Digital twinning manufacturing system

Abstract

The invention provides a manufacturing digital twin system, which is characterized in that a production process simulation module simulates different production scenes of a manufacturing process and production flow data under each production scene, and generates a simulation report; the workshop scheduling engine optimizes production task scheduling and resource allocation of a production workshop by using an operation planning optimization algorithm to generate a workshop task production scheduling scheme; obtaining a production scheduling result corresponding to the workshop task production scheduling scheme; a weekly production plan production scheduling module generates a weekly production plan according to the production scene provided by the production process simulation module and a production scheduling result corresponding to each workshop task production scheduling scheme; based on the enterprise resource state obtained in real time, dynamically adjusting the weekly production plan, and feeding back the dynamically adjusted weekly production plan to a production process simulation module and a workshop scheduling engine in real time, so that the production process simulation module performs simulation verification on the weekly production plan; and the workshop scheduling engine updates the workshop task production scheduling scheme according to the new weekly production plan.

Description

Technical Field

[0001] The present invention relates to the field of intelligent manufacturing technology, and in particular to a manufacturing digital twin system. Background Art

[0002] The existing manufacturing industry faces challenges such as personalized customer demands, short delivery cycles, high quality requirements, and significant cost pressures. Traditional production management models struggle to cope with complex production environments and volatile market demands, resulting in low production efficiency and significant resource waste. Existing production planning and scheduling systems (such as APS) and simulation tools (such as Siemens' PlantSimulation) face challenges in the construction machinery industry, including poor adaptability, complex operation, and limited integration capabilities. Summary of the Invention

[0003] The present invention provides a manufacturing digital twin system to address the shortcomings of existing engineering machinery manufacturing systems in the technology, such as insufficient adaptability, complex operation, high training costs, and limited integration capabilities with existing enterprise systems, and to achieve dynamic resource allocation and real-time data feedback.

[0004] The present invention provides a manufacturing digital twin system, comprising: A production process simulation module is configured to simulate different production scenarios of the production process and the production process data under each of the production scenarios, simulate the production process under each of the production scenarios based on the production process data, and generate a simulation report; The workshop scheduling engine is configured to optimize the production task scheduling and resource allocation of the production workshop using an operations optimization algorithm based on the production scenarios provided by the production process simulation module, the production process simulation results under each of the production scenarios, the real-time production environment of the workshop, and the received weekly production plan, and generate a workshop task scheduling plan; and feed the workshop task scheduling plan back to the production process simulation module in real time, so that the production process simulation module simulates and verifies the workshop task scheduling plan and obtains a scheduling result corresponding to the workshop task scheduling plan; The weekly production plan scheduling module is configured to generate a weekly production plan based on the production scenarios provided by the production process simulation module and the scheduling results corresponding to each of the workshop task scheduling plans; dynamically adjust the weekly production plan based on the real-time acquired enterprise resource status, and feed back the dynamically adjusted weekly production plan to the production process simulation module and the workshop scheduling engine in real time, so that the production process simulation module simulates and verifies the weekly production plan, and the workshop scheduling engine updates the workshop task scheduling plan according to the new weekly production plan.

[0005] According to the manufacturing digital twin system provided by the present invention, the manufacturing process in each of the production scenarios is simulated based on the production process data, and a simulation report is generated, specifically including: Based on discrete event simulation theory, simulate random events in the production process, including one or more of material flow, equipment status and worker operation; Predicting changes in the state of production factors caused by the random events at each discrete time point, where the production factors include one or more of production line layout, equipment configuration, logistics path, and personnel allocation; The prediction results of the random events and the changes in the status of the production factors are used as production process data to simulate the production process.

[0006] According to the manufacturing digital twin system provided by the present invention, after generating a production scheduling plan for shop floor tasks, the shop floor scheduling engine is further configured to: Based on the changes in production resources in the production workshop, the production task sequencing and resource allocation plan in the production workshop are dynamically adjusted to generate a new workshop task scheduling plan.

[0007] According to the manufacturing digital twin system provided by the present invention, the real-time production environment of the workshop includes entity attributes, process logic, and uncertainty factors; Based on the production scenarios provided by the production process simulation module, the production process simulation results under each production scenario, the real-time production environment of the workshop, and the received weekly production plan, the production task scheduling and resource allocation of the production workshop are optimized using an operations optimization algorithm, specifically including: Based on the real-time production environment of the workshop and the received weekly production plan, perform entity modeling, process modeling and uncertainty modeling on the production task scheduling and resource allocation of the workshop task scheduling plan to obtain the entity model, process model and uncertainty model for the production scheduling plan; Select the appropriate operations research optimization algorithm based on the scale, structure, optimization objectives, and complexity of constraints of the shop floor scheduling problem; Based on the production scenarios provided by the production process simulation module, the production process simulation results under each of the production scenarios, and the pre-set optimization goals and constraints, the operations optimization algorithm is used to solve the entity model, process model, and uncertainty model.

[0008] According to the manufacturing digital twin system provided by the present invention, the entity modeling includes: machine equipment modeling, material handling equipment modeling and workpiece modeling; The machine equipment modeling is used to model the processing capacity, time distribution, failure mode and probability of the machine equipment in the workshop to simulate the processing capacity and potential failure risks of the machine equipment; The material handling equipment modeling is used to model the speed, load capacity, and path planning rules of the material handling equipment to simulate the movement of the material handling equipment in the workshop and the impact of the handling tasks on the production process; The workpiece modeling is used to model the process route, process and processing time of each workpiece to simulate the impact of different processing paths and time requirements of each workpiece on the production process.

[0009] According to the manufacturing digital twin system provided by the present invention, the process modeling includes: production process modeling and constraint condition modeling; The production process modeling is used to construct a production process model, which is used to describe the flow process of workpieces in the workshop, including the connection sequence of processes, the location of the buffer area, the transportation path, and the collaborative relationship between equipment; The constraint modeling is used to add equipment availability, material supply conditions, and buffer area quantity restrictions to the production process model to ensure that the production process model can reflect various restrictions in actual production.

[0010] According to the manufacturing digital twin system provided by the present invention, the uncertainty factors include machine failure, raw material fluctuations, and order changes; The uncertainty modeling specifically includes: A probability model is established for uncertainty factors, and corresponding distribution functions are set to simulate the occurrence patterns of uncertainty factors in order to evaluate the performance of production scheduling plans under different scenarios.

[0011] According to the manufacturing digital twin system provided by the present invention, the operations optimization algorithm for selecting a production scheduling solution based on the scale, structure, optimization objectives, and complexity of constraints of the shop floor scheduling problem specifically includes: Select the category of initial algorithm based on the scale, structure, optimization objectives and complexity of constraints of the shop scheduling problem; Optimize the parameters of the selected initial algorithm, including parameter setting, initial solution generation, and neighborhood structure design; The effectiveness of the initial algorithm is verified through experiments, the performance of different initial algorithms on the same problem is compared, and the optimal initial algorithm is selected as the operations optimization algorithm for the production scheduling plan.

[0012] According to the manufacturing digital twin system provided by the present invention, the category of the initial algorithm is selected based on the scale, structure, optimization objectives and complexity of the constraints of the shop scheduling problem, specifically including: If the size of the shop scheduling problem is smaller than the preset size, the exact algorithm is used; If the scale of the shop scheduling problem is larger than the predetermined scale, a heuristic algorithm or a metaheuristic algorithm is used; If the objective function and constraints of the shop scheduling problem are linear, use the linear programming algorithm or the integer linear programming algorithm; If the objective function or constraints of the shop scheduling problem are nonlinear, a nonlinear programming algorithm or a mixed integer nonlinear programming algorithm is used; If the shop scheduling problem involves discrete decision variables, integer programming or mixed integer programming algorithms are used; If the shop scheduling problem involves only one optimization objective, a single-objective optimization algorithm is used; If there are multiple optimization objectives in the shop scheduling problem, a multi-objective optimization algorithm is used; If the constraints of the shop scheduling problem are simple constraints, a linear programming algorithm or an integer programming algorithm is used, wherein the simple constraints include a single entity constraint or a combination of multiple single entity constraints; If the constraints of the shop scheduling problem are complex constraints, a heuristic algorithm or a meta-heuristic algorithm is used, wherein the complex constraints include nonlinear constraints, dynamic constraints, and random constraints; If there are soft constraints and hard constraints in the shop scheduling problem, the constraint relaxation or penalty function method is used to transform the soft constraints into part of the objective function, where the soft constraints are constraints that can be partially violated and the hard constraints are constraints that must be strictly satisfied.

