Discrete workshop production simulation method driven by material flow element model

Abstract

The invention discloses a discrete workshop production simulation method driven by a material flow element model, and the method comprises the steps: firstly proposing the material flow element model to describe the flowing of a material, and refining a logistics decision function, a production behavior control function and control logic; secondly, designing a process interactive simulation algorithm based on the material flow element model to realize accurate production process simulation, including production simulation model construction and simulation operation calculation; defining a simulation model of a material flow connected graph as simulation input; then, an improved process interaction algorithm is designed based on the MFMM to achieve simulation run scheduling, which includes a clock propulsion mechanism and an MFMM-based process propulsion mechanism. According to the method, flexible modeling and accurate simulation of the workshop material flow are realized, and powerful technical support can be provided for production analysis and optimization of the discrete manufacturing industry.

Description

Technical Field

[0001] The invention belongs to discrete workshop production simulation technology, and in particular relates to a discrete workshop production simulation method driven by a material flow element model. Background Art

[0002] Against the backdrop of intensified competition and rapid technological development in the global manufacturing industry, the complexity of production processes continues to increase. Especially in discrete manufacturing workshops with multiple varieties and small batches, these complexities lead to frequent and dynamic changes in workshop material flow, posing huge challenges to the efficiency and flexibility of production systems [1]. Material flow modeling and simulation technology is an effective tool to address these challenges. It can deeply analyze the interactions between various links in the production process and optimize the production process, thereby improving overall efficiency, reducing waste and lowering operational risks [2]. Especially for material flow in discrete manufacturing systems, its dynamics are more significant and are affected by multiple factors such as production batches, equipment status and process scheduling. Therefore, a flexible and accurate modeling and simulation method is needed to achieve effective description and simulation of material flow in complex discrete manufacturing scenarios [3].

[0003] Many scholars at home and abroad have conducted in-depth research. For example, Jiang et al.

[14] proposed a seven-element modeling (SE) theory that can uniformly describe manufacturing systems, identify seven basic elements and introduce the concepts of logistics path network and service unit; Liu Huailan et al. [4] extended the VDEVS modeling method of hierarchical model and behavioral model based on the DEVS atomic model and five-dimensional model; Yin Chao et al. [5] established a dual-weighted directed production logistics network model with workstations and workshop storage units as nodes and logistics flow as edges; Zhang et al. [6] adopted a hierarchical modeling method to sort out the model data types and designed a spatiotemporal data model for training and verification on this basis; Xie et al. [7] refined the SE modeling process, proposed the MNOSE model and the model's operating mechanism; Liu Sai et al. [8] proposed a simulation parameter correction method based on the digital twin model, elaborated the simulation attribute update process in detail and developed a digital twin platform ; Li Liuzhen et al. [9] proposed to establish a production logistics EFSM component model based on the EFSM component theory, and mapped the model with the atomic model of DEVS; Liu Mingyuan et al.

[10] proposed a logical modeling and simulation method of the workshop production process based on a finite state machine by abstracting the key attributes and behaviors of the DMS workshop into a logical model; Fu Zhuorui et al.

[11] logically modeled the DMS workshop according to the hierarchical analysis method, and established a DT-DES platform data interaction mechanism on this basis; Glatt et al.

[12] applied a physical engine to simulate the interaction between workpieces and material handling systems in the production process, and proposed a material handling system of a physical workshop and its digital twin system based on physical simulation; Funke et al.

[13] used random block analysis to extract the internal structure of complex networks, and on this basis, modeled and predicted the material flow in the manufacturing system.

[0004] Discrete event simulation (DES) technology has been widely used in the production simulation process of discrete manufacturing workshops. Currently, the commonly used methods include event scheduling, process interaction and activity scanning

[14] . For example, Zhong Shouyi et al.

[16] used a graphical process interaction method to describe the relationship between processes in a discrete system; Huang Jianxin et al.

[17] implemented and discussed the process interaction algorithm using computer thread technology; Zhang et al.

[18] proposed a real-time event scheduling strategy based on multi-threading technology, and described in detail the implementation method and steps of real-time synchronization behavior; Guo Yongchen et al.

[19] used a five-step scanning method based on a multi-agent system to explore the simulation process control algorithm of a multi-agent system; Jiang

[14] used a segmented process interaction algorithm to form a multi-layer simulation architecture, thereby reducing the difficulty of implementing the simulation scheduling algorithm.

[0005] The above research has achieved certain results in the modeling and simulation analysis of workshop material flow, but there are still some shortcomings. The existing material flow model focuses on the description of macro structure and changes, but lacks in-depth expression of material flow and control process. This limitation limits the flexibility of the simulation model to a certain extent, thus affecting the accuracy of the simulation results.

[0006] References:

[0007] [1] Zhang Lin, Zhou Longfei. Modeling and simulation technology in manufacturing[J]. Journal of System Simulation, 2018, 30(06): 1997-2012. DOI: 10.16182 / j.issn1004731x.joss.201806001.

[0008] [2] Zhu Y, Cheng J, Liu Z, et al. Production logistics digital twins: Research profiling, application, challenges and opportunities [J]. Robotics and Computer-Integrated Manufacturing, 2023,84:102592.

