Emergency dynamic scheduling method for aviation material supply chain based on digital twin technology

Abstract

The present invention relates to the field of artificial intelligence decision-making technology and discloses a method for emergency dynamic scheduling of aviation material supply chains based on digital twin technology. The method comprises: constructing a digital twin model covering each physical node of the civil aviation material supply chain and mapping operational data; integrating the operational data and external environmental data of the digital twin model as input features, determining the probability of an emergency event through an emergency event identification model, and substituting this probability into a simulation model for deduction to quantitatively assess the potential disruption risk caused by the emergency event to the supply chain; based on the potential disruption risk, using a multi-objective dynamic optimization algorithm with transportation cost, time penalty, and inventory cost as optimization objectives to generate a dynamic scheduling plan for aviation materials in the supply chain; scheduling logistics execution equipment according to the dynamic scheduling plan, and obtaining task execution data fed back by the logistics execution equipment in real time to update the digital twin model. The present invention achieves efficient, accurate, and automated configuration of emergency resources.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence decision-making technology, and in particular to an emergency dynamic scheduling method for an aviation material supply chain based on digital twin technology. Background Art

[0002] In the civil aviation industry, supply chain management for aircraft parts (also known as aviation materials) is a key component in ensuring the smooth operation of flights. Aviation materials are diverse, valuable, and subject to highly variable demand. The stability and responsiveness of their supply chain directly impact airlines' operating costs and safety.

[0003] Existing technologies have the following shortcomings in aviation material supply chain management, especially in emergency dispatch under emergencies:

[0004] 1. Fragmented and delayed perception of supply chain status: Traditional supply chain management relies heavily on information systems such as enterprise resource planning (ERP) or warehouse management systems (WMS). However, these systems are often siloed, with data residing in separate "islands" (e.g., data from inventory, logistics, procurement, and maintenance systems are not interoperable). This makes it difficult for managers to obtain a comprehensive, real-time view of the supply chain. There is a significant delay in understanding key information such as the status of in-transit materials, remote warehouse inventory, and supplier production capacity, hindering rapid and accurate emergency decision-making.

[0005] 2. Emergency response is primarily reactive and lacks predictive capabilities: For emergencies such as natural disasters, public health outbreaks, and sudden supplier disruptions, the current emergency response model is largely reactive and post-event. After an incident occurs, managers rely on manual experience and limited information to coordinate and allocate resources. There is a lack of effective technical means to predict and quantitatively assess potential risks, making it impossible to prepare resources and conduct emergency plan drills before risks occur.

[0006] 3. Scheduling decisions rely on static models and a single optimization dimension: When emergency material allocation is required, traditional path planning or resource allocation algorithms are often based on static, idealized models. For example, these algorithms may base their calculations on fixed transit time and cost parameters, failing to incorporate dynamic variables such as real-time traffic conditions, temporary air traffic control regulations, and changing weather conditions into the optimization process. Furthermore, the optimization objectives are often relatively simple (e.g., focusing solely on minimizing cost or time), making it difficult to achieve a dynamic balance between multiple objectives such as cost, timeliness, and risk in emergency scenarios.

[0007] 4. Disconnect between digital simulation and physical execution: Although some advanced simulation software can simulate supply chains, these simulations are typically offline and based on pre-set scripts. There is a lack of standardized, automated interfaces between the "optimal" solutions derived from simulations and actual logistics execution equipment (such as automated guided vehicles (AGVs) and drones). Decision-making and feedback on execution status still require significant manual intervention, preventing the formation of a closed-loop adaptive system from virtual decision-making to physical execution and status feedback, significantly reducing response efficiency.

[0008] Therefore, how to build an aviation material supply chain emergency dispatch system that can perform real-time perception, active prediction, dynamic optimization and closed-loop execution is a technical challenge that needs to be solved urgently in this field. Summary of the Invention

[0009] In order to solve the problems existing in the above-mentioned prior art, the present invention provides a digital twin scheduling method that integrates predictive analysis, dynamic simulation and closed-loop control to achieve efficient, accurate and automated configuration of emergency resources.

[0010] In order to achieve the above-mentioned object of the invention, the technical solution provided by the present invention includes:

[0011] The emergency dynamic scheduling method for aviation material supply chain based on digital twin technology includes the following steps:

[0012] For a civil aviation material supply chain, a digital twin model covering each physical node is constructed, and the operating data of each physical node is mapped to the digital twin model through a real-time data synchronization mechanism;

[0013] The operational data of the digital twin model and external environment data are collected as input features, and the probability of an emergency event is determined through an emergency event identification model. The probability is then substituted into a simulation model for deduction to quantitatively assess the potential disruption risk caused by the emergency event to the supply chain;

[0014] Based on the potential disruption risk, a multi-objective dynamic optimization algorithm with transportation cost, time penalty, and inventory cost as optimization objectives is used to generate a dynamic scheduling plan for the aviation materials in the supply chain;

[0015] The logistics execution equipment is scheduled according to the dynamic scheduling plan, and the task execution data fed back by the logistics execution equipment is obtained in real time to update the digital twin model.