[0013] According to the manufacturing digital twin system provided by the present invention, the enterprise resource status includes: market demand data, production resource data and production status data, wherein the market demand data includes order quantity and delivery time, order priority, and market demand changes; the production resource data includes equipment availability, manpower availability and material supply; the production status data includes the complexity of production tasks, equipment maintenance plans and production bottlenecks.

[0014] According to the manufacturing digital twin system provided by the present invention, the production process simulation module is used to simulate the connection and interaction between production factors in the manufacturing entity. The production process simulation module adopts a discrete event system modeling tool and supports drag-and-drop simulation of the behavior of each production factor.

[0015] According to the manufacturing digital twin system provided by the present invention, the production process simulation module adopts a hierarchical modeling concept and supports the encapsulation of any custom atomic event object and the coupling of complex event objects.

[0016] According to the manufacturing digital twin system provided by the present invention, the engine packaging of the production process simulation module adopts a B / S architecture and is based on cloud native technology to support online modeling, simulation and optimization.

[0017] The manufacturing digital twin system provided by the present invention simulates different production scenarios of the production and manufacturing process and the production process data under each of the production scenarios through a production process simulation module, simulates the production and manufacturing process under each of the production scenarios based on the production process data, and generates a simulation report; the workshop scheduling engine optimizes the production task scheduling and resource allocation of the production workshop using an operations optimization algorithm based on the production scenarios provided by the production process simulation module, the production and manufacturing process simulation results under each of the production scenarios, the real-time production environment of the workshop and the received weekly production plan, and generates a workshop task scheduling plan; the workshop task scheduling plan is fed back to the production process simulation module in real time to enable the The production process simulation module simulates and verifies the workshop task scheduling plan to obtain the scheduling results corresponding to the workshop task scheduling plan; the weekly production plan scheduling module generates a weekly production plan based on the production scenario provided by the production process simulation module and the scheduling results corresponding to each workshop task scheduling plan; based on the real-time acquired enterprise resource status, the weekly production plan is dynamically adjusted, and the dynamically adjusted weekly production plan is fed back to the production process simulation module and the workshop scheduling engine in real time, so that the production process simulation module simulates and verifies the weekly production plan, and the workshop scheduling engine updates the workshop task scheduling plan according to the new weekly production plan. The present invention is based on discrete event simulation theory and is used to simulate the connection between production factors in manufacturing entities, providing a global factory perspective for cross-departmental communication. Machine learning and operations optimization technologies are used to discover patterns and trends from data, automatically find the optimal solution, and provide operators with feedback and control solutions for manufacturing entities. The production process simulation module simulates the actual production process in a virtual environment, providing an experimental and optimization platform for workshop scheduling such as process departments and weekly production planning scheduling such as operations departments. The process department uses these simulation data to test and improve the production layout and optimize the product manufacturing process, while the operations department guides actual production activities based on the simulation results, realizing continuous interaction and iterative evolution between physical and digital entities. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0019] Figure 1 This is one of the structural schematic diagrams of the manufacturing digital twin system provided by the present invention.

[0020] Figure 2 This is the second structural schematic diagram of the manufacturing digital twin system provided by the present invention.

[0021] Figure 3 This is the business architecture of the manufacturing digital twin system provided by the present invention.

[0022] Figure 4 This is the technical architecture of the manufacturing digital twin system provided by the present invention.

[0023] Figure 5 It is a flow chart of the simulation modeling of the fuel tank production process in the first embodiment provided by the present invention.

[0024] Figure 6 This is the problem-solving strategy adopted on-site in Example 1 provided by the present invention.

[0025] Figure 7 This is the algorithm framework used in the second embodiment provided by the present invention.

[0026] Figure 8 This is a comparison of the effectiveness of different algorithms and manual experience in Example 2 provided by the present invention.

[0027] Figure 9 It is the selectable constraint rule pool established in the third embodiment provided by the present invention.

[0028] Figure 10 This is the specific production scheduling result obtained in Example 3 provided by the present invention. DETAILED DESCRIPTION

[0029] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0030] The present invention is described in detail below with reference to the accompanying drawings. The specific operating methods in the method embodiments can also be applied to device embodiments or system embodiments. In the description of the present invention, unless otherwise specified, "at least one" includes one or more. "Multiple" refers to two or more. For example, at least one of A, B, and C includes: A exists alone, B exists alone, A and B exist at the same time, A and C exist at the same time, B and C exist at the same time, and A, B, and C exist at the same time. In the present invention, " / " means or, for example, A / B can mean A or B; "and / or" in this article is only a description of the association relationship of associated objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone.

[0031] The existing construction machinery industry faces a rapidly changing market environment, with customers demanding personalized products with shorter delivery cycles, higher quality, and lower costs. This requires companies to be highly flexible and able to identify optimal decisions based on numerous influencing factors. However, this complex global observation and long-term reasoning process exceeds the natural capabilities of the human brain. Against this backdrop, the manufacturing digital twin system has emerged. Through the interaction between virtual digital models and the physical production environment, it provides real-time feedback and iterative updates, offering real-world solutions to address business pain points.

[0032] Although every department in the manufacturing industry is working diligently to improve their own performance indicators, many of these targets conflict with one another. For example, inventory turnover and on-time delivery are in conflict, as are cost control and quality assurance. When all departments operate independently, it's difficult to achieve a holistic perspective and improve the company's overall performance.

[0033] Furthermore, to accommodate the high-variety, low-batch production model, production lines typically employ mixed-line production. Mixed-line products have uneven production cycles. While it's theoretically possible to achieve balanced production by combining different products, most companies lack effective tools to achieve this and typically rely on the manual experience of experienced technicians to schedule production. Traditional process informatization, on the other hand, simply embeds human experience into information systems. This leads to two major problems: first, there's no scientific evaluation method to validate the effectiveness of human experience; second, these copycat information systems are far less effective on the front lines than human experience, leading to resistance from manufacturing departments. Therefore, manufacturing companies need to provide assessment tools and methods that enable people across all departments to quickly identify the root causes of problems from a holistic perspective.

[0034] In view of this, an embodiment of the present invention provides a manufacturing digital twin system, which is implemented based on simulation and operations optimization, and its core purpose is to achieve continuous interaction and iterative evolution between physical entities and digital entities.

[0035] The present invention will be described in detail below with reference to specific embodiments.

[0036] In some specific embodiments of the present invention, Figure 1 As shown, this solution provides a manufacturing digital twin system, including: The production process simulation module 11 is configured to simulate different production scenarios of the production process and the production process data under each of the production scenarios, simulate the production process under each of the production scenarios based on the production process data, and generate a simulation report; The shop scheduling engine 12 is configured to optimize the production task scheduling and resource allocation of the production workshop using an operations optimization algorithm based on the production scenarios provided by the production process simulation module, the production process simulation results under each of the production scenarios, the real-time production environment of the workshop, and the received weekly production plan, and generate a shop task scheduling plan; and feed the shop task scheduling plan back to the production process simulation module in real time, so that the production process simulation module simulates and verifies the shop task scheduling plan and obtains a scheduling result corresponding to the shop task scheduling plan; The weekly production plan scheduling module 13 is configured to generate a weekly production plan based on the production scenario provided by the production process simulation module and the scheduling results corresponding to each of the workshop task scheduling plans; dynamically adjust the weekly production plan based on the real-time acquired enterprise resource status, and feed back the dynamically adjusted weekly production plan to the production process simulation module and the workshop scheduling engine in real time, so that the production process simulation module simulates and verifies the weekly production plan, and the workshop scheduling engine updates the workshop task scheduling plan according to the new weekly production plan.