[0009] [3]Reinhardt H,Weber M,Putz MA survey on automatic model generation for material flow simulation in discrete manufacturing[J].Procedia CIRP,2019,81:121-126.

[0010] [4] Liu Huailan, Yue Peng, He Chenhao, et al. Description method of virtual entity behavior model of discrete manufacturing workshop based on VDEVS[J]. Computer Integrated Manufacturing Systems, 2021, 27(10): 2950-2960. DOI: 10.13196 / j.cims.2021.10.019.

[0011] [5] Yin Chao, Gao Ling, Li Xiaobin. Modeling and analysis of discrete manufacturing workshop production logistics network based on complex network theory[J]. Computer Integrated Manufacturing Systems, 2020, 26(08): 2157-2169. DOI: 10.13196 / j.cims.2020.08.016.

[0012] [6] Zhang L, Guo Y, Qian W, et al. Modeling and online training method for digital twin workshop [J]. International Journal of Production Research, 2023, 61(12): 3943-3962.

[0013] [7]Xie J, Jiang H, Qin S, et al. A new description model for enabling more general manufacturing systems representation in digital twin[J]. Journal of Manufacturing Systems, 2024,72:475-491.

[0014] [8] Liu Sai, Guo Yu, Zhang Litong, et al. Parameter correction method for digital twin simulation of discrete manufacturing workshop[J]. Modular Machine Tool and Automated Machining Technology, 2023, (08): 1-5+12. DOI: 10.13462 / j.cnki.mmtamt.2023.08.001.

[0015] [9] Li Liuzhen, Jin Chao, Lin Tingyu, et al. Production logistics modeling and simulation of intelligent workshop manufacturing system based on EFSM[J]. Journal of System Simulation, 2023, 35(12): 2655-2668. DOI: 10.16182 / j.issn1004731x.joss.22-0802.

[0016]

[10] Liu Mingyuan, Xie Jiaxiang, Wu Hao, et al. Research on workshop logic modeling and simulation based on finite state machine[J]. Journal of System Simulation, 2023, 35(04): 853-861. DOI: 10.16182 / j.issn1004731x.joss.21-1218.

[0017]

[11] Fu Zhuorui, Zhao Ning, Zhu Yi. Design of flexible discrete event system modeling and simulation platform for digital twins[J / OL]. Computer Integrated Manufacturing Systems, 1-25[2024-03-14]. https: / / doi.org / 10.13196 / j.cims.2023.0564.

[0018]

[12] Glatt M, Sinnwell C, Yi L, et al. Modeling and implementation of a digital twin of material flows based on physics simulation[J]. Journal ofManufacturing Systems, 2021,58:231-245.

[0019]

[13] Funke T, Becker T. Complex networks of material flow inmanufacturing and logistics: Modeling, analysis, and prediction using stochasticblock models [J]. Journal of Manufacturing Systems, 2020, 56: 296-311.

[0020]

[14] Jiang H, Qin S, Fu J, et al. How to model and implement connections between physical and virtual models for digital twin application[J]. Journal of Manufacturing Systems, 2021,58:36-51.

[0021]

[15] Liu Yong, Wang Decai, Feng Zhengchao. Discrete event system simulation modeling and simulation strategy[J]. Journal of Southwest Normal University (Natural Science Edition), 2005, (06): 1019-1025. DOI: 10.13718 / j.cnki.xsxb.2005.06.017.

[0022]

[16] Zhong Shouyi. Graphical representation method for process interaction simulation [J]. Journal of System Simulation, 1992, (01): 31-38.

[0023]

[17] Huang Jianxin, Li Qun, Yu Wenguang, et al. Research on thread-based process interaction simulation framework [J]. Journal of System Simulation, 2011, 23(04): 652-658. DOI: 10.16182 / j.cnki.joss.2011.04.040.

[0024]

[18] Zhang

[0025]

[19] Guo Yongchen, Shen Yang. Multi-agent simulation control method based on discrete event simulation principle[J]. Computer Simulation, 2015, 32(06): 349-355.

[0026]

[20] Bahouth A,Crites S,Matloff N,et al.Revisiting the issue ofperformance enhancement of discrete event simulation software[C] / / 40th Annualsimulation symposium(ANSS'07).IEEE,2007:114-122. Summary of the invention

[0027] In order to achieve effective modeling and simulation of material flow in a workshop production process, the present invention provides a discrete workshop production simulation method driven by a material flow meta-model.

[0028] A discrete workshop production simulation method driven by a material flow meta-model of the present invention comprises the following steps:

[0029] Step 1: Material flow modeling based on the Material Flow Meta Model (MFMM).

[0030] (1) Definition of Material Flow Metamodel MFMM.

[0031] The material flow metamodel MFMM is defined as follows:

[0032] MFMM= <MSN in ,MSN out ,DE,F> (1)

[0033] Among them: MSN in It is the input service node of material flow; MSN outIt is the output service node of the material flow; DE is the directed edge of the material flow connecting two MSNs; F is the flow entity between the node and the edge, that is, the material.

