existing production change and reconstruction technologies in the production collaboration process of digital twin workshops, and propose a method for adaptive production change and reconstruction of digital twin workshop production collaboration. This method aims to effectively describe and associate the digital twin model of the models, data, status and behavior of the components such as equipment, workstations, work sections and production lines in the digital twin workshop production system. Through multi-level monitoring of the production process, real-time analysis of multi-source interference and intelligent decision-making for production reconstruction, it provides a basis for rapid production change, efficient collaboration, optimized analysis and accurate decision-making of subsequent production systems and their operation processes, thereby improving the flexibility and adaptability of production collaboration in digital twin workshops, as well as the operating efficiency and control capabilities during the production change and reconstruction process.

[0004] The digital twin workshop production collaboration and adaptive production change and reconstruction method proposed in the present invention constructs an efficient system for real-time monitoring, fluctuation detection and dynamic production change and reconstruction through independent and collaborative monitors, detectors and decision makers. This method can integrate multi-dimensional status data such as equipment parameters, material inventory, and personnel operations in real time, build a dynamic production process model, and realize full life cycle management of the production process. By analyzing the sources of production fluctuations under multi-source interference, identifying potential risks and providing real-time feedback, and optimizing production configuration and task allocation through joint decision makers, it ensures maximum resource utilization, taking into account production efficiency, cost control and product quality. This method can effectively improve the production collaboration efficiency and adaptive production change capabilities of the workshop, provide intelligent solutions for complex and changeable industrial production environments, and has important practicality and promotion value.

[0005] The present invention solves the technical problem by adopting the following technical solution: a method for adaptive production change and reconstruction of production collaboration in a digital twin workshop, comprising the following steps:

[0006] Step 1: Build a multi-level agent monitor, including the node layer, community layer, sub-layer layer, and network layer, to monitor and predict the production collaboration process of the digital twin workshop;

[0007] Step 2: Use multi-level subject monitors to comprehensively monitor the production process;

[0008] Step 3: Combine multi-level subject monitors and conduct production fluctuation analysis with online detectors for multi-source interference;

[0009] Step 4: Build a production reconstruction joint decision maker, which includes a community-layer decision maker and a network-layer decision maker to complete production configuration and task allocation.

[0010] The advantages of the present invention compared with the prior art are:

[0011] (1) This invention proposes a method for constructing a multi-level subject monitor for a digital twin workshop. It can effectively describe the demand allocation, OEE, and maximum completion time of the "equipment-workstation-work section-production line" in the workshop from the perspective of production collaboration, which is conducive to the digital twin virtual workshop to more realistically and comprehensively reflect the production status of the physical workshop, improve the monitoring effect of the physical workshop, and thus effectively improve the production efficiency of the digital twin workshop and optimize the allocation of production resources.

[0012] (2) The present invention proposes a method for adaptive reconstruction of production collaboration, which can effectively consider the production status of workshop equipment. When the workshop faces emergencies such as equipment failure and order changes, the method can quickly respond based on the equipment status fed back in real time by multi-level subject monitors. For example, once a key equipment failure is detected, the system immediately evaluates the role and impact of the equipment in the entire production process, replans the affected production tasks, and reasonably allocates them to other alternative equipment or workstations, minimizing the delay of the failure on the overall production schedule. At the same time, in response to the problem of order changes, by analyzing the priority of raw material demand in different work sections and workstations, key production links are prioritized to maintain production continuity. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 This is a structural diagram of the workshop production environment;

[0014] Figure 2 It is a structural diagram of a multi-level subject monitor;

[0015] Figure 3 A flow chart for dynamic optimization of the production process. DETAILED DESCRIPTION

[0016] In order to make the objectives, technical solutions, and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are intended only to illustrate the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other. To achieve the above-mentioned objectives, the present invention adopts the following technical solutions.