[0037] It should be noted that existing engineering machinery manufacturing solutions lack specific simulation support for the engineering machinery industry, are complex to operate and have high training costs, and have limited integration capabilities with the internal systems of engineering machinery manufacturing companies such as IoT, MES, ERP, etc., and cannot meet user needs for personalization, short delivery cycles, high quality, and low costs.

[0038] Therefore, the present invention sets a production process simulation module to simulate different production scenarios of the production and manufacturing process and the production process data under each of the production scenarios, sets a workshop scheduling engine, and optimizes the production task scheduling and resource allocation of the production workshop according to the production scenarios provided by the production process simulation module, the production and manufacturing process simulation results under each of the production scenarios, the real-time production environment of the workshop and the received weekly production plan using an operations optimization algorithm to generate a workshop task scheduling plan; sets a weekly production plan scheduling module to generate a weekly production plan according to the production scenarios provided by the production process simulation module and the scheduling results corresponding to each of the workshop task scheduling plans; dynamically adjusts the weekly production plan based on the enterprise resource status obtained in real time, and then feeds back the dynamically adjusted weekly production plan to the production process simulation module and the workshop scheduling engine in real time. The manufacturing digital twin system based on simulation and operations optimization provided by the present invention provides real-time feedback and iterative updates through the interaction between virtual digital models and physical production environments.

[0039] In some possible implementations of the present invention, simulating the production process in each of the production scenarios based on the production process data and generating a simulation report specifically includes: Based on discrete event simulation theory, random events in the manufacturing process are simulated, including one or more of material flow, equipment status, and worker operation; Predicting changes in the state of production factors caused by the random events at each discrete time point, where the production factors include one or more of production line layout, equipment configuration, logistics path, and personnel allocation; The prediction results of the random events and the changes in the status of the production factors are used as production process data to simulate the production process.

[0040] Specifically, this embodiment provides an implementation method for simulating the manufacturing process under various production scenarios. It utilizes discrete event simulation (DES) theory as a foundation. This theory treats events in the production process (such as material arrival, equipment failure, and task completion) as discrete time points and simulates the occurrence and processing of these events to recreate the dynamic behavior of the production system. Through this setup, the simulation module can highly recreate the actual production environment, including details such as equipment operation, personnel operations, material flow, and logistics distribution. By simulating the operation of different production scenarios, it verifies whether they meet order requirements, resource constraints, and quality requirements. Through simulation analysis, it quickly identifies bottlenecks in the production process (such as equipment failure, insufficient personnel, and material shortages), enabling detailed simulation and verification of the production process.

[0041] Furthermore, by simulating different optimization schemes (such as equipment layout adjustment, task sequencing optimization, resource allocation schemes, etc.) and comparing their performance indicators (such as production cycle, equipment utilization, cost, etc.), the system provides production managers with intuitive simulation results (such as Gantt charts, progress charts, animation displays, etc.), helping them to quickly evaluate the feasibility of the optimization scheme and support scientific decision-making.

[0042] Furthermore, the simulation module provides real-time feedback on dynamic changes in the production process (such as equipment failures, order changes, material shortages, etc.), providing data support for the dynamic adjustment of production plans; by simulating the adjusted production plan, it verifies whether it can effectively respond to changes in market demand and resource fluctuations, ensuring the feasibility of the adjustment plan.

[0043] In a possible implementation, random events such as material arrival, equipment failure, and worker status changes frequently occur during the production process, causing changes in state variables at discrete time points. To address the discreteness, uncertainty, dynamics, and computational complexity of production factor status changes, a production process simulation engine is provided, including modules for material generation, queuing, global scheduling, inventory, processing, and resource operations, to accurately simulate production and generate analytical reports.

[0044] In a possible embodiment, the production process simulation module is used to simulate the connection and interaction between production factors in a manufacturing entity, using a discrete event system modeling tool to support drag-and-drop simulation of the behavior of each of the production factors.

[0045] In a possible embodiment, the core engine of the production process simulation module is based on DEVS (Discrete Event System Specification) theory, adopts an object-oriented approach to build a model, and supports encapsulation, separation, and inheritance.

[0046] Specifically, based on DEVS theory, the simulation module supports hierarchical modeling. Users can decompose complex production systems into multiple subsystems (such as workstations, production lines, and workshops) and combine these subsystems into a complete production model through encapsulation technology.

[0047] In a possible embodiment, the production process simulation module adopts a hierarchical modeling concept and supports the encapsulation of any custom atomic event object and the coupling of complex event objects.

[0048] Specifically, the production process simulation module supports the encapsulation and reuse of atomic models, allowing users to define reusable production units (such as standard workstations and equipment templates) and quickly build new production scenarios through an inheritance mechanism.

[0049] In a possible embodiment, the engine is encapsulated as a dynamic link library to facilitate the construction of software of different architectures. Furthermore, the engine packaging of the production process simulation module adopts B / S architecture and is based on cloud-native technology, supporting online modeling, simulation and optimization.

[0050] It is worth noting that the engine of the production process simulation module is encapsulated on the basis of the B / S architecture and can be gradually developed and migrated to the desktop or mobile terminal.

[0051] Specifically, the engine of the production process simulation module adopts a B / S architecture based on cloud-native technology, allowing users to access simulation services over the internet. Network communication is primarily accomplished through HTTP and WebSocket protocols. The HTTP protocol is responsible for model parameter delivery, management, authentication, and report data transmission, while the WebSocket protocol is used for animation control, state transmission, and message exchange. In terms of technology selection, Java is used on the backend, the Vue framework is used on the front end, and the core computing engine uses C / C++ for precise memory and CPU / GPU scheduling. The Redis cluster provides high-speed caching for backend services. Simulation accuracy supports milliseconds, meeting animation requirements, and can even be extended to nanoseconds.

[0052] The production process simulation module utilizes a high-performance computing engine developed in C / C++, supporting millisecond and even nanosecond simulation accuracy, enabling precise simulation of time series and state changes during the production process. A three-stage parallel simulation approach (pre-scan, massively parallel execution, and conditional event processing) fully utilizes multi-core CPU or GPU resources to improve simulation efficiency. The production process simulation module interacts with the front-end via HTTP and WebSocket protocols, supporting online model execution, parameter adjustment, animation display, and result feedback.

[0053] Through the above settings, the production process simulation module provides online model operation, supports employees to search, run, share and discuss models online, and promotes cooperation and knowledge sharing; standardizes equipment parameters through a unified resource library, reduces modeling time, and improves the uniformity and accuracy of resource utilization between models; and can support multiple queuing methods and priority scheduling of cross-line transportation tasks; can complete the drawing of finite state machine transfer matrices of seven major categories of elements, and handle thousands of state and combination state transitions.

[0054] In a possible implementation, the simulation module provides an intuitive display of the production process, useful for training new employees and helping them quickly familiarize themselves with production processes and operating procedures. It also enables employees to search, run, share, and discuss simulation models online, promoting knowledge sharing and experience accumulation within the company. The simulation module also provides a holistic view of the factory floor, helping managers understand the production process holistically and breaking down information barriers between departments. Working in conjunction with operations research optimization algorithms, it provides optimization support for production planning, shop floor scheduling, equipment maintenance, and other aspects, ultimately achieving global optimization of the production system.

[0055] Furthermore, the production process simulation module is not only used to verify the feasibility of existing production plans, but also to provide optimization direction and verification environment for operations optimization algorithms by simulating different optimization plans (such as resource allocation, task sorting, equipment layout adjustment, etc.); the simulation results (such as production cycle, equipment utilization, number of work-in-progress, etc.) are used as feedback to provide a basis for the dynamic adjustment of production plans.

[0056] In some possible implementations of the present invention, after generating a production scheduling plan for the shop floor tasks, the shop floor scheduling engine is further configured to: Based on the changes in production resources in the production workshop, the production task sequencing and resource allocation plan in the production workshop are dynamically adjusted to generate a new workshop task scheduling plan.

