DDS-based Aircraft Wing Segment Testing Method and System

Abstract

The present invention provides a DDS-based aircraft wing section testing method and system, comprising: establishing a data communication link between a test platform, a DDS Agent, and one or more hardware test modules through DDS; instructing the test platform to send test instructions; simultaneously instructing the hardware test modules to perform wing test operations and obtain real-time test data; the hardware test modules feed back the real-time test data to the test platform; during the test instruction sending and real-time test data feedback process, calculating the communication quality Q value and dynamically adjusting the data distribution strategy based on the Q value; the test platform receives and analyzes the real-time test data, simultaneously performs fault diagnosis, determines the cause of the fault, and locates the fault location on the digital twin model in the test platform. By integrating digital twins, machine learning, and DDS communication optimization, the present invention significantly improves the intuitiveness, fault diagnosis efficiency, and data transmission robustness of aircraft wing section testing.

Description

Technical Field

[0001] The present invention relates to the technical field of aircraft testing, and in particular to a DDS-based aircraft wing section testing method and system. Background Art

[0002] The development and testing of complex systems in modern industrial control, autonomous driving, aerospace, and other fields often involve a large number of distributed devices and nodes. Traditional test communication architectures, such as client-server models or message queue-based methods, are often limited by centralized architectures and struggle to meet the demands of scenarios with large node counts, frequent data exchanges, and stringent real-time requirements.

[0003] As an advanced communication middleware, the Data Distribution Service (DDS) provides high-performance, high-reliability, and low-latency data communication services for distributed systems through a decentralized publish-subscribe model and flexible Quality of Service (QoS) strategies. In DDS, publishers and subscribers are independent of each other, and data is organized and managed by topics. Subscribers can directly access relevant data based on topics without having to worry about the data's source or path. This enables a data-centric communication model, effectively alleviating existing issues with centralized architectures and improving communication efficiency.

[0004] During factory testing, aircraft are tested at the segment level. Wing segments include multiple functional integrity and state consistency checks, including wing rotation angle, fuel tank temperature, cable continuity, AC measurements, and more. Currently, these tests rely on testers performing separate operations and reading results on different areas, such as flaps, ailerons, and fuel tanks. This leads to numerous problems:

[0005] First, there is a lack of intuitive and comprehensive status monitoring and alarm methods. Currently, wing section test data is often recorded and presented in the form of scattered numerical values or independent charts, making it difficult for testers to intuitively grasp the integrated status and dynamic behavior of the entire wing section in real time. This increases labor costs and lengthens the testing cycle.

[0006] Secondly, fault analysis is difficult and inefficient. Limited by the severe fragmentation of test data records and the lack of real-time correlation, tracing the root cause of a fault and understanding its propagation path becomes extremely difficult. This is especially true for faults caused by multiple factors or with subtle manifestations, where manual troubleshooting can easily miss key links.

[0007] Furthermore, the reliability of critical data during transmission is difficult to guarantee. When network conditions are poor, there's no mechanism for quantitatively assessing real-time network communication quality, making it impossible to proactively detect and adjust transmission strategies. This can lead to delayed delivery of critical test instructions and loss or corruption of important status data, directly impacting the accuracy of test results.

[0008] The Chinese patent document with publication number CN118964184A discloses a vehicle testing system and testing method based on DDS simulation service. The vehicle testing system includes: an industrial computer, which is deployed with a test framework, and the test framework is used to respond to test requests and obtain test cases corresponding to the test requests; a DDS server, which is connected to the industrial computer through a switch, and the DDS server is deployed with a simulation service program, and the DDS server is used to receive test cases sent by the test framework and build a simulation service program for a virtual target test scenario based on the test cases; the hardware under test is connected to the DDS server in communication, and is used to receive the simulation service program for the virtual target test scenario and execute the test program based on the simulation service program; the test cases include one or more of factory flash test cases, self-test trial cases and OTA test cases. By building a virtual test scenario, the test requirements for various working conditions can be met, thereby improving the reliability of the test results.

[0009] The existing technology still has the above problems. Therefore, there is an urgent need for a DDS-based aircraft wing section testing method and system. Summary of the Invention

[0010] In view of the defects in the prior art, the object of the present invention is to provide a method and system for testing aircraft wing sections based on DDS.

[0011] According to the present invention, a DDS-based aircraft wing section testing method is provided, comprising:

[0012] Step S1: Establishing a data communication link between the test platform, the DDS Agent, and one or more hardware test modules through DDS;

[0013] Step S2: instructing the test platform to send a test instruction, which is forwarded to one or more hardware test modules via the DDS Agent; and instructing the hardware test modules to perform wing test operations to obtain real-time test data.