[0017] like Figure 1 As shown, the present invention provides a method for adaptive production change and reconstruction of production collaboration in a digital twin workshop, and the specific implementation method is as follows:

[0018] Step 1: Construction of multi-level subject monitor, multi-level subject monitor such as Figure 2 As shown, the specific implementation is as follows:

[0019] The digital twin workshop is disassembled to construct a hierarchical architecture of equipment, work stations, work sections, and production lines. The digital twin workshop covers the four-layer structure of equipment, work stations, work sections, and production lines of the physical workshop, which correspond to the node layer, community layer, sub-layer layer, and network layer of the monitor respectively.

[0020] a) Construct the node layer based on the device. Specifically, based on the digital twin method, build the geometric, physical, behavioral, and rule models of the device, and propose The monitoring variable set:

[0021] ,

[0022] in, Respectively represent devices progress Actual production capacity, actual production volume, rated production capacity, rated production volume, expected production capacity, expected production volume, It also proposes a method for predicting equipment production status based on the equipment's maximum makespan and overall equipment effectiveness (OEE), enabling node-level monitoring of production demand and actual production load.

[0023] When the actual production volume of the equipment meets the expected production volume, the equipment stops working. The maximum completion time is ,Right now ,in, Indicates the functions required for the current task, and Respectively represent devices The actual production volume and the expected production volume.

[0024] Based on the maximum completion time of the equipment and the comprehensive utilization efficiency of the equipment, the comprehensive utilization efficiency of the equipment is usually The equipment production status prediction method takes into account the maximum completion time of the equipment based on the equipment comprehensive utilization efficiency OEE. ,The equipment production status prediction method is specifically as follows, including availability , performance factor , quality factor ,but:

[0025] ,

[0026] in, is the maximum completion time of the equipment, For device targeting The rated production capacity of the function, The equipment production cycle.

[0027] b) Construct a community layer based on the workstation, specifically: propose a prediction method for the average OEE and OEE balance of the community, and the maximum completion time of the community, to achieve community-level monitoring of the production demand and actual production load of each community in the workstation; the workstation is composed of communities with the same functions Node and their related relationships to form communities , is the index number of the community, where For the community The node set in For the community The number of midpoints.

[0028] Calculation method for the average OEE of the community: ;

[0029] OEE balance calculation method: ;

[0030] Calculation method of the maximum completion time of a community: ;

[0031] Calculation method of the maximum completion time balance of the community: ;

[0032] c) Construct sub-layers based on work sections. Specifically, a prediction method for the maximum completion time of a work section is proposed to achieve sub-graph-level monitoring of the production demand and actual load of the work section sub-graph. The sub-graph is composed of workstation clusters with different functions. ,in Indicates the different functions of the community, Indicates the number of the community, the former is the section index number, the latter Number the communities within the section.

[0033] Calculation method for the maximum completion time of a work section:

[0034] ;

[0035] in, For subgraph The last community to complete its production task.

[0036] d) Network layer construction, based on the production line, specifically: propose a prediction method for the overall network OEE and the maximum completion time of the production line, and realize the network-level prediction of the overall production demand and actual load of the production line. The network consists of all the sub-graphs of the workshop production line. .

[0037] Calculation method for the overall network OEE:

[0038] ;

[0039] in, is the set of all work-section subgraphs in the network layer, is the set of all work section communities in the sublayer, Represents a community Average OEE.

[0040] Production line maximum completion time prediction method:

[0041] ;

[0042] in, is the set of all subgraphs in the network layer, Represents a work section subgraph The maximum completion time.

[0043] Step 2: Figure 1 Production process monitoring, specific implementation methods are as follows Figure 3 As shown, the details are as follows:

[0044] Use the multi-level subject monitor constructed in step 1 to monitor the production process: the workshop produces according to the production plan and determines whether the task sequence is 0. If it is 0, the production is terminated; if it is not 0, the inspection task is checked to see whether it can be completed on time. If so, the multi-level subject monitor is continued to be used for monitoring; if not, the online detector is used to perform online detection of multiple types of events in the workshop.