[0034] In the material flow network, the material point of the equipment is defined as MSN, including the docking point of the logistics equipment, the loading and unloading point of the processing equipment, and the service point of the processing equipment. The definition formula is as follows:

[0035] MSN= <State,DEs,F,T,δ in ,δ out ,α run ,P,Q> (2)

[0036] Where: State represents the state of the node, including waiting, serving, and blocking; DEs is the set of directed edges of material flow, and one MSN can be associated with multiple MSNs; F represents the flow entity in the current node; T is the task execution time of MSN; δ in is the logistics input decision; δ out is the logistics output decision, α run Represents production behavior; P represents the production equipment to which MSN belongs, and Q is the queue at the material point for storing blocked materials.

[0037] The material logistics process is defined as DE, including handling, conveying and transportation. Different equipment types lead to different control logics, so α is used. run Represents different production behaviors, and the definition formula is:

[0038] DE= <MSN in ,MSN out ,Dirction,E,State,Rule,T,α run > (3)

[0039] Where: MSN in It is the input service node, MSN out is the output service node, Dirction represents the direction of the edge, including forward, reverse and bidirectional, E is the execution device, State is the state of the directed edge, including waiting, serving and blocking, Rule is the logistics rule, T is the execution time on the edge, α run It is production behavior control.

[0040] (2) Material flow metamodel MFMM decision and control logic.

[0041] Logistics input decision function δ in Determine whether the material is transferred from the current material flow metamodel to the next material flow metamodel, which is used for the initialization of the material flow metamodel;δ inIt consists of logistics condition judgment and condition change; among them, logistics condition judgment is further divided into basic judgment and custom judgment; basic judgment is mainly based on MSN / DE status, queuing rules and execution conflict judgment; execution conflict is used to analyze whether there will be logistics equipment execution conflict and production equipment execution conflict, and custom judgment allows some judgment conditions to be flexibly set according to specific production needs and scenarios to better adapt to the complex and changeable production environment; according to whether the logistics conditions are met, it can be divided into satisfied condition change and unsatisfied condition change, among which MSN blocking is production equipment execution conflict, and DE blocking is logistics equipment execution conflict; logistics output decision function δ out It determines the flow of materials from DE to MSN out .

[0042] Production behavior control function α run It is the behavioral abstraction of the actual production equipment, defining and controlling the production behavior of the basic element MSN / DE of the material flow metamodel, which includes the logic of the production execution and post-execution of the equipment to which MSN / DE belongs. run It refers to the processing, assembly and disassembly activities of production equipment, which is essentially a production time probability function; in DE, α run It is the movement process of the executing equipment, including the transportation process of the conveying equipment, the transportation process of the transportation equipment trajectory planning and the loading and unloading process of the handling equipment motion control. In essence, it is the movement time function of the executing equipment; the production time probability function is a probability time function formed by adding a certain probability distribution to the execution time of the production task of the processing equipment.

[0043] The kinematics solution, dynamics solution, trajectory calculation and conveyor belt and AGV trajectory planning of the robot arm are produced according to the probabilistic time; the production behavior control outputs a random time value t under known conditions; after the production is completed, different post-production execution logics are called according to the executed production elements.

[0044] The logistics decision and production behavior control of the material flow meta-model ultimately change the state of MSN / DE. The internal behavior of MSN / DE and the state transition of external input events are modeled through the finite state machine to describe the state transition conditions of the model during its life cycle; the state of MNS / DE includes waiting, serving and blocking. i is the conversion event of MNS, b i It is the conversion event of DE.

[0045] Step 2: Material flow simulation based on the material flow metamodel MFMM; including production simulation model construction and simulation operation calculation.

[0046] (1) Simulation model generation.

[0047] In order to describe various complex simulation input parameters as a unified simulation model, an MFG based on MFMM is defined as follows:

[0048] MFG=<Sr,Task> = <MFS1,MFS2,...,MFS n > (4)

[0049] The simulation input parameters include the static resource information Sr of the production workshop and the dynamic scheduling task information Task; based on the SE model, the production time and logistics time, layout information, logistics path, and delivery method of the equipment are described as static resources of the workshop; the production scheduling plan is to plan and arrange specific tasks of specific production activities for specific production processes or links, including processes, workpiece numbers, processing quantities, task equipment, start time, and end time; through the matching of the production scheduling plan and the static resources of the workshop, a process material flow connection subgraph MFS based on MFMM is formed, as shown in the following formula;

[0050] MFS= <MFMM1,MFMM2,...,MFMM n > (5)

[0051] The production process of materials is composed of multiple production processes, that is, a combination of multiple process material flow connection sub-graphs, which ultimately form the production simulation model MFG of the material's entire life cycle.

[0052] (2) Simulation operation scheduling based on material flow metamodel MFMM.

[0053] Design a process interaction algorithm based on MFMM to implement the simulation clock advancement mechanism in DES and the process interaction mechanism based on MFMM.

[0054] The simulation life cycle of the flow entity is the process; each flow entity has an MFG about the entire production process of the material, and the MFMM model is the basic unit of simulation execution; the simulation clock is advanced to the minimum unconditional delayed resurrection time Ta of the material in the future event table FEL, and the MFMM model of all materials in the entire FEL is updated within the Ta time period, and the next Ta is calculated and updated. The simulation is continuously looped until all materials complete the production task, forming a segmented simulation algorithm.

[0055] In order to realize the time advancement of complex simulation systems, for the process interaction method, the time complexity of the priority queue in obtaining Ta is O(1); in FEL, the MFMM production events Ta of the materials are sorted by their size. If Ta is the same, they are sorted according to the order of occurrence and priority of the events.