[0057] Specifically, this embodiment provides an implementation method for generating a new workshop task scheduling plan by obtaining changes in production resources in the production workshop and dynamically adjusting the production task sorting and resource allocation plan in the production workshop based on the changes, thereby forming a new workshop task scheduling plan.

[0058] Specifically, the workshop scheduling engine consists of two core parts: a digital simulator and an operations optimization algorithm, which are used to optimize production task scheduling and resource allocation at the workshop level.

[0059] In some possible implementations of the present invention, the real-time production environment of the workshop includes entity attributes, process logic, and uncertainty factors; Based on the production scenarios provided by the production process simulation module, the production process simulation results under each production scenario, the real-time production environment of the workshop, and the received weekly production plan, the production task scheduling and resource allocation of the production workshop are optimized using an operations optimization algorithm, specifically including: Based on the real-time production environment of the workshop and the received weekly production plan, perform entity modeling, process modeling and uncertainty modeling on the production task scheduling and resource allocation of the workshop task scheduling plan to obtain the entity model, process model and uncertainty model for the production scheduling plan; Select the appropriate operations research optimization algorithm based on the scale, structure, optimization objectives, and complexity of constraints of the shop floor scheduling problem; Based on the production scenarios provided by the production process simulation module, the production process simulation results under each of the production scenarios, and the pre-set optimization goals and constraints, the operations optimization algorithm is used to solve the entity model, process model, and uncertainty model.

[0060] Specifically, the shop floor scheduling engine includes entity modeling, process modeling, and uncertainty modeling.

[0061] In some possible implementations of the present invention, the entity modeling includes: machine equipment modeling, material handling equipment modeling, and workpiece modeling; The machine equipment modeling is used to model the processing capacity, time distribution, failure mode and probability of the machine equipment in the workshop to simulate the processing capacity and potential failure risks of the machine equipment; The material handling equipment modeling is used to model the speed, load capacity, and path planning rules of the material handling equipment to simulate the movement of the material handling equipment in the workshop and the impact of the handling tasks on the production process; The workpiece modeling is used to model the process route, process and processing time of each workpiece to simulate the impact of different processing paths and time requirements of each workpiece on the production process.

[0062] Specifically, entity modeling includes equipment modeling: modeling the machinery and equipment in the workshop, including parameters such as their processing capabilities, failure modes, maintenance cycles, and availability; material handling equipment modeling: modeling material handling equipment such as AGVs (automatic guided vehicles) and forklifts, and clarifying their speed, load capacity, path planning rules, etc.; workpiece modeling: modeling production tasks (such as workpieces) and defining their process routes, procedures, processing time, priority, etc.

[0063] In some possible implementations of the present invention, the process modeling includes: production process modeling and constraint condition modeling; The production process modeling is used to construct a production process model, which is used to describe the flow process of workpieces in the workshop, including the connection sequence of processes, the location of the buffer area, the transportation path, and the collaborative relationship between equipment; The constraint modeling is used to add equipment availability, material supply conditions, and buffer area quantity restrictions to the production process model to ensure that the production process model can reflect various restrictions in actual production.

[0064] Specifically, process modeling includes production process definition: building a production process model to describe the flow process of workpieces in the workshop, including process connection, buffer area location, transportation path, equipment collaboration, etc.; constraint setting: considering equipment availability, material supply, buffer area quantity limit, process sequence and other constraints to ensure the rationality of the production process.

[0065] In some possible implementations of the present invention, the uncertainty factors include machine failure, raw material fluctuations, and order changes; The uncertainty modeling specifically includes: A probability model is established for uncertainty factors, and corresponding distribution functions are set to simulate the occurrence patterns of uncertainty factors in order to evaluate the performance of production scheduling plans under different scenarios.

[0066] Specifically, uncertainty modeling includes probability models: for uncertain factors such as machine failures, raw material fluctuations, and order changes, probability models are established, and appropriate distribution functions (such as Poisson distribution and normal distribution) are set to simulate their occurrence patterns; multi-scenario evaluation: by simulating production operations under different scenarios, the performance of the scheduling plan in various situations is evaluated to improve the robustness of the scheduling plan.

[0067] In some possible implementations of the present invention, the operations research optimization algorithm for selecting a production scheduling solution based on the scale, structure, optimization objectives, and complexity of constraints of the shop scheduling problem specifically includes: Select the category of initial algorithm based on the scale, structure, optimization objectives and complexity of constraints of the shop scheduling problem; Optimize the parameters of the selected initial algorithm, including parameter setting, initial solution generation, and neighborhood structure design; The effectiveness of the initial algorithm is verified through experiments, the performance of different initial algorithms on the same problem is compared, and the optimal initial algorithm is selected as the operations optimization algorithm for the production scheduling plan.

[0068] Specifically, this embodiment provides an implementation method for an operations optimization algorithm for selecting a production scheduling plan. By selecting the category of the initial algorithm, tuning the parameters of the initial algorithm, and then verifying the effectiveness of the initial algorithm through experiments, the optimal initial algorithm is obtained as the operations optimization algorithm for the production scheduling plan.

[0069] Specifically, the algorithm selection basis can be based on the problem scale, structure, optimization objectives and constraint complexity to select appropriate operations optimization algorithms, such as traditional algorithms and heuristic and meta-heuristic algorithms. Objective function construction can be achieved by clarifying the optimization objectives and converting them into mathematical functions, such as minimizing the production cycle or cost, and considering various cost factors to construct a cost function. Constraint processing, by identifying constraints such as processing technology, resources, and time, ensures that the processing sequence meets the requirements, considers equipment and manpower time constraints, and handles delivery dates, maintenance time windows, etc. In the solution process, an algorithm is used to solve the problem. Specifically, it can be used through iterative search to find the optimal solution or approximate solution, and adopt local and global search strategies to avoid falling into local optimality. Parameter optimization, that is, optimizing algorithm parameters, can be specifically determined through experimental design and analysis techniques to improve algorithm performance.

[0070] In a possible embodiment, a suitable operations research optimization algorithm, such as a genetic algorithm, a simulated annealing algorithm, a hybrid algorithm, etc., may be selected according to the problem scale, structure, optimization goal, and constraint complexity.

[0071] Furthermore, the optimization objectives (such as minimizing production cycle, maximizing equipment utilization, minimizing cost, etc.) are converted into mathematical functions, and the cost function is constructed by considering various cost factors.

[0072] Then, mathematical models are used to process constraints such as process, resources, and time to ensure that the processing sequence meets the requirements, consider the time constraints of equipment and manpower, and handle delivery dates, maintenance time windows, etc.

[0073] On this basis, we use the selected optimization algorithm to find the optimal solution or a near-optimal solution through iterative search. We employ local and global search strategies to avoid being trapped in local optima. Through experimental design and analysis techniques, we optimize the algorithm parameters, determine the optimal value range, and improve the algorithm performance.

[0074] In some possible implementations of the present invention, the type of initial algorithm is selected based on the scale, structure, optimization objectives, and complexity of constraints of the shop scheduling problem, specifically including: If the size of the shop scheduling problem is smaller than the preset size, the exact algorithm is used; If the scale of the shop scheduling problem is larger than the predetermined scale, a heuristic algorithm or a metaheuristic algorithm is used; If the objective function and constraints of the shop scheduling problem are linear, use the linear programming algorithm or the integer linear programming algorithm; If the objective function or constraints of the shop scheduling problem are nonlinear, a nonlinear programming algorithm or a mixed integer nonlinear programming algorithm is used; If the shop scheduling problem involves discrete decision variables, integer programming or mixed integer programming algorithms are used; If the shop scheduling problem involves only one optimization objective, a single-objective optimization algorithm is used; If there are multiple optimization objectives in the shop scheduling problem, a multi-objective optimization algorithm is used; If the constraints of the shop scheduling problem are simple constraints, a linear programming algorithm or an integer programming algorithm is used, wherein the simple constraints include a single entity constraint or a combination of multiple single entity constraints; If the constraints of the shop scheduling problem are complex constraints, a heuristic algorithm or a meta-heuristic algorithm is used, wherein the complex constraints include nonlinear constraints, dynamic constraints, and random constraints; If there are soft constraints and hard constraints in the shop scheduling problem, the constraint relaxation or penalty function method is used to transform the soft constraints into part of the objective function, where the soft constraints are constraints that can be partially violated and the hard constraints are constraints that must be strictly satisfied.