[0016] Preferably, the physical nodes include: raw material supply nodes at the supplier layer, central or branch subsidiary warehouse nodes at the regional warehouse layer, and airport site nodes at the maintenance point layer.

[0017] Preferably, the real-time data synchronization mechanism includes: dynamically allocating data synchronization frequency using a hierarchical weighted synchronization strategy according to the importance level of the physical node in the supply chain.

[0018] Preferably, the operating data includes at least one or a combination of the following:

[0019] Real-time inventory data of the node, including aviation material code, quantity and shelf life information;

[0020] Transportation status data of materials in transit, including information on quantity in transit and estimated arrival time;

[0021] The node's storage or transportation equipment health data.

[0022] Preferably, the external environment data includes: meteorological warning data and / or major social event data representing the macro-social operating status.

[0023] Preferably, the step of performing deduction by the simulation model includes:

[0024] Based on the probability of occurrence of the emergency, boundary conditions of the simulation are set, wherein the boundary conditions include the number of simulation iterations and the probability distribution of key variables;

[0025] In each iteration, random sampling is performed according to the boundary conditions to simulate the inventory disruption of the supply chain and the feasibility of the transportation path under the scenario of the emergency;

[0026] After the iteration is completed, all deduction results are counted to generate a risk heat map that represents the spatial distribution of risks.

[0027] Preferably, the key variables include at least one or a combination of the following: transportation delay time; inventory availability or loss rate; equipment failure rate.

[0028] Preferably, the multi-objective dynamic optimization algorithm uses real number coding to represent the scheduling path node sequence, and performs optimization through an elite retention strategy and a dynamic fitness function combined with real-time traffic data.

[0029] Preferably, the task execution data includes:

[0030] The real-time geographic location coordinates or task completion status of the logistics execution equipment;

[0031] The code and quantity information of the aviation materials carried or operated by the logistics execution equipment;

[0032] The self-diagnostic data is collected and fed back by the sensors carried by the logistics execution equipment.

[0033] Preferably, the method for updating the digital twin model includes:

[0034] Update the corresponding operation data in the digital twin model according to the task execution data to maintain the state consistency between the physical entity and the virtual model;

[0035] Based on the updated operating data, when it is detected that the actual execution state deviates from the preset scheduling target, re-planning of the dynamic scheduling solution is triggered.

[0036] Beneficial effects

[0037] 1. This invention breaks down the data silos of traditional information systems by building a digital twin model covering the entire supply chain. It uniformly maps and visualizes real-time operational data from physical entities such as suppliers, warehouses, in-transit materials, and maintenance sites. This enables managers to obtain a comprehensive, dynamic, and accurate "living map" of the supply chain, significantly enhancing their overall situational awareness.

[0038] 2. This invention incorporates machine learning and simulation technology, integrating internal and external supply chain data to proactively predict potential risks such as supply disruptions and demand surges, and quantitatively assess their potential impact. This enables the initiation of emergency response plans and the preparation of resources to be based on data-driven scientific forecasts rather than post-hoc manual judgment, significantly enhancing the resilience and foresight of the supply chain.

[0039] 3. The multi-objective dynamic optimization algorithm employed by this invention incorporates dynamic environmental factors such as real-time traffic and weather as decision variables. It intelligently balances multiple mutually constrained objectives, such as cost, time, and inventory, to generate the optimal dynamic scheduling solution under the current real-world conditions. Compared to traditional static, single-objective algorithms, the decision-making results of this invention are more realistic and more cost-effective.

[0040] 4. This invention establishes an automated path from decision-making solutions in the digital twin model to the execution of instructions by logistics equipment in the physical world. It also continuously corrects and updates the digital twin model by obtaining real-time feedback. This closed-loop control of "perception-decision-execution-feedback" not only greatly improves the automation and efficiency of emergency response, but also enables the entire system to self-learn and continuously optimize in a real-world environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is a flow chart of an emergency dynamic scheduling method for civil aviation material supply chain based on digital twin technology provided in a preferred embodiment of the present invention. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments herein are only used to explain the present invention, rather than to limit the present invention.