[0056] In order to refine the clock management of each material, the simulation clock is divided into the simulation master clock SimClock and the entity simulation sub-clock SubClock; SimClock is the main clock of the entire simulation system, responsible for managing the advancement of the global simulation time, controlling the entire simulation process, determining the simulation start time, simulation step step, end time and the time multiplier with the real world, and the SubClock of all materials are synchronized with it; after the process advancement is executed, SubClock calculates the material Ta for each simulation step update.

[0057] For the process advancement mechanism based on MFMM, the main purpose is to update the material process Ta. The update of Ta is realized through the operation mechanism of MFMM. When the simulation clock advances the step length step, all material processes are traversed to execute the MFMM model. By judging the current material position, the MSN is determined. out Push strategy or DE push strategy, and finally calculate and update Ta and FEL of SimClock.

[0058] MSN out Promotion process:

[0059] In order to simulate the production and processing of materials, that is, the material conversion process, the α run Or update the delay state; in the MFMM model, the delay state is related to the state of MSN and DE. The idle and service states represent unconditional delays, that is, Ta is a finite time value, and the blocking represents conditional delays, that is, Ta is an infinite value. In the MFMM model, δ in and δ out To change the delay status, the process is as follows:

[0060] 1) If it is a conditional delay, execute MSN out δ out , and reconfigure the delay state.

[0061] 2) If it is an unconditional delay, execute MSN out Alpha run , the remaining time of production activities is calculated through the internal processing time probability function or a deeper process simulation algorithm, and the unconditional delay time of the flow entity is updated.

[0062] 3) Determine whether the current task is completed. If the current task is completed, set the next task as the DE advancement task.

[0063] DE pushes the process to realize material unloading, loading and transportation, that is, the material flow process. The push process is as follows:

[0064] 1) Determine the delay status of the current MFMM.

[0065] 2) If it is a conditional delay, execute MSN in δ in , reconfigure the delay state.

[0066] 3) If it is an unconditional delay, execute α of the directed edge DE run , through α run The internal AGV tracking motion algorithm, robotic arm motion control algorithm, etc. calculate the remaining motion time and update the Ta of the flow entity.

[0067] 4) Determine whether the current task is completed. If the current task is completed, set the next task to MSN out Advance the task.

[0068] The beneficial technical effects of the present invention are:

[0069] The present invention proposes a discrete manufacturing workshop production process simulation method based on a material flow metamodel to solve the problems of low modeling flexibility and poor simulation accuracy caused by the lack of flow and control process description in the material flow model. First, a material flow metamodel (MFMM) is constructed based on the seven-element unified modeling theory to describe logistics decisions and behavior control, thereby capturing the dynamic characteristics of material flow. Subsequently, an MFG simulation model is constructed based on the MFMM and a process interaction simulation method based on the MFMM is designed. The method can realize the material flow and behavior control process in the simulation, and realize effective and flexible calculation of the simulation. Finally, a simulation system is developed based on this method, and the accuracy and flexibility of the method are verified by examples.

[0070] This invention realizes flexible modeling and accurate simulation of workshop material flow, which can provide strong technical support for production analysis and optimization in discrete manufacturing industry. Future research work will focus on the analysis and evaluation of simulation results to realize intelligent simulation analysis of material flow. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 It is a graphic representation of the material flow element model of the present invention.

[0072] Figure 2 It is the execution logic of logistics decision and production behavior control of the present invention.

[0073] Figure 3 It is the finite state machine model of MSN / DE of the present invention.

[0074] Figure 4 Schematic diagram of material flow simulation method based on MFMM.

[0075] Figure 5 It is the advancement mechanism of the simulation clock and the process advancement mechanism based on MFMM.

[0076] Figure 6 Advance the flow graph for output level nodes and directed edges.

[0077] Figure 7 Model and simulate systems for material flows.

[0078] Figure 8 It is the MFG of the entire production process of a certain material.

[0079] Fig. 9 This is a comparison chart of the simulation average error. DETAILED DESCRIPTION

[0080] The present invention is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0081] A discrete workshop production simulation method driven by a material flow meta-model of the present invention comprises the following steps:

[0082] 1. Material flow modeling based on material flow metamodel MFMM.

[0083] (1) Definition of Material Flow Metamodel MFMM.

[0084] The basic components of a discrete manufacturing workshop include people, machines, materials, methods, and environments. Their relationships can be described as production organization relationships and logistics organization relationships. The former can organize production factors from a hierarchical perspective. The latter can describe the interactive relationship between production organization elements and form a material flow network structure. The basic elements of the material flow network structure are the material location of the equipment and the logistics form of the workpiece.

[0085] Therefore, the present invention defines the material flow relationship in the production workshop as a material flow meta model (MFMM) for the interaction logic of production organization elements, which is composed of a quaternary group of input and output material flow service nodes (MSN), material directed edges (DE) and flow entities (F), such as Figure 1 As shown, the material flow metamodel MFMM is defined as follows:

[0086] MFMM= <MSN in ,MSN out ,DE,F> (1)

[0087] Among them: MSN in It is the input service node of material flow; MSN out It is the output service node of the material flow; DE is the directed edge of the material flow connecting two MSNs; F is the flow entity between the node and the edge, that is, the material.