[0045] After obtaining the summarized supply and demand relationship through the online detector, check whether the current production line can meet the demand. If so, use the community layer decision maker to perform production configuration and task allocation. The community layer decision maker mainly redistributes the task load within a single production community within the community layer.

[0046] If not, the network-layer decision maker is used to perform production configuration and task allocation. The adaptive reconstruction plan is then checked to see if it meets the adaptive reconstruction requirements, that is, whether it minimizes the network's OEE balance. If not, the network-layer decision maker is used to perform production configuration and task allocation. If so, production continues according to the new production plan. The network-layer decision maker primarily performs adaptive production changes and reconstruction across multiple clusters of production lines in the network layer.

[0047] Step 3: Online detector for multi-source interference, the specific implementation method is as follows: for production fluctuations in the production process, online detection of the following three types of events is implemented.

[0048] Event 1: Device performance degradation or failure: Detect whether the device is faulty or aging. If the device is normal, summarize and analyze the supply and demand relationship at that time. If there is a problem, Event 1 is detected, and it is necessary to identify the faulty device and check whether the faulty device can be immediately restored. If so, restore the faulty device. If not, delete the faulty device.

[0049] Event 2: Rework of defective products: Check whether there are defective products that need to be reworked. If there are no defective products, summarize and analyze the supply and demand relationship at that time. If there is a problem, Event 2 is detected, and it is necessary to identify the cluster that causes the product defects, then update the corresponding workshop task sequence, and summarize and analyze the supply and demand relationship at that time.

[0050] Event 3: New order arrives: Check whether the demand has changed. If not, summarize and analyze the supply and demand relationship at that time. If it has changed, event 3 is detected, and the corresponding workshop task sequence needs to be updated, and the supply and demand relationship at that time is summarized and analyzed.

[0051] Step 4: Build a joint decision maker for production reconfiguration. The specific implementation is as follows:

[0052] The production reconfiguration joint decision maker includes a community-level decision maker and a network-level decision maker. The community-level decision maker redistributes tasks based on the demands that the current production line can meet. The task redistribution process is based on the production capacity of the current equipment as follows:

[0053] ;

[0054] in, and communities Any node within Expected production volume and actual production capacity, For the community expected production volume.

[0055] When the current production line cannot meet demand, the network-layer decision maker uses a reinforcement learning algorithm to reconfigure the production line's limited resources. The reinforcement learning algorithm includes a state space, an action space, and feedback.

[0056] The state space contains the real-time production status and production collaboration network topology ,Right now:

[0057] ;

[0058] in, It is a graph convolution function, based on which the graph convolution of the current production status and the production collaboration network topology, i.e., the state space, is realized.

[0059] The action space contains two action strategies, namely equipment function reconstruction and equipment maintenance.

[0060] The steps of equipment function reconstruction are to select the community with the smallest average OEE, select the node with the lowest similarity in the community, adjust the production function of the node, and adjust the position of the community with the smallest average OEE to the community with the longest maximum completion time; equipment maintenance is to reduce the abnormal maximum completion time of the node through maintenance.

[0061] Feedback is to minimize the network OEE balance, the formula is:

[0062] ;

[0063] in, is the set of all subgraphs in the network layer, It is the collection of all work section communities in the sublayer.

[0064] In summary, the present invention proposes a method for adaptive production change and reconstruction of production collaboration in a digital twin workshop. The method includes four steps: constructing a multi-level subject monitor for the production process, using the multi-level subject monitor to monitor the production process, using the online detector to analyze production fluctuations, and using a joint decision maker to perform production configuration and task allocation. On the one hand, a method for constructing a multi-level subject monitor for a digital twin workshop is proposed, which can effectively describe the production status of "equipment-workstation-work section-production line" in the workshop; on the other hand, a method for adaptive reconstruction considering production collaboration is proposed, which effectively considers the production status of the workshop and responds quickly based on the equipment status fed back in real time by the multi-level subject monitor.

[0065] The contents not described in detail in the specification of the present invention belong to the prior art known to those skilled in the art.