[0043] Example 1

[0044] This embodiment provides a method for emergency dynamic scheduling of aviation material supply chain based on digital twin technology in civil aviation. Figure 1 As shown, this method can be applied by civil aviation enterprises to manage their aviation material supply chains, especially in the face of emergencies such as natural disasters, sudden equipment failures, and major social events, to ensure the stability and timeliness of the supply of key aviation materials. The specific steps of the method described in this embodiment are as follows:

[0045] S1. For a civil aviation material supply chain, a digital twin model covering each physical node is constructed, and the operating data of each physical node is mapped to the digital twin model through a real-time data synchronization mechanism.

[0046] The digital twin model is a high-fidelity mapping of the physical supply chain in a virtual space, capable of reflecting the status of physical entities in real time. In existing technologies, the implementation of a digital twin model is typically a comprehensive software project integrating multiple technologies. Its construction first relies on an Internet of Things (IoT) platform, such as Microsoft's Azure IoT Hub or Amazon's AWS IoT Core, to collect real-time operational data from physical sensors, PLCs (programmable logic controllers), and business systems such as ERP and WMS via protocols such as MQTT and OPC-UA. This data is then used to drive a virtual geometric model created in 3D modeling software (such as Autodesk Forge and Dassault Systèmes 3DEXPERIENCE) or a real-time rendering engine (such as Unity and Unreal Engine). To imbue this model with intelligence, it also requires integration with specialized simulation and analysis software (such as ANSYS, MATLAB / Simulink) or a Python-based machine learning platform to perform physical property simulation, performance prediction, and behavioral analysis. Many major cloud service providers also offer dedicated digital twin platforms, such as Microsoft's Azure Digital Twins and Amazon's AWS IoT TwinMaker. These platforms simplify the integration of data, models, and analytical capabilities through standardized model definition languages ​​(such as DTDL) and data interfaces, thereby creating a dynamic virtual replica that can reflect, diagnose, and predict the state of physical entities in real time. Since this section is not the focus of this article, it will not be discussed in detail.

[0047] In a specific application scenario, the physical nodes can be modeled hierarchically. For example, these include: raw material supply nodes at the supplier level, such as the manufacturer of specific aviation materials; central or subsidiary warehouse nodes at the regional warehouse level, responsible for the storage and transfer of aviation materials; and airport site nodes at the maintenance point level, representing the end-users of aviation materials, such as the maintenance repair (MRO) hangars at major airports. By meticulously modeling these nodes at different levels and with different functions, the network topology of the entire supply chain can be fully reproduced.

[0048] To maintain state consistency between the digital twin model and the physical supply chain, this embodiment employs a real-time data synchronization mechanism to map the operational data of each physical node to the digital twin model. The MQTT protocol can be used, which features low latency (e.g., latency can be less than 100ms), meeting the dynamic nature of the supply chain. To optimize system resources, this synchronization mechanism can be flexible. For example, a hierarchical weighted synchronization strategy can be used to dynamically allocate data synchronization frequency based on the importance of the physical nodes in the supply chain. For example, central warehouse nodes storing critical circulatory parts (e.g., engines and APUs) have the highest importance and can be assigned the highest synchronization frequency (e.g., seconds). Airport site nodes storing common consumable parts (e.g., screws and seals) can have their data synchronization frequency appropriately reduced (e.g., minutes).

[0049] The operational data is the basis of the digital twin model and may include at least one or a combination of the following:

[0050] Real-time inventory data of the node, which may specifically include the aviation material code (for example, a code according to ATA specifications), quantity, and shelf life information (for aviation materials with a storage limit, such as rubber products and chemicals).

[0051] The transportation status data of materials in transit describes the status of aviation materials that have been sent from one node but have not yet arrived at another node. Specifically, it may include the quantity in transit and the estimated time of arrival (ETA) information.

[0052] The health data of the node's storage or transportation equipment, such as the battery level and wear degree of key components of the stacker crane in the automated warehouse, the drone or smart vehicle performing transportation tasks, etc.

[0053] S2. Gather the operating data and external environment data of the digital twin model as input features, determine the probability of an emergency through an emergency identification model, and substitute the probability into a simulation model for deduction to quantitatively assess the potential disruption risk caused by the emergency to the supply chain.

[0054] Among them, the external environment data may include meteorological warning data that directly affects logistics (such as typhoon, blizzard, and heavy fog warnings) and / or major social event data that characterizes the macro-social operating status (such as traffic control caused by major international conferences, regional public health events, etc.).