[0075] In a possible embodiment, for dynamic adjustment and feedback of the shop scheduling engine, the shop scheduling engine can be integrated with systems such as MES and ERP to obtain production status data (such as equipment failures, order changes, material shortages, etc.) in real time, thereby achieving real-time data integration; the scheduling plan can be dynamically adjusted based on real-time data to ensure the smooth execution of production tasks and achieve dynamic adjustment; by feeding the optimization results back to the simulation module for verification, a closed-loop optimization process is formed, the scheduling plan is continuously improved, and an effective feedback mechanism is implemented.

[0076] The embodiment of the present invention optimizes production task scheduling by setting up a shop scheduling engine and optimizing task sequencing and resource allocation.

[0077] Specifically, task sequencing can be optimized based on factors such as workpiece priority, delivery date, and processing time, reducing production cycle time and waiting time. By rationally allocating equipment, manpower, and material resources, resource utilization can be improved, equipment idleness or overload can be avoided, and resource allocation can be optimized.

[0078] Improve production efficiency by reducing waiting time and increasing equipment utilization.

[0079] Specifically, by optimizing task sequencing and resource allocation, the waiting time of workpieces in the production process can be reduced and production efficiency can be improved; by rationally arranging the use of equipment, production interruptions caused by equipment idleness or failure can be avoided and the overall utilization of equipment can be improved.

[0080] Effectively deal with uncertainty through dynamic adjustment capabilities and multi-scenario assessment.

[0081] Specifically, the workshop scheduling engine can respond to uncertainties in the production process (such as equipment failure, order changes, material shortages, etc.) in real time, dynamically adjust the scheduling plan, and ensure the continuity of production; by simulating production operations under different scenarios, evaluate the robustness of the scheduling plan, and improve the ability to respond to emergencies.

[0082] Through scientific decision support and collaborative optimization, decision support and collaboration are achieved.

[0083] Specifically, the scheduling plan generated by the optimization algorithm provides scientific decision-making support for production managers and reduces the uncertainty of human experience; by working in collaboration with the simulation module and the weekly production planning and scheduling module, the global optimization of the production system is achieved.

[0084] Improve production transparency through visual display and real-time monitoring and adjustment.

[0085] Specifically, the scheduling results can be displayed intuitively through Gantt charts, progress charts, etc., to help managers monitor production progress in real time and identify potential problems in a timely manner; at the same time, managers can quickly adjust the scheduling plan based on real-time monitoring data to ensure the smooth execution of production tasks.

[0086] In some possible implementations of the present invention, the enterprise resource status includes: market demand data, production resource data and production status data, wherein the market demand data includes order quantity and delivery time, order priority, and market demand changes; the production resource data includes equipment availability, manpower availability and material supply; and the production status data includes the complexity of production tasks, equipment maintenance plans and production bottlenecks.

[0087] Specifically, the weekly production planning and scheduling module uses operations optimization and big data analysis technology to achieve dynamic resource allocation and real-time data feedback, ensuring tasks are completed on time and improving capacity utilization. The weekly production planning and scheduling module can achieve the following functions: Intelligent production scheduling: Specifically, technologies such as simulated annealing and genetic algorithms can be used to automatically consider order priorities and resource fulfillment rates to generate the optimal plan; Real-time adjustment and dynamic scheduling: Monitor production variables, such as line changes, and adjust production schedules instantly to maintain flexibility and responsiveness; Production forecasting and simulation: Based on historical data, predict future trends, conduct simulation tests, and evaluate the availability of production scheduling strategies; Visual management: The intuitive interface displays production progress and resource utilization, supporting real-time monitoring and adjustment; Data-driven optimization: Collect and analyze data, continuously optimize production scheduling strategies, form a closed loop, and improve efficiency.

[0088] The manufacturing digital twin system provided by the embodiment of the present invention has a lightweight production process simulation module. Through a highly flexible discrete event simulation framework, it adopts the most advanced DEVS theoretical framework, truly restores the hierarchical modeling idea, supports the encapsulation of any custom atomic event object and the coupling of event objects of any level of complexity, and effectively realizes the inheritance and separation of system structure and behavior; adopts high-speed parallel simulation advancement technology and the most advanced three-stage parallel simulation method to pre-scan the occurrence time of unconditional events and form a unified time advancement plan; executes unconditional events in parallel on a large scale; and executes conditionally dependent events through multiple processes, thereby making full use of multi-core CPU or GPU resources and maximizing the operating efficiency of the simulation system.

[0089] Furthermore, in shop-floor scheduling engines, accurately describing entity attributes and process logic is crucial. Machines and equipment require detailed descriptions of their processing capabilities, time distribution, failure modes, and probabilities; material handling equipment, such as automated guided vehicles (AGVs), requires clear speeds, load capacities, and path planning rules; and workpieces require defined process routes and processing parameters. The production process model must be accurately constructed, including the sequence of process connections, precedence constraints, parallel process processing methods, and material handling path coordination. Constraints such as equipment availability, material supply, and buffer size limitations must also be considered. The integration and innovation of intelligent optimization algorithms are key to improving performance. For example, hybrid algorithms can be developed by combining the strengths of artificial bee colonies, genetic algorithms, and variable neighborhood algorithms. Innovative improvements can be made to traditional algorithms, such as by improving the encoding of genetic algorithms and designing new algorithm structures based on a two-layer structure of processes and machines. Adaptive search strategies can improve search efficiency and convergence. Based on the quality and distribution of solutions, the search step size, crossover and mutation probabilities, or pheromone update rules can be dynamically adjusted to enable rapid exploration of the solution space in the early stages and refined optimization in the later stages. For multi-objective shop-floor scheduling problems, properly handling the objective function is crucial. Clarify the relationships between multiple objectives and transform them into comprehensive evaluation indicators or non-dominated solutions using weight coefficients or Pareto optimization methods. At the same time, accurately address process, resource, and time constraints and incorporate them into algorithm design.

[0090] Furthermore, in the weekly production planning and scheduling module, AI optimization algorithms, specifically utilizing simulated annealing and genetic algorithms, handle complex production environments and balance production cycles, equipment load, and production capacity. Dynamic scheduling and adaptive control adjust production plans in real time, quickly responding to equipment failures or order changes, and intelligently optimizing production processes. Big data prediction and decision support: Production data is collected, deep learning analysis is performed, demand and bottlenecks are predicted, and production tasks are planned in advance. Through human-machine collaborative intelligent decision-making, AI outputs scheduling plans, allowing managers to fine-tune based on experience, combining the efficiency of AI with human judgment. These technological breakthroughs make the system uniquely competitive in the market.

[0091] Understandably, existing solutions on the market generally face adaptability issues in the construction machinery industry. For example, the APS system and Siemens' Plant Simulation have not been ideally applied in construction machinery. The main reasons include a lack of specific simulation support for the construction machinery industry, complex operations and high training costs, and limited integration capabilities with IoT, MES, ERP, and other systems.

[0092] In view of this, in some specific embodiments of the present invention, such as Figure 2As shown, this solution provides a manufacturing digital twin system based on simulation and operations optimization. Based on discrete event simulation theory, it simulates the connections between production factors in manufacturing entities, providing a holistic factory perspective for cross-departmental communication. Machine learning and operations optimization technologies are used to discover patterns and trends in data, automatically find optimal solutions, and provide operators with feedback and control solutions for manufacturing entities. Core modules such as lightweight production process simulation, weekly production planning and scheduling, and shop floor scheduling engines are promoted and applied, providing new ideas for the innovation of intelligent manufacturing and digital lean management systems.