[0088] In the material flow network, the material point of the equipment is defined as MSN, including the docking point of the logistics equipment, the loading and unloading point of the processing equipment, and the service point of the processing equipment. The definition formula is as follows:

[0089] MSN= <State,DEs,F,T,δ in ,δ out ,α run ,P,Q> (2)

[0090] Where: State represents the state of the node, including waiting, serving, and blocking; DEs is the set of directed edges of material flow, and one MSN can be associated with multiple MSNs; F represents the flow entity in the current node; T is the task execution time of MSN; δ in is the logistics input decision; δ out is the logistics output decision, α run Represents production behavior; P represents the production equipment to which MSN belongs, and Q is the queue at the material point for storing blocked materials.

[0091] The material logistics process is defined as DE, including handling (such as loading and unloading by robotic arms), conveying (such as conveyor belt conveying) and transportation (such as AGV transportation). Different equipment types lead to different control logics, so α is used. run Represents different production behaviors, and the definition formula is:

[0092] DE= <MSN in ,MSN out ,Dirction,E,State,Rule,T,α run > (3)

[0093] Where: MSN in It is the input service node, MSN out is the output service node, Dirction represents the direction of the edge, including forward, reverse and bidirectional, E is the execution equipment such as robotic arms, AGV, etc., State is the state of the directed edge, including waiting, service and blocking, Rule is the logistics rule (such as first-in-first-out, last-in-first-out, etc.), T is the execution time on the edge, α run It is production behavior control.

[0094] (2) Material flow metamodel MFMM decision and control logic.

[0095] In the process of material flow and conversion in discrete manufacturing workshops, we are faced with the problem of describing the relationship between material flow logistics decision-making and production behavior control. To this end, the present invention proposes an operation logic for production behavior control and logistics decision-making of MFMM, such as Figure 2 shown.

[0096] Logistics input decision function δ in Determines whether the material is transferred from the current material flow metamodel to the next material flow metamodel, and is used for initialization of the material flow metamodel. Figure 2 (a) as shown; δ in It consists of logistics condition judgment and condition change; among them, logistics condition judgment is further divided into basic judgment and custom judgment; basic judgment is mainly based on MSN / DE status, queuing rules and execution conflict judgment; execution conflict is used to analyze whether there will be logistics equipment execution conflict and production equipment execution conflict, and custom judgment allows some judgment conditions to be flexibly set according to specific production needs and scenarios to better adapt to the complex and changeable production environment; according to whether the logistics conditions are met, it can be divided into satisfied condition change and unsatisfied condition change, among which MSN blocking is production equipment execution conflict, and DE blocking is logistics equipment execution conflict; logistics output decision function δ out It determines the flow of materials from DE to MSN out The logistics condition judgment and condition change are simpler than input decision, such as Figure 2 (b) as shown.

[0097] Production behavior control function α run It is the behavioral abstraction of the actual production equipment, defining and controlling the production behavior of the basic element MSN / DE of the material flow metamodel, which includes the logic of production execution and post-execution of the equipment to which MSN / DE belongs, such as Figure 2 (c). In MSN run It refers to the processing, assembly and disassembly activities of production equipment, which is essentially a production time probability function; in DE, α run It is the movement process of the executing equipment, including the transportation process of the conveying equipment, the transportation process of the transportation equipment trajectory planning and the loading and unloading process of the handling equipment motion control. In essence, it is the movement time function of the executing equipment; the production time probability function is to add a certain probability distribution (such as Poisson distribution, exponential distribution, normal distribution, triangular distribution, etc.) to the execution time of the production task of the processing equipment to form a probability time function.

[0098] The kinematics solution, dynamics solution, trajectory calculation and conveyor belt and AGV trajectory planning of the robot arm are produced according to the probabilistic time; the production behavior control outputs a random time value t under known conditions; after the production is completed, different post-production execution logics are called according to the executed production elements.

[0099] The logistics decision and production behavior control of the material flow meta-model eventually change the state of MSN / DE. The internal behavior of MSN / DE and the state transition of external input events are modeled through the finite state machine to describe the state transition conditions of the model during its life cycle; the state of MNS / DE includes waiting, serving and blocking. The state machine model is as follows Figure 3 As shown, in the state machine model, a i is the conversion event of MNS, b i is the transition event of DE. The state transition process of the MNS / DE finite state model is shown in Table 1.

[0100] Table 1 Conversion events of MSN / DE

[0101]

[0102]

[0103] 2. Material flow simulation based on material flow metamodel MFMM.

[0104] In the manufacturing workshop, in order to realize the preparation work before the execution of manufacturing, the production schedule generated by the production planning system (APS, Advanced Planning and Scheduling) is sent to the virtual workshop for simulation verification, simulating the deficiencies of the production schedule under various working conditions, and further promoting the optimization of the production schedule. To this end, the present invention designs a material flow simulation method based on MFMM, which includes the construction of a production simulation model and simulation operation calculation, such as Figure 4 As shown in the figure, a simulation model of Material Flow Graph (MFG) is defined based on MFMM as the simulation input; then, an improved process interaction algorithm is designed based on MFMM to realize simulation operation scheduling, which includes a clock advancement mechanism and a process advancement mechanism based on MFMM, which reduces the difficulty of implementing the simulation scheduling algorithm and improves the fine-grainedness of the simulation calculation process.