[0055] These aggregated data are input into an emergency event identification model. This model can be a pre-trained machine learning model, such as a classification or regression model based on a gradient boosted decision tree (GBDT) or a deep neural network (DNN). By analyzing multi-dimensional input features, the model can calculate the probability of occurrence of specific emergencies (such as "a core supplier stops production for some reason" or "a major transportation artery is interrupted") in real time. It should be understood that the specific structure and training process of the emergency event identification model are not the focus of the present invention. Those skilled in the art can design it based on the purpose of the present invention and the existing technology, and the present invention does not make further requirements.

[0056] Subsequently, the probability is substituted into a simulation model for deduction to quantitatively assess the potential disruption risk caused by the emergency to the supply chain. The deduction steps of the simulation model may specifically include:

[0057] S201. Based on the probability of occurrence of the emergency, boundary conditions for the simulation are set. The boundary conditions include the number of simulation iterations and the probability distribution of key variables. For example, if the probability of an airport being closed due to a snowstorm is identified as 70%, this can be used as a key boundary condition.

[0058] The key variables may include at least one or a combination of the following: transportation delay time (for example, its probability distribution can be set to conform to the gamma distribution); inventory availability or loss rate (for example, the loss rate of aviation materials in a warehouse under a fire scenario may conform to the beta distribution); equipment failure rate (for example, the failure rate of transport drones will increase significantly in bad weather).

[0059] S202. In each iteration, random sampling is performed according to the boundary conditions to simulate the inventory disruption of the supply chain and the feasibility of the transportation route under the scenario of the emergency.

[0060] S203. After the iteration is complete, all deduction results are counted to generate a risk heat map that represents the spatial distribution of risk. For example, the expected probability of "aircraft material supply disruption" occurring at each airport during the public health period can be calculated, and a risk heat map representing the spatial distribution of risk can be generated to intuitively display the most vulnerable and dangerous links in the supply chain network.

[0061] S3. Based on the potential disruption risk, a multi-objective dynamic optimization algorithm with transportation cost, time penalty, and inventory cost as optimization targets is used to generate a dynamic scheduling plan for the aviation materials in the supply chain.

[0062] Among them, transportation cost refers to the direct logistics cost generated by implementing the scheduling plan; time penalty refers to the virtual cost generated by failure to meet aviation material demand on time. The more urgent the demand, the higher the time penalty coefficient; inventory cost includes holding cost and out-of-stock cost.

[0063] In one optional embodiment, the multi-objective dynamic optimization algorithm uses real-number encoding to represent the sequence of dispatch path nodes and performs optimization using a modified genetic algorithm or particle swarm optimization algorithm. To improve optimization efficiency and the practical feasibility of the solution, the multi-objective dynamic optimization algorithm uses real-number encoding to represent the sequence of dispatch path nodes (for example, [warehouse W1 → maintenance point M2 → maintenance point M3]) and optimizes using an elite retention strategy (for example, directly copying the top 5% of individuals in each generation to the next generation to accelerate convergence and avoid falling into local optimality) and a dynamic fitness function that incorporates real-time traffic data.

[0064] When evaluating the quality of a transportation route, the fitness function not only considers its static distance, but also accesses the map service API in real time to obtain the real-time congestion situation of the route, thereby calculating a more accurate estimated transportation time.

[0065] It should be understood that the output of the algorithm is a specific dynamic scheduling plan, for example: instructing warehouse A to use path R1, call vehicle V1, and urgently transport aviation material P1 to the maintenance point at airport B, and requiring it to arrive before time point T1.

[0066] Step S4: Schedule the logistics execution equipment according to the dynamic scheduling plan, obtain the task execution data fed back by the logistics execution equipment in real time, and update the digital twin model.

[0067] It should be understood that flow execution devices can be driverless trucks, delivery drones, automated guided vehicles (AGVs) in warehouses, etc. Specifically, a protocol conversion module can be included to convert the JSON-formatted optimization solution generated by the upper-level system into instructions that can be recognized by the lower-level device, such as the ROS message format for AGVs or MAVLink track files for drones.

[0068] The purpose of this step is to realize a closed-loop control system. In some preferred embodiments, the task execution data may specifically include:

[0069] The real-time geographic location coordinates of the logistics execution equipment (such as obtained through GPS or Beidou system) or the task completion status (such as "picked up", "in transit", "signed for").

[0070] The code and quantity information of the aviation materials carried or operated by the logistics execution equipment can be automatically obtained through the RFID or barcode scanner integrated in the equipment.

[0071] The logistics execution equipment collects and feeds back its own diagnostic data through sensors, such as tire pressure, fuel / battery level of vehicles, anemometer reading of drones, etc.