[0093] The following is an analysis of the business architecture and technical architecture. Figure 2 The manufacturing digital twin system based on simulation and operations optimization is described in detail.

[0094] Figure 2 The business architecture of the manufacturing digital twin system provided, such as Figure 3 As shown, simulation, process engineering, operations, and operational research algorithms together form the core of the manufacturing digital twin system. Their interaction forms a closed-loop optimization process. Simulation technology simulates actual production processes in a virtual environment, including the interactions and flows of production lines, workers, equipment, and logistics. This provides a platform for experimentation and optimization for process engineering and operations departments. The process engineering department uses this simulation data to test and improve production layouts and optimize product manufacturing processes. Operations departments, on the other hand, use simulation results to guide actual production activities, such as material demand forecasting, production planning, and job scheduling optimization.

[0095] Operational research algorithms play a decision-making support role in this process. They analyze simulation data and apply optimization techniques to solve complex production decision-making problems, such as resource allocation and scheduling optimization, to achieve overall improvements in labor efficiency and financial indicators. This collaborative working model of simulation, process, operations, and operational research algorithms not only improves production efficiency and product quality, but also reduces costs and risks, making the entire manufacturing process more flexible and responsive to market changes. Through this integrated business architecture, the manufacturing digital twin system can provide enterprises with continuous process improvement and decision support, driving the development of intelligent and automated manufacturing operations.

[0096] Figure 4 The technical architecture of the manufacturing digital twin system provided, such as Figure 4 As shown, the overall structure is divided into data layer, business layer, presentation layer, and communication and deployment components.

[0097] At the data layer, the system collects and stores necessary data through the data center, including persistent data and logging, to ensure data integrity and availability. The business layer handles core business logic, including simulation services, scheduling services, and reporting services. It interacts with the front end through POST and GET requests, and leverages the data center to integrate data from various sources to support intelligent decision-making.

[0098] The presentation layer is responsible for presenting data from the business layer in a user-friendly manner. It uses Vue.js and Element UI to build an interactive interface, allowing users to easily interact with the system. Internal system components communicate via HTTP, using Nginx as a reverse proxy to ensure high availability and load balancing, and providing third-party interfaces for inter-system integration.

[0099] In terms of deployment, the system uses Docker containerization technology to improve portability and scalability, while using JWT for permission control to ensure the security of the system.

[0100] Overall, this technical architecture realizes a clearly layered, modular and efficient manufacturing digital twin system, supports complex simulation and operations optimization tasks, provides enterprises with intelligent production decision support, and helps achieve operational excellence.

[0101] To further illustrate the practical value of the embodiments of the present invention in the application of engineering machinery manufacturing, the specific contents of the present invention are further verified through three specific embodiments below.

[0102] Example 1 Lightweight production process simulation - Sany Heavy Machinery Huaxiang fuel tank assembly production line: The following problems exist on the fuel tank production line of Huaxiang Factory: Question 1: From a process perspective, the theoretical production cycle of a fuel tank is about one day, but in reality, the production of a fuel tank takes seven days. As a result, the 2010 factory was unable to produce according to the 3+1 order, resulting in the contradictory phenomenon of frequent fuel tank shortages and stagnant inventory.

[0103] Question 2: There are a large number of oil tanks in production piled up on site, especially before and after the cleaning and shielding process. Normally, there are about 100 oil tanks waiting to be processed.

[0104] The goal of this case study was to identify the core causes of long fuel tank production cycles and work-in-process (WIP) accumulation. By identifying these core issues, engineers hoped to develop targeted solutions and optimize fuel tank production efficiency.

[0105] Application: Simulation modeling: Process engineers simulate the fuel tank production process. Figure 5As shown, the model results are analyzed to locate the core problems of the production line.

[0106] Problem Analysis: Through simulations and analysis of the current production line, the research team identified the shield cleaning process as the bottleneck affecting production efficiency. This was consistent with the subjective assessment of on-site manufacturing personnel. The reason for this bottleneck was that the number of workers assigned to this process (3) was far less than the planned number of workers (8).

[0107] In order to solve the problem of low production efficiency of downstream processes caused by bottleneck workstations, the strategy adopted on site is to produce in staggered shifts, such as Figure 6 As shown, each process can be produced efficiently in its own shift.

[0108] However, this solution presented a problem: the tanks would take half a day to move to the next process, and this waiting time for the guide tanks accounted for over 90% of the entire production cycle. The on-site solution was the main cause of the long production cycle and the accumulation of tank work-in-progress before and after the cleaning and shielding process.

[0109] Summary of results: To address this core issue, process personnel developed two solutions. By comparing the results of the solutions through simulation, they ultimately chose to merge the production of Tank Lines 2 and 3, combining the two shifts. Following these improvements, the Huaxiang Tank Line's production cycle was shortened by 28%, and the average daily work-in-progress was reduced by 30%. After rigorous financial accounting based on energy consumption, capital utilization, depreciation, and other factors, a total of 2.614 million yuan in cost reduction benefits were generated annually. The core reason for choosing this solution was that it met five evaluation dimensions and offered better economic benefits, including: By combining the two lines of workers into one shift, without any additional investment, the gaps between the number of cleaning and shielding personnel and the planned number of people can be filled; From the perspective of process feasibility, the No. 3 fuel tank production line can meet the production needs of large, medium and small fuel tanks; The line connection plan can be executed on site; Simulate the solution to meet the production capacity requirements of 6,000 fuel tanks on the production line; The production cycle of simulated fuel tanks was shortened by 2 days, and the average daily work-in-progress was reduced by 200 pieces; This directly reduces the operating time of one coating line and greatly saves energy.

[0110] Example 2 Shop floor scheduling engine: Sany Heavy Machinery's mini excavator light tower factory assembly line: Issues resolved: The assembly line of the micro-excavator lighthouse factory faces complex workshop scheduling problems during the production process. Traditional scheduling methods are difficult to effectively deal with a variety of constraints and optimization goals, resulting in low production efficiency, high costs, and difficulty in guaranteeing delivery time. Specifically, it manifests itself as: unreasonable resource allocation: the allocation of machinery, equipment, manpower and materials fails to fully consider the actual needs of the assembly line, resulting in the coexistence of idle equipment and overload, unbalanced manpower allocation, and untimely material supply, which affects the continuity of production. Task sequencing is chaotic: the processing and assembly sequence of workpieces lacks scientific planning, and the connection between processes is not smooth, resulting in a lot of waiting time and extended production cycle. Insufficient response to uncertainty: There is a lack of effective response strategies for uncertain factors such as machine failures and order changes, and production plans are frequently adjusted, resulting in a significant decline in production efficiency.

[0111] Application: Based on the actual business scenarios of the production line and the simulation analysis results of the digital simulator, the operations optimization algorithm is constructed by comprehensively considering factors such as constraints and bottlenecks. In terms of algorithm selection, the best algorithm is selected according to its characteristics and optimization results to ensure that the algorithm can best adapt to the production needs of the micro-excavation assembly line. The algorithm framework is shown in Figure 7 .

[0112] For example, the hybrid artificial bee colony algorithm uses an artifact encoding approach and a NEH-based initial solution generation method to improve the quality of the initial population. By embedding a genetic algorithm and introducing genetic operations into the artificial bee colony algorithm's search process, global search capabilities are enhanced. A variable neighborhood search mechanism based on inserting / exchanging neighborhoods is also established to expand local search capabilities. The adaptive large neighborhood algorithm adaptively changes its search strategy based on solution quality and search progress. Initially, it rapidly explores the solution space by expanding the neighborhood, and later, it refines the solution by narrowing the neighborhood. This effectively avoids falling into local optima and improves the algorithm's global optimization capability and convergence speed.

