A method for assembling aircraft wings based on digital twins

Abstract

The present invention discloses an aircraft wing assembly method based on digital twins, which includes an assembly preparation stage, a virtual assembly stage, an assembly and detection stage, and a quality tracing stage, and the specific steps are: determining a production plan; the workshop information management system allocates personnel and parts according to the production plan; RFID collects the name information of workers, parts, and workstations and enters them into the management system; UWB collects the real-time location information of workers, tools, and parts; constructs a virtual workshop, and workers use augmented reality glasses to perform virtual assembly; workers assemble in the real workshop according to the virtual assembly results; during the assembly process, the augmented reality glasses provide assembly suggestions in real time according to the guidance of remote engineers; after the assembly is completed, a deep learning algorithm is used to detect the assembly quality; the workshop information management system is used to trace the assembly process of the assembly parts that fail the inspection; and an expert system is used to guide and improve the quality defects. The method improves the assembly quality and assembly efficiency of the aircraft wing.

Description

Technical Field

[0001] The present invention relates to the field of aircraft assembly, and more particularly to an aircraft wing assembly method based on digital twins. Background Art

[0002] The assembly of complex equipment products such as aircraft is the final stage and key link in achieving product functions and performance. It is an important factor affecting product R&D quality and performance. Assembly quality largely determines the final quality of the product. At present, the technical status of domestic aircraft assembly production sites is complex. It is inconvenient for assembly personnel to consult materials, provide on-site information feedback, and collaborate remotely. It is easy to encounter problems such as inadequate assembly and assembly rework. Therefore, the assembly efficiency is low and the assembly quality is poor, which cannot meet the requirements of high-efficiency and high-quality aircraft assembly. Therefore, it is crucial to propose an efficient aircraft wing assembly method.

[0003] The current patent disclosures and literature show that: 1) Patent (CN109872080A) is a knowledge management method and system for aircraft assembly coordination rules. It explicitly expresses implicit assembly coordination knowledge by constructing a reasonable knowledge expression framework, then constructs a knowledge model for knowledge acquisition, and finally constructs an assembly coordination rule base to achieve effective management of assembly coordination rules. This method schedules the aircraft assembly process by constructing a rule base. The constructed rule base is relatively fixed and cannot meet the complex and changeable production needs of assembly workshops; 2) Patent (CN110991056A) is a method for aircraft assembly line job scheduling based on genetic variable neighborhood algorithm. By establishing a resource-constrained aircraft assembly line job scheduling model, the job scheduling problem in actual production is converted into a mathematical model problem for optimization and solution, with the goal of minimizing the total construction period of the assembly operation. At the same time, considering the constraints of the front and back, resource constraints and space constraints, a sub-segment job scheduling model for the aircraft assembly line is constructed, and finally an improved genetic algorithm is used to implement the scheduling of the aircraft assembly line. Variable neighborhood algorithm solution, this method transforms the assembly production scheduling problem into a mathematical model problem, and applies the genetic variable neighborhood algorithm to solve it, but ignores the problem that the assembly process is difficult to digitize and the assembly environment is complex and changeable. The results are prone to fall into local optimality and are difficult to guide assembly production; 3) Patent (CN108564304A) A parameterized generation method for operation descriptions for aircraft assembly, which classifies the operation descriptions and defines key attributes to assist process personnel to quickly find the required operation description templates, defines the operation template descriptions with parameters, guides process personnel to edit parameter descriptions, improves the standardization and rationality of process files, and obtains the resource names in the manufacturing resource library and all resource names in the process link where the process personnel are located to achieve rapid selection and filling of the parameter part. This method classifies the operation descriptions and solves the complexity of the operation to a certain extent, but there is no solution for special assembly problems. It still relies on communication between process personnel and engineers at each level, which affects production efficiency.

[0004] In summary, although the existing research results and methods can improve the efficiency of aircraft assembly to a certain extent, there are no good solutions to problems such as remote collaboration and document access among workers. Therefore, the above methods are difficult to meet the requirements of high-efficiency and high-quality aircraft assembly.