[0066] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A method for adaptive production change and reconstruction in a digital twin workshop production collaboration, characterized by: The following steps are involved: Step 1: Build a multi-level agent monitor, including the node layer, community layer, sub-layer layer, and network layer, to monitor and predict the production collaboration process of the digital twin workshop; Step 2: Use multi-level subject monitors to comprehensively monitor the production process; Step 3: Combine multi-level subject monitors and conduct production fluctuation analysis with online detectors for multi-source interference; Step 4: Build a production reconstruction joint decision maker, which includes a community-layer decision maker and a network-layer decision maker to complete production configuration and task allocation.

2. The method for adaptive production change and reconstruction of a digital twin workshop production collaboration according to claim 1, characterized in that: Step 1 includes: disassembling the digital twin workshop and building a hierarchical architecture of equipment, workstations, work sections, and production lines, corresponding to the node layer, community layer, sub-layer layer, and network layer respectively; Equipment-based node layer construction: Based on the equipment's overall utilization efficiency (OEE) and the equipment's maximum completion time, the equipment production status prediction method is used to achieve node-level monitoring of production demand and actual production load; Construction of a community layer based on workstations: Based on the prediction method of the average equipment utilization efficiency (OEE) and the balance degree of equipment utilization efficiency (OEE) of the community, as well as the maximum completion time of the community, the production demand and actual production load of each community in the process section can be monitored at the community level; Construction of sub-layers based on work sections: Based on the prediction method of the maximum completion time of the work section, sub-graph-level monitoring of the production demand and actual production load of the work section community is achieved; Production line-based network layer construction: Based on the prediction method of the overall equipment utilization efficiency (OEE) of the network and the maximum completion time of the production line, network-level prediction of the overall production demand and actual production load of the production line is achieved.

3. The method for adaptive production change and reconstruction of a digital twin workshop production collaboration according to claim 2, characterized in that: Constructing the node layer specifically involves: constructing the geometric, physical, behavioral, and rule models of the device, and proposing The monitoring variable set: ； in, Respectively represent devices progress Actual production capacity, actual production volume, rated production capacity, rated production volume, expected production capacity, expected production volume, Indicates time; and proposes a method for predicting equipment production status based on the equipment's maximum completion time and equipment comprehensive utilization efficiency (OEE), to achieve node-level monitoring of production demand and actual production load; When the actual production volume of the equipment meets the expected production volume, the equipment stops working. The maximum completion time is ,Right now ,in, Indicates the functions required for the current task, and Respectively represent devices Actual production volume and expected production volume; Comprehensive utilization efficiency of equipment for: ； in, is the maximum completion time of the equipment, For device targeting The rated production capacity of the function, For the equipment production cycle, is the availability rate, is the performance factor, is the quality factor.

4. The method for adaptive production change and reconstruction of a digital twin workshop production collaboration according to claim 2, characterized in that: Construct a community layer, specifically: the workstation consists of Node and their related relationships to form communities , is the index number of the community, where For the community The node set in For the community Number of midpoints; Calculation method for the average equipment comprehensive utilization efficiency (OEE) of the group: ; Equipment comprehensive utilization efficiency OEE balance calculation method: ; Calculation method of the maximum completion time of a community: ; Calculation method of the maximum completion time balance of the community: .

5. The method for adaptive production change and reconstruction of a digital twin workshop production collaboration according to claim 2, characterized in that: Construct sub-layers, specifically: the sub-map consists of workstation communities with different functions, ,in, Indicates the different functions of the community, , Both represent the number of the community, among which the former is the section index number, the latter Number the communities within the section; Calculation method for the maximum completion time of a work section: ； in, For subgraph The last community to complete its production task.

6. The method for adaptive production change and reconstruction of a digital twin workshop production collaboration according to claim 2, characterized in that: The network layer is constructed as follows: the network consists of all the subgraphs of the workshop production line ; Calculation method for the overall network OEE: ； in, is the set of all work-section subgraphs in the network layer, is the set of all work section communities in the sublayer, Represents a community Average OEE; Production line maximum completion time prediction method: ； in, Represents a work section subgraph The maximum completion time.