[0105] (1) Simulation model generation.

[0106] In order to describe various complex simulation input parameters as a unified simulation model, an MFG based on MFMM is defined as follows:

[0107] MFG=<Sr,Task> = <MFS1,MFS2,...,MFS n > (4)

[0108] The simulation input parameters include the static resource information Sr of the production workshop and the dynamic scheduling task information Task, such as Figure 4As shown. Based on the SE model, the production time and logistics time, layout information, logistics path, and delivery method of the equipment are described as static resources in the workshop; the production scheduling plan is to plan and arrange specific tasks of specific production activities for specific production processes or links, including processes, workpiece numbers, processing quantities, task equipment, start time, and end time; through the matching of production scheduling plans and workshop static resources, a process material flow connection subgraph (Material Flow Subgraph, MFS) based on MFMM is formed, as shown below;

[0109] MFS= <MFMM1,MFMM2,...,MFMM n > (5)

[0110] The production process of materials is composed of multiple production processes, that is, a combination of multiple process material flow connection sub-graphs, which ultimately form the production simulation model MFG of the material's entire life cycle.

[0111] (2) Simulation operation scheduling based on material flow metamodel MFMM.

[0112] Design a process interaction algorithm based on MFMM to implement the simulation clock advancement mechanism in DES and the process interaction mechanism based on MFMM, such as Figure 5 shown.

[0113] The simulation life cycle of the flow entity (material) is used as the process; each flow entity has an MFG about the entire production process of the material, and the MFMM model is used as the basic unit of simulation execution; the simulation clock is advanced to the minimum unconditional delayed resurrection time Ta of the material in the Future Event List (FEL), and the MFMM model of all materials in the entire FEL is updated within the Ta time period, and the next Ta is calculated and updated. The simulation is continuously looped until all materials complete the production tasks, forming a segmented simulation algorithm.

[0114] The simulation clock advancement mechanism is as follows Figure 5 (a) is shown. In order to achieve time advancement of complex simulation systems, the management of FEL is crucial. In the study of Bahouth

[20] , when the data structure of FEL is a priority queue, it performs better in terms of time complexity and space complexity. For the process interaction method, the time complexity of the priority queue in obtaining Ta is O(1); in FEL, the MFMM production events Ta of the materials are sorted by size. If Ta is the same, they are sorted according to the order of occurrence and priority of the events.

[0115] In order to refine the clock management of each material, the simulation clock is divided into the simulation master clock SimClock and the entity simulation sub-clock SubClock; SimClock is the main clock of the entire simulation system, responsible for managing the advancement of the global simulation time, controlling the entire simulation process, determining the simulation start time, simulation step step, end time and the time multiplier with the real world, and the SubClock of all materials are synchronized with it; after the process advancement is executed, SubClock calculates the material Ta for each simulation step update.

[0116] For the process advancement mechanism based on MFMM, updating the Ta of the material process is its main purpose, and the update of Ta is achieved through the operation mechanism of MFMM, such as Figure 5 (b) When the simulation clock advances the step length step, all material processes are traversed to execute the MFMM model. By judging the current material position, the MSN model is executed. out Push strategy or DE push strategy, and finally calculate and update Ta and FEL of SimClock.

[0117] MSN out Promoting the process as Figure 6 To simulate the production and processing of materials, i.e. the material conversion process, the delay state of MFMM in the current flow entity is judged to execute α run Or update the delay state; in the MFMM model, the delay state is related to the state of MSN and DE. The idle and service states represent unconditional delays, that is, Ta is a finite time value, and the blocking represents conditional delays, that is, Ta is an infinite value. In the MFMM model, δ in and δ out To change the delay status, the process is as follows:

[0118] 1) If it is a conditional delay, execute MSN out δ out , and reconfigure the delay state.

[0119] 2) If it is an unconditional delay, execute MSN out Alpha run , the remaining time of production activities is calculated through the internal processing time probability function or a deeper process simulation algorithm, and the unconditional delay time of the flow entity is updated.

[0120] 3) Determine whether the current task (such as machining, welding, assembly and other production activities) is completed. If the current task is completed, set the next task as the DE advancement task.

[0121] DE promotion process Figure 6As shown in (b), material unloading, loading and transportation are realized, that is, the material flow process, and the promotion process is as follows:

[0122] 1) Determine the delay status of the current MFMM.

[0123] 2) If it is a conditional delay, execute MSN in δ in , reconfigure the delay state.

[0124] 3) If it is an unconditional delay, execute α of the directed edge DE run , through α run The internal AGV tracking motion algorithm, robotic arm motion control algorithm, etc. calculate the remaining motion time and update the Ta of the flow entity.

[0125] 4) Determine whether the current task (such as handling, transportation and transmission tasks) is completed. If the current task is completed, set the next task to MSN out Advance the task.

[0126] 3. System development and case verification.

[0127] This embodiment builds a material flow modeling and simulation system based on B / S architecture to realize the encapsulation of the proposed method, such as Figure 7 A modeling and simulation experiment was carried out using a certain intelligent manufacturing workshop as an example, and finally the simulation results were compared and analyzed with the commercial software Flexsim to verify the effectiveness of the material flow modeling and simulation of the present invention.