[0005] The combination of digital twin technology and aircraft wing assembly provides a possibility to further improve the efficiency and quality of aircraft wing assembly, and is an effective means to achieve efficient aircraft assembly. The aircraft wing assembly method based on digital twins deploys personnel and parts according to the production plan, collects the name information and real-time location information of workers, parts to be assembled, and workstations, and builds a virtual workshop. Workers use augmented reality glasses to perform virtual assembly according to the assembly process, and perform real assembly according to the virtual assembly results. During the assembly process, engineers use augmented reality glasses to remotely give assembly suggestions. After the assembly is completed, deep learning algorithms are used to detect the assembly quality. For assembly parts that fail the inspection, the workshop information management system is used to trace their assembly process, and expert systems are used to guide and improve quality defects, providing an effective technical approach for efficient and high-quality assembly of aircraft wings. Summary of the invention

[0006] The purpose of the present invention is to solve the above-mentioned problems and provide an aircraft wing assembly method based on digital twins. First, the production plan is determined, and the workshop information management system allocates personnel and parts according to the production plan. RFID collects the name information of workers, parts to be assembled, and workstations and enters them into the management system. UWB collects the real-time location information of workers, tools, and parts to be assembled to build a virtual workshop. Workers use augmented reality glasses to perform virtual assembly according to the assembly process and perform real assembly according to the virtual assembly results. During the assembly process, the augmented reality glasses provide assembly suggestions in real time according to the guidance of remote engineers. After the assembly is completed, a deep learning algorithm is used to detect the assembly quality. The workshop information management system is used to trace the assembly process of the assembly parts that fail the inspection, and an expert system is used to guide and improve the quality defects. This method has convenient remote collaboration, virtual assembly-real assembly-inspection-quality traceability full-cycle assembly quality control, and the assembly process is visualized. The actual assembly can be guided according to the virtual assembly results, truly realizing virtual control of the real, and effectively improving the assembly quality and efficiency of the aircraft wing.

[0007] The present invention achieves the above-mentioned purpose through the following technical solutions:

[0008] A method for assembling an aircraft wing based on digital twins, the technical solution of which includes four stages, namely, an assembly preparation stage, a virtual assembly stage, an assembly and inspection stage, and a quality traceability stage, including the following steps:

[0009] (1) The factory determines the production plan based on the order situation;

[0010] (2) The workshop information management system allocates personnel and parts according to the production plan to ensure that all resources required for production are complete;

[0011] (3) RFID collects the name information of workers, parts to be assembled, and workstations and enters it into the management system for deployment in the workshop information management system. UWB collects the real-time location information of workers, tools, and parts to be assembled to facilitate workers' assembly operations.

[0012] (4) Build a virtual workshop, where workers use augmented reality glasses to perform virtual assembly according to the assembly process, and perform real assembly in the real workshop based on the virtual assembly results. During the real assembly process, they use augmented reality glasses to refer to the assembly suggestions provided by remote engineers;

[0013] (5) After the assembly is completed, the deep learning algorithm is used to detect the assembly quality, with a focus on the connection parts of the assembly;

[0014] (6) Use the workshop information management system to trace the assembly process of the assembly parts that failed the inspection, and find out the steps and specific operators that affect the assembly quality;

[0015] (7) Use expert systems to analyze assembly quality defects, and use the conclusions drawn to repair unqualified parts and guide subsequent assembly to prevent similar defects from occurring again.

[0016] As a further improvement of the present invention, in step (1), the number of personnel and parts required for production at each workstation during the assembly process, as well as the estimated start time of each process are determined according to the production plan and the equipment usage of each workstation.

[0017] As a further improvement of the present invention, in step (2), the workshop information management system coordinates the resources required for each workstation according to the production plan combined with the specific conditions of equipment maintenance, personnel rest, and parts inventory, ensures the production needs of different workstations in the same period, and coordinates the production progress of each workstation.

[0018] As a further improvement of the present invention, in step (3), an RFID integrated machine is used to collect name information from the RFID chips of workers, parts to be assembled, and workstations and enter the information into the management system, which is used for the workshop information management system to monitor resource utilization and make adjustments. The UWB terminal identifies the UWB tags of workers, tools, and parts to be assembled for real-time positioning. Workers can view the location of tools and parts in real time, which facilitates assembly operations.

[0019] As a further improvement of the present invention, a virtual workshop is constructed in step (4), and workers use augmented reality glasses to perform virtual assembly according to the assembly process. During the virtual assembly process, the workers can further understand the assembly process and practice assembly difficulties. The virtual assembly results are used to guide the assembly in the real workshop. When workers encounter assembly problems during the real assembly process, they can use augmented reality glasses to communicate and discuss remotely with engineers in real time.