7. The method for adaptive production change and reconstruction of a digital twin workshop production collaboration according to claim 1, characterized in that: In step 2, the multi-level subject monitor performs production process monitoring, including: the workshop produces according to the production plan and determines whether the task sequence is 0. If it is 0, production ends; if it is not 0, it checks whether the task can be completed on time. If so, it continues to monitor using the multi-level subject monitor; if not, it uses the online detector to perform online detection of multiple types of events in the workshop; After obtaining the aggregated supply and demand relationship through the online detector, it is checked whether the current production line can meet the demand. If so, the cluster-level decision maker is used to configure production and allocate tasks. The cluster-level decision maker redistributes the task load within a single cluster for the production clusters within the cluster layer. If not, the network layer decision maker is used to perform production configuration and task allocation, and then check whether the adaptive reconstruction plan meets the adaptive reconstruction requirements, that is, whether the OEE balance of the comprehensive utilization efficiency of network equipment is minimized. If not, the network layer decision maker is continued to be used for production configuration and task allocation. If so, production continues according to the new production plan, in which the network layer decision maker performs adaptive production change reconstruction among multiple communities for the production lines in the network layer.

8. The method for adaptive production change and reconstruction of a digital twin workshop production collaboration according to claim 1, characterized in that: Step 3 includes: implementing online detection of the following three types of events for production fluctuations during the production process; Event 1: Device performance degradation or failure: Detect whether the device is faulty or aging. If the device is normal, summarize and analyze the current supply and demand relationship. If there is a problem, Event 1 is detected, and the faulty device needs to be identified and checked for immediate recovery. If so, the device is restored. If not, the device is deleted. Event 2: Defective Product Rework: Checks for defective products that require rework. If not, summarizes and analyzes the current supply and demand relationship. If a problem exists, Event 2 is detected, and the cluster causing the product defect needs to be identified. The corresponding workshop task sequence is then updated, and a summary analysis of the current supply and demand relationship is conducted. Event 3: New order arrives: Check whether the demand has changed. If not, summarize and analyze the supply and demand relationship at that time. If it has changed, event 3 is detected, and the corresponding workshop task sequence needs to be updated, and the supply and demand relationship at that time is summarized and analyzed.

9. The method for adaptive production change and reconstruction of a digital twin workshop production collaboration according to claim 7, characterized in that: In step 4: The production reconfiguration joint decision maker includes a community-level decision maker and a network-level decision maker. The community-level decision maker redistributes tasks based on the demands that the current production line can meet. The task redistribution process is based on the production capacity of the current equipment as follows: ； in, and communities Any node within Expected production volume and actual production capacity, For the community Expected production volume; When the current production line cannot meet demand, the network-layer decision maker uses a reinforcement learning algorithm to reconfigure the production line's limited resources. The reinforcement learning algorithm includes a state space, an action space, and feedback. The state space contains the real-time production status and production collaboration network topology ,Right now: ； in, It is a graph convolution function, based on which the graph convolution of the current production status and the production collaboration network topology, i.e., the state space, is realized.

10. The method for adaptive production change and reconstruction of a digital twin workshop production collaboration according to claim 9, characterized in that: The action space contains two action strategies, namely, equipment function reconstruction and equipment maintenance; The steps for equipment function reconstruction are to select the cluster with the lowest average equipment overall utilization efficiency (OEE), select the node with the lowest similarity in the cluster, adjust the production function of the node, and adjust the cluster with the lowest average equipment overall utilization efficiency (OEE) to the cluster with the longest maximum completion time; equipment maintenance is to reduce the abnormal maximum completion time of the node through maintenance; Feedback is to minimize the OEE balance of network equipment comprehensive utilization efficiency. The formula is: ； in, is the set of all subgraphs in the network layer, It is the collection of all work section communities in the sublayer.

Images