[0113] To address production line pain points and user needs, an automated scheduling engine for micro-excavator assembly lines, centered around an operations optimization algorithm, was built to enhance the scientific nature of production line planning decisions and improve production efficiency and resource utilization. The algorithm was successfully embedded into the production and processing workflow of the micro-excavator lighthouse factory assembly line. Through integration with the Manufacturing Operations Management (MOM) system, the algorithm results are transmitted to the MOM system in real time. Based on the algorithm's results, the MOM system locks the production queue and sends relevant instructions to the on-site execution control center, thereby guiding assembly line processing. Since its implementation, this project has been operating stably on the micro-excavator assembly line for 19 months, fully demonstrating the reliability and practicality of this scheduling engine.

[0114] Summary of results: After the implementation of this project, the production efficiency of the mini excavator assembly line has been significantly improved. Compared with the AS embedded module and manual experience, the production time of this scheduling solution based on the operations optimization algorithm has been shortened by more than 10%. Figure 8 .

[0115] This improvement is due to the algorithm's optimization of task sequencing. By effectively reducing waiting times between processes, unnecessary stops of workpieces on the production line are avoided, making the entire production process more compact and smooth. At the same time, the algorithm rationally adjusts resource allocation. In terms of equipment utilization, by precisely matching equipment processing capacity with production tasks, idle equipment is fully utilized and the burden on overloaded equipment is alleviated, thereby improving overall equipment utilization. In terms of labor efficiency, scientific task allocation is carried out according to employee skills and work efficiency, ensuring that each employee can maximize their work efficiency in the appropriate position, reducing labor waste. Most importantly, after a financial audit, this optimization result has brought a cumulative economic benefit of more than 2.8 million yuan, fully demonstrating the great value of this production scheduling engine in improving production efficiency and economic benefits.

[0116] Example 3: Weekly production plan and scheduling: Sany Heavy Machinery mini excavator assembly line: After conducting a field survey on the mini excavator assembly line, we found the following major problems on the production line: Problem 1: According to worker feedback, the production plan exhibits an imbalance in workload. During certain periods, production workloads are heavy, requiring overtime to ensure on-time completion; while during other periods, workloads are relatively light, allowing workers to rest and downtime. This significant difference in work experience leads to an uneven distribution of production capacity within the production plan.

[0117] Problem 2: During the production process, some models sometimes encounter a shortage of key components, requiring temporary adjustments to production plans. This on-site production volatility is significant, seriously affecting production efficiency.

[0118] Question 3: When developing production plans, planners must consider multiple constraints, including order delivery deadlines, resource fulfillment rates, production calendars, and balanced production schedules for complex models. They often spend considerable time evaluating each of these constraints, a laborious and time-consuming process that often makes it difficult to simultaneously meet all constraints.

[0119] Problem 4: Due to frequent fluctuations in market demand and resource availability, production plans require frequent adjustments. These adjustments often trigger chain reactions. For example, increased demand can lead to resource shortages, and prioritizing urgent orders can impact the delivery of other orders. Planners must reassess constraints with each production plan adjustment, resulting in inefficient response and significant workload.

[0120] Application: To address these pain points, we developed a weekly production planning module based on precise solution algorithms and operations optimization algorithms. This module converts the core scheduling considerations for marketing orders, business procurement, manufacturing, and other scenarios into a series of algorithmic constraints. To address multiple indicator requirements such as capacity balance and resource utilization, we developed a multi-objective optimization algorithm based on quadratic programming and metaheuristic algorithms. This ensures that constraints are automatically met while rapidly iterating towards multiple objectives, finding high-quality solutions, and outputting reasonable production plans.

[0121] We have established a pool of optional constraint rules, such as Figure 9 As shown, planners can freely combine and match production lines according to the production scheduling requirements of different production lines at different production stages to meet the production scheduling requirements of different production lines in complex production environments such as mixed-line production workshops.

[0122] After the constraints are matched, the planner can click OK to automatically schedule production with one click. The scheduling results will be output in seconds while all the constraints are met. The specific scheduling results are as follows: Figure 10 shown.

[0123] Depend on Figure 10 It can be seen that the daily production capacity of the production plan output by the algorithm is relatively balanced. According to the production line replacement and line clearing scenarios at the No. 8 and No. 25 production and manufacturing sites, the algorithm reduced production capacity scheduling, flexibly adapted to the on-site conditions, and ensured production stability.

[0124] To address the frequent changes in marketing orders and resource arrivals, the weekly production planning and scheduling module is connected to data platforms such as the production management system (SAP). This allows for real-time data pull and rescheduling within seconds, providing immediate feedback and timely response. The system also retains manual maintenance functionality to flexibly address emergencies such as data errors or untimely updates. This improves the system's usability while allowing users of different roles to efficiently intervene and make decisions based on their needs. The red area in the image above represents the production plan adjustments made by planners in response to changes in resource arrivals caused by extreme weather events such as typhoons.

[0125] Summary of results: The weekly production planning and scheduling module has been applied in the mini-excavator assembly line of Sany Heavy Industry. The specific production benefits are summarized as follows: Improved inventory turnover rate: Through the precise scheduling of key resources (such as important parts and structural parts), inventory turnover rate increased by 10%.

[0126] Improved capacity utilization: Through a multi-objective optimization algorithm for capacity balancing and resource utilization, the production plan output by the weekly production planning and scheduling module effectively reduced the equipment idle rate in the workshop, fully released the overall capacity of the production line, and increased capacity utilization by 5%.

[0127] Improved production scheduling efficiency: Even in the face of sudden order changes and resource bottlenecks, the system can quickly adjust the production schedule to reduce production delays and resource waste caused by improper planning, and improve overall production scheduling efficiency by 50%.

[0128] Through the above embodiments, it can be seen that the manufacturing digital twin system provided by the present invention creatively integrates the simulation framework and the operations and scheduling module, uses the simulation module to simulate the actual situation of the production line, and then uses the operations module to optimize and make decisions on the production line, forming a closed-loop system for building digital models, simulation optimization, and decision-making control; a lightweight production process simulation module that is simple to use, fully technically independent and controllable, and has a low threshold is designed. It embodies the actual production experience and execution details of front-line production and process personnel, and develops a simulation engine that conforms to the actual manufacturing scenarios of the engineering technology industry. Process personnel and front-line workers can simulate, model, analyze, and optimize the workstations, workshops, and production lines they are responsible for, just like professional modelers, without the need to learn additional professional knowledge such as simulation, programming, and operations research. It will completely get rid of dependence on foreign simulation technology and have the potential for independent and controllable technological development. This will significantly lower the threshold for simulation, providing an effective simulation tool for small and medium-sized enterprises (SMEs) that lack the capacity to establish simulation teams. It can meet their needs for rapid production process verification, resource optimization, and improved production efficiency, helping SMEs enhance their competitiveness in the fiercely competitive market. The system also features hybrid algorithm fusion and co-evolution, integrating an artificial bee colony algorithm with a genetic algorithm (GA), using artifact encoding and NEH-based initial solution generation to improve initial population quality. A GA is embedded to enhance global search capabilities, while a variable neighborhood search mechanism based on insertion and exchange neighborhoods is established to expand local search capabilities. An adaptive search strategy is also employed to dynamically adjust parameters. The adaptive large neighborhood search algorithm dynamically adjusts the search neighborhood size. During the search process, damage and repair operators are introduced based on solution quality and search progress, adaptively changing the search strategy. Initially, the neighborhood is expanded to rapidly explore the solution space, while later, the neighborhood is narrowed to fine-tune the solution, avoiding local optima and improving the algorithm's global optimization capabilities and convergence speed. Multi-level visualization and interactive design: The system's visual interface design is another innovation. The weekly production planning and scheduling module provides diverse visualization tools for users at different levels. For production managers, the system provides intuitive displays of production scheduling execution, such as Gantt charts and progress charts; for operators, it provides detailed operational suggestions and real-time monitoring data. This multi-level design not only improves the ease of use of the system, but also allows users of different roles to quickly obtain key information according to their respective needs, and to carry out efficient manual intervention and decision-making; the combination of flexible production planning and real-time scheduling, the weekly production planning and scheduling module innovatively combines flexible production planning with real-time scheduling functions. Flexible production planning allows companies to flexibly respond to changes in orders and market demand, while real-time scheduling ensures that production tasks can be quickly adjusted during the production process, reducing delays and waste of resources. The system breaks the traditional fixed production scheduling model and can automatically optimize and dynamically update production scheduling plans based on real-time data, greatly improving the agility and stability of production.