[0020] As a further improvement of the present invention, after the assembly in step (5) is completed, a convolutional neural network algorithm is used to detect the assembly quality, with a focus on detecting the connection parts of the assembly. Before the algorithm is applied, a large number of assembly photos are required for learning and training.

[0021] As a further improvement of the present invention, in step (6), the workshop information management system is used to trace the assembly process of the assembly parts that fail the inspection, thereby finding out the steps and specific operators that affect the assembly quality, analyzing the rationality of the assembly steps, and providing targeted training to the operators.

[0022] As a further improvement of the present invention, in step (7), the cause of the defect and the specific modification plan are obtained based on the inference engine and knowledge base of the expert system, which are used for rework of unqualified parts and guidance of subsequent assembly. Workers modify the unreasonable assembly process according to the suggestions of the expert system.

[0023] Compared with the prior art, the present invention has the following beneficial effects:

[0024] The present invention improves the intelligence level of aircraft wings by establishing a workshop information management system and equipping workers with wearable devices, deploys personnel and parts according to the production plan, collects the name information and real-time location information of workers, parts to be assembled, and workstations, and workers can use augmented reality glasses to perform virtual assembly according to the assembly process and perform real assembly according to the virtual assembly results. During the assembly process, engineers can remotely give assembly suggestions through augmented reality glasses, and after the assembly is completed, a deep learning algorithm is used to detect the assembly quality. The workshop information management system is used to trace the assembly process of assembly parts that fail the inspection, and an expert system is used to guide and improve quality defects, which effectively solves the problems of inconvenience in assemblers consulting materials, on-site information feedback, and remote collaboration, and realizes virtual assembly-real assembly-inspection-quality tracing full-cycle assembly quality control, improves the visualization of the assembly process, can guide the actual assembly according to the virtual assembly results, truly realizes virtual control of the real, and effectively improves the assembly quality and efficiency of aircraft wings. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0026] Figure 1 It is an execution flow chart of a digital twin-based aircraft wing assembly method according to the present invention;

[0027] Figure 2 It is a schematic diagram of a networking mode of a workshop information management system of a digital twin-based aircraft wing assembly method according to the present invention;

[0028] Figure 3 It is a virtual assembly example diagram of an aircraft wing assembly method based on digital twinning according to the present invention;

[0029] Figure 4 It is a schematic diagram of a deep learning algorithm of a digital twin-based aircraft wing assembly method described in the present invention;

[0030] Figure 5 This is an expert system architecture diagram of an aircraft wing assembly method based on digital twinning described in the present invention. DETAILED DESCRIPTION

[0031] The present invention is further described in detail below in conjunction with the accompanying drawings. The accompanying drawings described herein are part of this application and are used to further explain the present invention, but do not constitute a limitation of the present invention.

[0032] like Figure 1-Figure 5 As shown, a digital twin-based aircraft wing assembly method mainly includes four stages, namely, assembly preparation stage, virtual assembly stage, assembly and inspection stage, and quality traceability stage, including the following steps:

[0033] (1) The factory determines the production plan based on the order situation, determines the personnel and parts required for each workstation in the assembly process, and the estimated start time of each process, which facilitates the workshop information management system to allocate resources.

[0034] (2) The workshop information management system allocates personnel and parts according to the production plan, ensures the production needs of different workstations in the same period, coordinates the production progress of each workstation, ensures high production efficiency, and prevents backlogs of personnel or orders.

[0035] (3) Use an RFID integrated machine to collect name information from the RFID chips of workers, parts to be assembled, and workstations and enter it into the management system. This information is used in the workshop information management system to monitor resource utilization and make reasonable allocations. The UWB terminal identifies the UWB tags of workers, tools, and parts to be assembled for real-time positioning. Workers can view the location of tools and parts in real time, facilitating assembly operations.

[0036] (4) Build a virtual workshop, where workers use augmented reality glasses to perform virtual assembly according to the assembly process. During the virtual assembly process, workers can further understand the assembly process, practice assembly difficulties, and perform assembly in the real workshop based on the virtual assembly results. During the real assembly process, if workers encounter assembly problems, they can use augmented reality glasses to communicate and discuss with engineers remotely in real time, and refer to the assembly suggestions provided by the remote engineers.

[0037] (5) After assembly, the convolutional neural network algorithm is used to detect the assembly quality, with a focus on the connection parts of the assembly. A large number of assembly photos are required for learning before the algorithm is applied.

[0038] (6) For assembly parts that fail the inspection, the workshop information management system is used to trace the assembly process, find out the steps and specific operators that affect the assembly quality, analyze the rationality of the assembly steps, and provide targeted training for the operators to prevent similar assembly failure incidents from occurring.