[0128] According to the actual production situation of the manufacturing workshop, the system is used to graphically configure the equipment layout, logistics path and production area of ​​the workshop. In order to facilitate the MFMM display, the graphical modeling process of MFMM is constructed, such as Figure 7 The lower left sub-figure shows the construction process of MSN and DE. MSN is created on the device, and DE is created by connecting MSNs of different devices through directed lines. Figure 7 The logistics equipment binding shown in the lower right small picture eventually forms an MFMM.

[0129] (1) Simulation input configuration.

[0130] The workshop contains a variety of processing equipment, AGV, robotic arms, conveyor belts, stackers, vertical warehouses and other workshop resources. According to the actual production situation of the workshop, the developed system is used to graphically configure the production area, resource layout and logistics path, and the material MFMM model is built based on the workshop production resources. Figure 7 shown.

[0131] The intelligent manufacturing workshop has a total of 4 processes in processing area 2, storage area, laser marking area, and visual inspection area. The production scheduling plan generated by the production planning system (APS, Advanced Planning and Scheduling) is shown in Table 2, and the loading and unloading, logistics time and processing time of the equipment are configured, as shown in Table 3. The materials are set to be shipped out at intervals of 60s, and the rules of the MSN queue are set to first-in-first-out. The production scheduling plan is used as the system input to generate the MFG of logistics, as shown in Figure 8 shown.

[0132] Table 2 Production schedule

[0133]

[0134]

[0135] Table 3 Equipment time parameter table

[0136]

[0137] (2) Analysis of simulation results.

[0138] According to the above simulation input, 20 simulations were performed using the material flow simulation engine and Flexsim to obtain the average utilization of the relevant equipment. The production time obtained by the material flow simulation and Flexsim were 50730s and 50937s respectively. The average utilization of all equipment is shown in Table 4, and the average blocking rate is shown in Table 5.

[0139] Table 4 Equipment utilization comparison table (%)

[0140]

[0141] Table 5 Comparison of equipment blocking rate (%)

[0142]

[0143]

[0144] The relative error of production time, the relative error of average utilization rate and the average blocking rate error are as follows: Fig. 9 As shown. The comparative analysis of simulation results shows that after 20 simulations with Flexsim, the average equipment utilization, blocking rate and production time indicators have errors but are within a reasonable range, with a high degree of fit with actual production and small errors in key indicators. Compared with the modeling flexibility of Flexsim, the MFMM model has more advantages in structure, decision control logic, simulation model generation and operation, and can better adapt to complex situations and changes.