[0014] Step S3: the hardware test module feeds back real-time test data to the test platform via DDS;

[0015] Step S4: During the test command transmission and real-time test data feedback process, the DDS Agent calculates the communication quality Q value and dynamically adjusts the distribution strategy according to the Q value;

[0016] Step S5: Based on the adjusted distribution strategy, the test platform receives and analyzes the real-time test data, and performs fault diagnosis to determine the cause of the fault and locate the fault location on the digital twin model in the test platform.

[0017] Preferably, the step S1 includes:

[0018] Step S1.1: Configuring the test platform, the DDS Agent, and all hardware test modules within the same DDS domain; wherein there may be one or more hardware test modules, configured at specific locations on the wing as needed;

[0019] Step S1.2: Establish data communication links between the test platform, the DDS Agent and the hardware test modules in sequence through DDS. Specifically, establish data communication links with publish-subscribe relationships between the test platform and the DDS Agent, and between the DDS Agent and one or more hardware test modules based on predefined DDS topics, for forwarding test instructions and real-time test data. The real-time test data includes sequential test data and static test data.

[0020] Preferably, the step S4 includes:

[0021] Step S4.1: The DDS Agent monitors the key performance indicators (KPIs) of the data communication link in real time;

[0022] Step S4.2: normalize the monitored key performance indicators (KPIs) to obtain normalized key performance indicator scores;

[0023] Step S4.3: Preset weight coefficients, perform weighted summation on the normalized key performance indicator scores, and calculate the communication quality Q value;

[0024] Step S4.4: Set high quality threshold Q high and a low quality threshold Q low , determine the threshold interval of the Q value;

[0025] Step S4.5: adjusting the distribution strategy of the preset hierarchical data according to the interval of the Q value;

[0026] Among them, the key performance indicators KPIs include: end-to-end delay L, packet loss rate P, delay jitter J, effective throughput T eff and the proxy buffer occupancy rate B occ .

[0027] Preferably, step S4 includes:

[0028] Step S4.2 includes: normalizing the key performance indicators KPIs and calculating the normalized scores of the key performance indicators respectively; specifically,

[0029] Calculating latency scores : Set the maximum delay and ideal delay ; The scoring function is designed as the time delay score When the delay score is 1, the delay score is When it is 0, it decreases linearly in the middle:

[0030]

[0031] Calculating packet loss score : Set the maximum tolerable packet loss rate ; The scoring function is designed to score packet loss in When it is 1, the packet loss score is When 0:

[0032]

[0033] Calculating Jitter Score : Set the maximum jitter and ideal jitter , the scoring function is designed to score jitter in When the jitter score is 1, When it is 0, it decreases linearly in the middle:

[0034]

[0035] Calculating Throughput Score : Set the minimum throughput required for the current test task , the scoring function is designed to score the throughput in When the throughput score is 1, it decreases proportionally when it is lower than the minimum throughput:

[0036]

[0037] Calculating buffer scores : Set the buffer occupancy warning threshold and critical thresholds , the scoring function is designed as the number of buffer scores in 1, the buffer score is When it is 0, it decreases linearly in the middle:

[0038]

[0039] In the above function, min(·) means calculating the minimum value, and max(·) means calculating the maximum value;

[0040] Step S4.3 includes: calculating the Q value; specifically,

[0041] Set weight coefficient , let the weight coefficient satisfy ; then the Q value is

[0042]

[0043] in, The weight coefficients corresponding to the KPIs respectively;

[0044] Step S4.5 includes: selecting a distribution strategy for hierarchical data according to the interval of the Q value, and correspondingly adjusting at least one of the data publishing rate, the internal message cache of the DDS Agent, and modifying the service quality of the DDS; specifically,

[0045] when When , it indicates that the current network link status is excellent, then the adjustment strategy is: remove or relax the restrictions on data publishing rate, reduce cache occupancy, set the maximum data transmission throughput, and minimize the end-to-end communication delay;

[0046] when When , it indicates that the current network link quality has deteriorated to a certain extent. The adjustment strategy is: take preventive measures to reduce the frequency of publishing non-critical or delay-tolerant data, appropriately increase cache, improve QoS service quality, actively alleviate network pressure, and give priority to ensuring the transmission quality of important data;

[0047] when When , it indicates that the current network link quality has seriously deteriorated. The adjustment strategy is: adopt the most conservative transmission strategy, reserve the transmission channel for the highest priority data and ensure priority transmission, expand its internal message cache capacity to the preset maximum allowable value, and ensure the reliable transmission of the highest priority core instructions and key safety status data.

[0048] Preferably, step S5 includes:

[0049] Step S5.1: The test platform receives and analyzes real-time test data. If abnormal real-time test data is analyzed, the process proceeds to step S5.2. If normal real-time test data is analyzed, the process continues to receive data.