[0129] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0130] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0131] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A manufacturing digital twin system, characterized in that: include: A production process simulation module is configured to simulate different production scenarios of the production process and the production process data under each of the production scenarios, simulate the production process under each of the production scenarios based on the production process data, and generate a simulation report; a shop scheduling engine configured to optimize the production task scheduling and resource allocation of the production workshop using an operations optimization algorithm based on the production scenarios provided by the production process simulation module, the production process simulation results under each of the production scenarios, the real-time production environment of the workshop, and the received weekly production plan, and generate a shop task scheduling plan; Feedback the workshop task production scheduling plan to the production process simulation module in real time, so that the production process simulation module simulates and verifies the workshop task production scheduling plan and obtains the production scheduling result corresponding to the workshop task production scheduling plan; A weekly production planning and scheduling module is configured to generate a weekly production plan based on the production scenario provided by the production process simulation module and the scheduling results corresponding to each of the workshop task scheduling plans; Based on the real-time acquired enterprise resource status, the weekly production plan is dynamically adjusted, and the dynamically adjusted weekly production plan is fed back to the production process simulation module and the workshop scheduling engine in real time, so that the production process simulation module simulates and verifies the weekly production plan, and the workshop scheduling engine updates the workshop task scheduling plan according to the new weekly production plan.

2. The manufacturing digital twin system according to claim 1, characterized in that: The simulation of the production process in each production scenario based on the production process data and the generation of a simulation report specifically include: Based on discrete event simulation theory, simulate random events in the production process, including one or more of material flow, equipment status and worker operation; Predicting changes in the state of production factors caused by the random events at each discrete time point, where the production factors include one or more of production line layout, equipment configuration, logistics path, and personnel allocation; The prediction results of the random events and the changes in the status of the production factors are used as production process data to simulate the production process.

3. The manufacturing digital twin system according to claim 1, characterized in that: After generating the production schedule for the shop floor tasks, the shop floor scheduling engine is further configured to: Based on the changes in production resources in the production workshop, the production task sequencing and resource allocation plan in the production workshop are dynamically adjusted to generate a new workshop task scheduling plan.

4. The manufacturing digital twin system according to claim 1, characterized in that: The real-time production environment of the workshop includes entity attributes, process logic, and uncertainty factors; Based on the production scenarios provided by the production process simulation module, the production process simulation results under each production scenario, the real-time production environment of the workshop, and the received weekly production plan, the production task scheduling and resource allocation of the production workshop are optimized using an operations optimization algorithm, specifically including: Based on the real-time production environment of the workshop and the received weekly production plan, perform entity modeling, process modeling and uncertainty modeling on the production task scheduling and resource allocation of the workshop task scheduling plan to obtain the entity model, process model and uncertainty model for the production scheduling plan; Select the appropriate operations research optimization algorithm based on the scale, structure, optimization objectives, and complexity of constraints of the shop floor scheduling problem; Based on the production scenarios provided by the production process simulation module, the production process simulation results under each of the production scenarios, and the pre-set optimization goals and constraints, the operations optimization algorithm is used to solve the entity model, process model, and uncertainty model.

5. The manufacturing digital twin system according to claim 4, characterized in that: The entity modeling includes: machine equipment modeling, material handling equipment modeling and workpiece modeling; The machine equipment modeling is used to model the processing capacity, time distribution, failure mode and probability of the machine equipment in the workshop to simulate the processing capacity and potential failure risks of the machine equipment; The material handling equipment modeling is used to model the speed, load capacity, and path planning rules of the material handling equipment to simulate the movement of the material handling equipment in the workshop and the impact of the handling tasks on the production process; The workpiece modeling is used to model the process route, process and processing time of each workpiece to simulate the impact of different processing paths and time requirements of each workpiece on the production process.

6. The manufacturing digital twin system according to claim 4, characterized in that: The process modeling includes: production process modeling and constraint condition modeling; The production process modeling is used to construct a production process model, which is used to describe the flow process of workpieces in the workshop, including the connection sequence of processes, the location of the buffer area, the transportation path, and the collaborative relationship between equipment; The constraint modeling is used to add equipment availability, material supply conditions, and buffer area quantity restrictions to the production process model to ensure that the production process model can reflect various restrictions in actual production.

7. The manufacturing digital twin system according to claim 4, characterized in that: The uncertainties mentioned include machine failure, raw material fluctuations, and order changes; The uncertainty modeling specifically includes: A probability model is established for uncertainty factors, and corresponding distribution functions are set to simulate the occurrence patterns of uncertainty factors in order to evaluate the performance of production scheduling plans under different scenarios.

8. The manufacturing digital twin system according to claim 4, characterized in that: The operations optimization algorithm for selecting a production scheduling solution based on the scale, structure, optimization objectives, and complexity of constraints of the shop floor scheduling problem specifically includes: Select the category of initial algorithm based on the scale, structure, optimization objectives and complexity of constraints of the shop scheduling problem; Optimize the parameters of the selected initial algorithm, including parameter setting, initial solution generation, and neighborhood structure design; The effectiveness of the initial algorithm is verified through experiments, the performance of different initial algorithms on the same problem is compared, and the optimal initial algorithm is selected as the operations optimization algorithm for the production scheduling plan.

9. The manufacturing digital twin system according to claim 8, characterized in that: The initial algorithm category is selected based on the scale, structure, optimization objectives, and complexity of constraints of the shop floor scheduling problem. Specifically, it includes: If the size of the shop scheduling problem is smaller than the preset size, the exact algorithm is used; If the scale of the shop scheduling problem is larger than the predetermined scale, a heuristic algorithm or a metaheuristic algorithm is used; If the objective function and constraints of the shop scheduling problem are linear, use the linear programming algorithm or the integer linear programming algorithm; If the objective function or constraints of the shop scheduling problem are nonlinear, a nonlinear programming algorithm or a mixed integer nonlinear programming algorithm is used; If the shop scheduling problem involves discrete decision variables, integer programming or mixed integer programming algorithms are used; If the shop scheduling problem involves only one optimization objective, a single-objective optimization algorithm is used; If there are multiple optimization objectives in the shop scheduling problem, a multi-objective optimization algorithm is used; If the constraints of the shop scheduling problem are simple constraints, a linear programming algorithm or an integer programming algorithm is used, wherein the simple constraints include a single entity constraint or a combination of multiple single entity constraints; If the constraints of the shop scheduling problem are complex constraints, a heuristic algorithm or a meta-heuristic algorithm is used, wherein the complex constraints include nonlinear constraints, dynamic constraints, and random constraints; If there are soft constraints and hard constraints in the shop scheduling problem, the constraint relaxation or penalty function method is used to transform the soft constraints into part of the objective function, where the soft constraints are constraints that can be partially violated and the hard constraints are constraints that must be strictly satisfied.

10. The manufacturing digital twin system according to any one of claims 1 to 9, characterized in that: The enterprise resource status includes: market demand data, production resource data and production status data, wherein the market demand data includes order quantity and delivery time, order priority, and market demand changes; the production resource data includes equipment availability, manpower availability and material supply; the production status data includes the complexity of production tasks, equipment maintenance plans and production bottlenecks.

11. The manufacturing digital twin system according to claim 1, characterized in that: The production process simulation module is used to simulate the connection and interaction between production factors in a manufacturing entity. The production process simulation module adopts a discrete event system modeling tool and supports drag-and-drop simulation of the behavior of each production factor.

12. The manufacturing digital twin system according to claim 1, characterized in that: The production process simulation module adopts a hierarchical modeling concept and supports the encapsulation of any custom atomic event object and the coupling of complex event objects.

13. The manufacturing digital twin system according to claim 1, characterized in that: The engine packaging of the production process simulation module adopts B / S architecture and is based on cloud native technology, supporting online modeling, simulation and optimization.

Images