[0072] In some other preferred embodiments, the method for updating the digital twin model includes:

[0073] S401. Update the corresponding operational data in the digital twin model based on the task execution data to maintain state consistency between the physical entity and the virtual model. For example, when GPS data indicates that a transport vehicle has arrived at its destination, the status of the in-transit materials in the digital twin model is updated from unfinished to completed, and the inventory data of the corresponding node is also updated.

[0074] S402: Based on the updated operational data, if the actual execution status deviates from the preset scheduling target (for example, a vehicle significantly deviates from the planned route and time due to temporary traffic control, or the actual transportation time exceeds the plan by a preset percentage), the dynamic scheduling plan is triggered to replan to adapt to unexpected changes in the real world. At this point, the system returns to step S3 and uses the latest supply chain status as input to re-optimize the calculations and generate a new, more appropriate scheduling plan.

[0075] The basic principles, main features, and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the foregoing embodiments. The foregoing embodiments and descriptions are merely illustrative of the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. The emergency dynamic scheduling method of aviation material supply chain based on digital twin technology is characterized by: The steps include: For a civil aviation material supply chain, a digital twin model covering each physical node is constructed, and the operating data of each physical node is mapped to the digital twin model through a real-time data synchronization mechanism; The operational data of the digital twin model and external environment data are collected as input features, and the probability of an emergency event is determined through an emergency event identification model. The probability is then substituted into a simulation model for deduction to quantitatively assess the potential disruption risk caused by the emergency event to the supply chain; Based on the potential disruption risk, a multi-objective dynamic optimization algorithm with transportation cost, time penalty, and inventory cost as optimization objectives is used to generate a dynamic scheduling plan for the aviation materials in the supply chain; Scheduling logistics execution equipment according to the dynamic scheduling plan, and obtaining task execution data fed back by the logistics execution equipment in real time to update the digital twin model; The steps of deducing the simulation model include: Based on the probability of occurrence of the emergency, boundary conditions of the simulation are set, wherein the boundary conditions include the number of simulation iterations and the probability distribution of key variables; In each iteration, random sampling is performed according to the boundary conditions to simulate the inventory disruption of the supply chain and the feasibility of the transportation path under the scenario of the emergency; After the iteration is completed, all deduction results are counted to generate a risk heat map that represents the spatial distribution of risks.

2. The method for emergency dynamic scheduling of aviation material supply chain based on digital twin technology according to claim 1, characterized in that: The physical nodes include: raw material supply nodes at the supplier level, center or branch warehouse nodes at the regional warehouse level, and airport site nodes at the maintenance point level.

3. The method for emergency dynamic scheduling of aviation material supply chain based on digital twin technology according to claim 1, characterized in that: The real-time data synchronization mechanism includes: dynamically allocating data synchronization frequency using a hierarchical weighted synchronization strategy according to the importance level of the physical node in the supply chain.

4. The method for emergency dynamic scheduling of aviation material supply chain based on digital twin technology according to claim 1, characterized in that: The operation data includes at least one or more of the following: Real-time inventory data of the node, including aviation material code, quantity and shelf life information; Transportation status data of materials in transit, including information on quantity in transit and estimated arrival time; The node's storage or transportation equipment health data.

5. The method for emergency dynamic scheduling of aviation material supply chain based on digital twin technology according to claim 1, characterized in that: The external environment data includes: meteorological warning data and / or major social event data that characterizes the macro-social operating status.

6. The method for emergency dynamic scheduling of aviation material supply chain based on digital twin technology according to claim 1, characterized in that: The key variables include at least one or a combination of the following: transportation delay time; inventory availability or loss rate; equipment failure rate.

7. The method for emergency dynamic scheduling of aviation material supply chain based on digital twin technology according to claim 1, characterized in that: The multi-objective dynamic optimization algorithm uses real number coding to represent the scheduling path node sequence, and performs optimization through an elite retention strategy and a dynamic fitness function combined with real-time traffic data.

8. The method for emergency dynamic scheduling of aviation material supply chain based on digital twin technology according to claim 1, characterized in that: The task execution data includes: The real-time geographic location coordinates or task completion status of the logistics execution equipment; The code and quantity information of the aviation materials carried or operated by the logistics execution equipment; The self-diagnostic data is collected and fed back by the sensors carried by the logistics execution equipment.

9. The method for emergency dynamic scheduling of aviation material supply chain based on digital twin technology according to claim 1, characterized in that: The method for updating the digital twin model includes: Update the corresponding operation data in the digital twin model according to the task execution data to maintain the state consistency between the physical entity and the virtual model; Based on the updated operating data, when it is detected that the actual execution state deviates from the preset scheduling target, re-planning of the dynamic scheduling solution is triggered.

Images