[0039] (7) Use expert systems to analyze assembly quality defects such as Figure 5 , the cause of the defect and the specific modification plan are obtained, which are used for rework of unqualified parts and guidance of subsequent assembly. Appropriate modifications can also be made to unreasonable assembly processes based on the expert system's suggestions to prevent similar defects from occurring.

[0040] like Figure 2As shown, the networking mode of the workshop information management system of a certain type of aircraft wing assembly workshop of an aircraft manufacturing plant is taken as an example to further illustrate the working principle and characteristics of the present invention. The workshop information management system program is installed on the computer, the Internet of Things screwdriver and the RFID desktop card issuer are connected to the computer through the serial port, the RFID all-in-one machine is connected to the computer through the network cable, the UWB anchor base station, the Internet of Things riveting gun, the handheld barcode scanning gun, and the augmented reality glasses are connected to the computer using industrial WIFI. The above hardware is connected to the computer to form a workshop information management system, wherein the Internet of Things screwdriver model is SD-NC3500L-S, the RFID desktop card issuer model is RFD5003-H, and the R The model of the FID all-in-one machine is RFD-U8910, the model of the industrial WIFI is TL-WAR1200L, the model of the UWB anchor base station is JETREE8000F5, the model of the IoT rivet gun is Gesipa-Taurus2, the model of the handheld barcode scanner is RFD-U7910, and the model of the augmented reality glasses is Hololens2. The RFID component collects the name information of the workers, parts to be assembled, and workstations, and the UWB component collects the real-time location information of the workers, tools, and parts to be assembled, and enters it into the workshop information management system. The workshop information management system deploys personnel and parts based on the above information and combined with the production plan to ensure that all the assembly resources required for production are complete.

[0041] like Figure 3 As shown, the working principle and characteristics of the present invention are further explained by taking the virtual assembly in the wing box assembly of a certain type of aircraft as an example. The virtual assembly process flow instruction panel is displayed to the workers through augmented reality glasses. The workers perform virtual assembly according to the specific steps on the panel and guide the actual assembly according to the virtual assembly results. Remote technical engineers can also guide and monitor the virtual assembly process.

[0042] like Figure 4As shown in the figure, it is the structural diagram of the convolutional neural network assembly quality detection algorithm, where the convolution kernel size of C1 is set to 3×3, the depth is 5, the convolution kernel size of C3 is set to 3×3, the depth is 10, the initial weights of the convolution kernels of C1 and C3 are set to random numbers with a mean of 0 and a variance of 0.1 that satisfy the normal distribution, the bias is initialized to a constant of 0.1, and both C1 and C3 use the ReLu activation function to delinearize their calculation results. S2 and S4 are both maximum pooling layers, and their convolution kernel size is set to 2×2, the moving step is 2, and local response normalization is performed after the pooling operation to enhance the generalization ability of the model. The convolution layer and the pooling layer do not use all 0 filling. The number of nodes in the fully connected layer F5 is set to 128, the number of nodes in F6 is set to 84, and the number of nodes in the output layer is determined according to the specific detection object. The processed image is used as the input of the convolutional neural network. After a series of convolution and pooling operations, the deep abstract features of the image are extracted and generated. These feature matrices are flattened into one-dimensional vectors as the input of the full-link layer, and the full-link layer completes the classification task. Finally, the probability distribution is obtained through the Softmax layer, thereby obtaining the classification results of qualified quality, quality defects, and defect types. Finally, the defective connection parts are picked out, and their assembly process is traced through the workshop information management system to find out the steps and specific operators that affect the assembly quality, and the cause of the defect and the specific modification plan are obtained through the expert system's reasoning engine and knowledge base, which are used for rework of unqualified parts and guidance of subsequent assembly.

[0043] In summary, in actual production, the production plan is first determined; the workshop information management system allocates personnel and parts according to the production plan; RFID collects the name information of workers, parts to be assembled, and workstations and enters them into the management system; UWB collects the real-time location information of workers, tools, and parts to be assembled; workers use augmented reality glasses to perform virtual assembly according to the assembly process; workers perform real assembly according to the virtual assembly results; during the assembly process, augmented reality glasses provide real-time assembly suggestions according to the guidance of remote engineers; after the assembly is completed, the deep learning algorithm is used to detect the assembly quality; the workshop information management system is used to trace the assembly process of the assembly parts that fail the inspection; the expert system is used to guide and improve the quality defects. This method effectively solves the problems of inconvenience for assembly personnel to consult materials, on-site information feedback, and remote collaboration, and realizes the full cycle assembly quality control of virtual assembly-real assembly-inspection-quality traceability, improves the visualization of the assembly process, and can guide the actual assembly according to the virtual assembly results, truly realizing virtual control of reality, and effectively improving the assembly quality and assembly efficiency of aircraft wings.