Claims

1. A discrete workshop production simulation method driven by a material flow meta-model, characterized in that: The following steps are involved: Step 1: Material flow modeling based on material flow metamodel MFMM; (1) Definition of Material Flow Metamodel (MFMM); The material flow metamodel MFMM is defined as follows: MFMM= <MSN in ,MSN out ,DE,F> (1) Among them: MSN in It is the input service node of material flow; MSN out It is the output service node of the material flow; DE is the directed edge of the material flow connecting two MSNs; F is the flow entity between the node and the edge, i.e., the material; In the material flow network, the material point of the equipment is defined as MSN, including the docking point of the logistics equipment, the loading and unloading point of the processing equipment, and the service point of the processing equipment. The definition formula is as follows: MSN= <State,DEs,F,T,δ in ,d out ,a run ,P,Q< (2) Where: State represents the state of the node, including waiting, serving, and blocking; DEs is the set of directed edges of material flow, and one MSN can be associated with multiple MSNs; F represents the flow entity in the current node; T is the task execution time of MSN; δ in is the logistics input decision; out is the logistics output decision, α run represents production behavior; P represents the production equipment to which MSN belongs, and Q is the queue at the material point for storing blocked materials; The material logistics process is defined as DE, including handling, conveying and transportation. Different equipment types lead to different control logics, so α is used. run Represents different production behaviors, and the definition formula is: DE= <MSN in ,MSN out ,Direction,E,State,Rule,T,α run > (3) Where: MSN in It is the input service node, MSN out is the output service node, Dirction represents the direction of the edge, including forward, reverse and bidirectional, E is the execution device, State is the state of the directed edge, including waiting, serving and blocking, Rule is the logistics rule, T is the execution time on the edge, α run It is production behavior control; (2) Material flow meta-model MFMM decision and control logic; Logistics input decision function δ in Determine whether the material is transferred from the current material flow metamodel to the next material flow metamodel, which is used for the initialization of the material flow metamodel;δ in It consists of logistics condition judgment and condition change; among them, logistics condition judgment is further divided into basic judgment and custom judgment; basic judgment is mainly based on MSN / DE status, queuing rules and execution conflict judgment; execution conflict is used to analyze whether there will be logistics equipment execution conflict and production equipment execution conflict, and custom judgment allows some judgment conditions to be flexibly set according to specific production needs and scenarios to better adapt to the complex and changeable production environment; according to whether the logistics conditions are met, it can be divided into satisfied condition change and unsatisfied condition change, among which MSN blocking is production equipment execution conflict, and DE blocking is logistics equipment execution conflict; logistics output decision function δ out It determines the flow of materials from DE to MSN out ; Production behavior control function α run It is the behavioral abstraction of the actual production equipment, defining and controlling the production behavior of the basic element MSN / DE of the material flow metamodel, which includes the logic of the production execution and post-execution of the equipment to which MSN / DE belongs. run It refers to the processing, assembly and disassembly activities of production equipment, which is essentially a production time probability function; in DE, α run It is the motion process of the execution equipment, including the transportation process of the conveying equipment, the transportation process of the transportation equipment trajectory planning, and the loading and unloading process of the handling equipment motion control. In essence, it is the motion time function of the execution equipment; the production time probability function is a probability time function formed by adding a certain probability distribution to the execution time of the production task of the processing equipment; The kinematics solution, dynamics solution, trajectory calculation and conveyor belt and AGV trajectory planning of the robot arm produced according to the probability time; the production behavior control output is a random time value t under known conditions; after the production execution is completed, different post-production execution logic is called according to the executed production elements; The logistics decision and production behavior control of the material flow meta-model ultimately change the state of MSN / DE. The internal behavior of MSN / DE and the state transition of external input events are modeled through the finite state machine to describe the state transition conditions of the model during its life cycle; the state of MNS / DE includes waiting, serving and blocking. i is the conversion event of MNS, b i is the conversion event of DE; Step 2: Material flow simulation based on the material flow metamodel MFMM; including production simulation model construction and simulation operation calculation; (1) Simulation model generation; In order to describe various complex simulation input parameters as a unified simulation model, an MFG based on MFMM is defined as follows: MFG=<Sr,Task>=<MFS1,MFS2,...,MFS n > (4) The simulation input parameters include the static resource information Sr of the production workshop and the dynamic scheduling task information Task; based on the SE model, the production time and logistics time, layout information, logistics path, and delivery method of the equipment are described as static resources of the workshop; the production scheduling plan is to plan and arrange specific tasks of specific production activities for specific production processes or links, including processes, workpiece numbers, processing quantities, task equipment, start time, and end time; through the matching of the production scheduling plan and the static resources of the workshop, a process material flow connection subgraph MFS based on MFMM is formed, as shown in the following formula; MFS=<MFMM1,MFMM2,...,MFMM n > (5) The production process of materials is composed of multiple production processes, that is, multiple process material flow connection sub-graphs are combined to finally form the production simulation model MFG of the material's entire life cycle; (2) Simulation operation scheduling based on material flow metamodel MFMM; Design a process interaction algorithm based on MFMM, implement the simulation clock advancement mechanism in DES and the process interaction mechanism based on MFMM; The simulation life cycle of the flow entity is used as a process; each flow entity has an MFG for the entire production process of the material, and the MFMM model is used as a basic unit of simulation execution; the simulation clock is advanced to the minimum unconditional delayed resurrection time Ta of the material in the future event table FEL, and the MFMM model of all materials in the entire FEL is updated within the Ta time period, and the next Ta is calculated and updated. The simulation is continuously looped until all materials complete the production task, forming a segmented simulation algorithm; In order to realize the time advancement of complex simulation systems, the time complexity of the priority queue in obtaining Ta in the process interaction method is O(1); in FEL, the materials are sorted by the size of the MFMM production events Ta. If Ta is the same, they are sorted according to the order of occurrence and priority of the events; In order to refine the clock management of each material, the simulation clock is divided into the simulation master clock SimClock and the entity simulation sub-clock SubClock; SimClock is the main clock of the entire simulation system, responsible for managing the advancement of the global simulation time, controlling the entire simulation process, determining the simulation start time, simulation step step, end time and the time multiplier with the real world, and all material SubClocks are synchronized with it; after the process advancement is completed, SubClock calculates the material Ta updated for each simulation step; For the process advancement mechanism based on MFMM, the main purpose is to update the material process Ta. The update of Ta is realized through the operation mechanism of MFMM. When the simulation clock advances the step length step, all material processes are traversed to execute the MFMM model. By judging the current material position, the MSN is determined. out Push strategy or DE push strategy, finally calculate and update Ta and FEL of SimClock; MSN out Promotion process: In order to simulate the production and processing of materials, that is, the material conversion process, the α run Or update the delay state; in the MFMM model, the delay state is related to the state of MSN and DE. The idle and service states represent unconditional delays, that is, Ta is a finite time value, and the blocking represents conditional delays, that is, Ta is an infinite value. In the MFMM model, δ in and δ out To change the delay status, the process is as follows: 1) If it is a conditional delay, execute MSN out Delta out , and reconfigure the delay state; 2) If it is an unconditional delay, execute MSN out Alpha run , the remaining time of production activities is calculated through the internal processing time probability function or a deeper process simulation algorithm, and the unconditional delay time of the flow entity is updated; 3) Determine whether the current task is completed. If the current task is completed, set the next task as the DE advancement task; DE pushes the process to realize material unloading, loading and transportation, that is, the material flow process. The push process is as follows: 1) Determine the delay status of the current MFMM; 2) If it is a conditional delay, execute MSN in Delta in , reconfigure the delay state; 3) If it is an unconditional delay, execute α of the directed edge DE run , through α run The internal AGV tracking motion algorithm and the robot arm motion control algorithm calculate the remaining motion time and update the Ta of the flow entity; 4) Determine whether the current task is completed. If the current task is completed, set the next task to MSN out Advance the task.

Images