[0044] 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 the same. Although the present invention has been described in detail with reference to the above embodiments, a person skilled in the art can still modify or make equivalent substitutions to the specific implementation schemes of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention are within the scope of protection of the claims of the present invention to be approved.

Claims

1. A method for assembling an aircraft wing based on digital twins, characterized in that: The aircraft wing assembly method based on digital twins has four stages, namely, assembly preparation stage, virtual assembly stage, assembly and inspection stage, and quality traceability stage, which includes the following specific steps: (1) The factory determines the production plan based on the order situation; (2) The workshop information management system allocates personnel and parts according to the production plan to ensure that all resources required for production are complete; (3) RFID collects the name information of workers, parts to be assembled, and workstations and enters it into the management system for allocation by the workshop information management system. UWB collects the real-time location information of workers, tools, and parts to be assembled to facilitate workers' assembly operations; (4) Build a virtual workshop, where workers use augmented reality glasses to perform virtual assembly according to the assembly process, and perform real assembly in the real workshop based on the virtual assembly results. During the real assembly process, they use augmented reality glasses to refer to the assembly suggestions provided by remote engineers; (5) After assembly, deep learning algorithms are used to detect assembly quality, with a focus on the connection parts of the assembly. (6) For assembly parts that fail the inspection, the workshop information management system is used to trace the assembly process to find out the steps and specific operators that affect the assembly quality. (7) Expert systems are used to analyze assembly quality defects, and the conclusions drawn are used to repair unqualified parts and guide subsequent assembly to prevent similar defects from occurring again.

2. The aircraft wing assembly method based on digital twinning according to claim 1, characterized in that: In step (1), the number of personnel and parts required for production at each workstation during the assembly process, as well as the estimated start time of each process, are determined according to the production plan and the equipment usage of each workstation.

3. The aircraft wing assembly method based on digital twinning according to claim 1, characterized in that: In step (2), the workshop information management system coordinates the resources required for each workstation according to the production plan combined with the specific conditions of equipment maintenance, personnel rest, and parts inventory, ensures the production needs of different workstations in the same period, and coordinates the production progress of each workstation.

4. The aircraft wing assembly method based on digital twinning according to claim 1, characterized in that: In step (3), an RFID integrated machine is used to collect name information from the RFID chips of workers, parts to be assembled, and workstations and enter it into the management system for the workshop information management system to monitor resource utilization and make adjustments. The UWB terminal identifies the UWB tags of workers, tools, and parts to be assembled for real-time positioning. Workers can view the location of tools and parts in real time to facilitate assembly operations.

5. The aircraft wing assembly method based on digital twinning according to claim 1, characterized in that: In step (4), a virtual workshop is constructed, and workers use augmented reality glasses to perform virtual assembly according to the assembly process. During the virtual assembly process, workers can further understand the assembly process and practice assembly difficulties. The virtual assembly results are used to guide the assembly in the real workshop. When workers encounter assembly problems during the real assembly process, they can use augmented reality glasses to communicate and discuss with engineers remotely in real time.

6. The aircraft wing assembly method based on digital twinning according to claim 1, characterized in that: After the assembly in step (5) is completed, the convolutional neural network algorithm is used to detect the assembly quality, with a focus on the connection parts of the assembly. Before the algorithm is applied, a large number of assembly photos are required for learning and training.

7. The aircraft wing assembly method based on digital twinning according to claim 1, characterized in that: In step (6), the workshop information management system is used to trace the assembly process of the assembly parts that failed the inspection, thereby finding out the steps and specific operators that affect the assembly quality, analyzing the rationality of the assembly steps, and providing targeted training for the operators.

8. The aircraft wing assembly method based on digital twinning according to claim 1, characterized in that: In step (7), the expert system's reasoning engine and knowledge base are used to derive the cause of the defect and the specific modification plan, which is used to repair the unqualified parts and guide the subsequent assembly. The workers modify the unreasonable assembly process according to the expert system's suggestions.

Images