[0050] Step S5.2: The test platform extracts time series data features and static data features from the abnormal real-time test data;

[0051] Step S5.3: Inputting the time series data features into the time series processing branch in the pre-trained machine learning model for processing to obtain a time series processing result;

[0052] Step S5.4: Inputting the static data features into the static data processing branch in the pre-trained machine learning model for processing to obtain a static processing result;

[0053] Step S5.5: Fusion of the temporal processing result and the static processing result;

[0054] Step S5.6: Based on the fusion results, a prediction of the cause of the fault is output. At the same time, combined with real-time test data, the test platform simulates the actual condition of the wing to form a digital twin model, locates the fault location on the digital twin model, and highlights and annotates it.

[0055] According to the present invention, a DDS-based aircraft wing section testing system is provided, comprising:

[0056] A test platform, a DDS Agent, and one or more hardware test modules, wherein the test platform, the DDS Agent, and the hardware test modules are sequentially connected via a data communication link established by DDS;

[0057] Test platform: Built on digital twins, the test platform is used to load and run digital twin models of wing sections. It receives and analyzes real-time test data fed back by the hardware test module through the internal test platform DDS communication module, achieving real-time mapping between the wing's physical hardware status and the virtual model's status. It is used to monitor test data and generate visual alarms. It integrates intelligent fault diagnosis capabilities, using machine learning to automatically track and visually locate fault chains. It is also used to manage real-time test data and send test instructions.

[0058] DDS Agent: Connected to the test platform and hardware test modules, it forwards test instructions and real-time data bidirectionally based on DDS topics. It sets a communication quality evaluation (Q) value to assess network communication quality in real time and dynamically adjusts data distribution strategies based on the Q value to ensure data transmission reliability.

[0059] Hardware test module: It is configured at a specific position on the wing and is used to actually measure or control specific parameters of the aircraft wing according to the test instructions issued by the test platform, and feed back real-time test data to the test platform through DDS.

[0060] Preferably, the test platform includes:

[0061] The test platform's DDS communication module is used to create DDS publishers and subscribers, associate with the DDS Agent through DDS topics, and send, receive, and parse data based on a predefined standardized command frame format.

[0062] Wing Twin Model Module: This module is used to load, manage, and run the 3D digital twin model of the wing segment. It updates the model status based on real-time test data received by the test platform's DDS communication module, synchronizing the wing physical test with the virtual wing segment test status.

[0063] Monitoring and alarm module: used to verify the real-time test data feedback from the hardware test module and the preset data range, and trigger visual alarms when abnormalities are found;

[0064] Intelligent Fault Diagnosis Module: Based on machine learning technology, it is used to establish a mapping relationship between fault modes and data features by cleaning historical data and training models. When the monitoring and alarm module detects an anomaly, it automatically extracts the abnormal data features, inputs them into the pre-trained model for calculation, predicts the cause of the fault, and traces its propagation path. The fault source and related components are highlighted on the visual interface of the wing twin model module, and diagnostic suggestions are output.

[0065] Test management module: used for configuration, import, storage, and execution control of test cases, and is responsible for collecting test process data, alarm records, and diagnostic results.

[0066] Preferably, the intelligent fault diagnosis module includes:

[0067] Pre-trained fault diagnosis model submodule: This module trains a fault classification model using a hybrid input neural network model architecture based on historical time-series test data and static test data, along with their corresponding known fault cause labels. The core function of the fault classification model is to map input data features to specific fault cause categories, and to predict fault causes for abnormal test data.

[0068] Test data extraction submodule: During the test period, it is used to obtain the currently monitored abnormal test data from the monitoring and alarm module in real time and perform feature engineering processing, including calculating statistical features and performing normalization processing;

[0069] Fault location submodule: used to receive the fault category prediction result output by the pre-trained fault diagnosis model for the current test data feature vector; when the prediction result indicates a specific fault, it is used to parse and map this fault category information to the specific components or three-dimensional spatial positions in the digital twin model of the aircraft wing section, trigger the wing twin model module, and highlight and mark the located components or areas on the visual interface.

[0070] Preferably, the hardware testing module includes:

[0071] Hardware test DDS communication module: used to handle DDS connection and data transmission and reception, sending data in a predefined standard format; according to the test instructions, it is used to identify whether a specific field is relevant to itself, and if so, perform the test; if not, discard the message;

[0072] Physical measurement and control module: includes sensors and actuators for directly interacting with the aircraft wing sections, performing measurement or control operations, and feeding back real-time measurement data;

[0073] Data preprocessing module: Filters and calibrates the raw data from the physical measurement and control module, and then sends it through the hardware test DDS communication module;

[0074] It also includes multiple test channels, each test channel corresponds to a different control word and is used to perform different test operations.

[0075] Preferably, the DDS Agent includes:

[0076] Data forwarding module: used to bidirectionally forward test platform and hardware test module instructions and real-time test data. It resides in the same DDS domain as the test platform and hardware test modules. Hardware test modules of the same category are associated with the same topic as the DDS Agent. When the DDS Agent receives a test instruction from the test platform, it forwards data to all hardware test modules associated with the topic simultaneously.

[0077] Q-value real-time calculation module: used to calculate the communication quality evaluation index Q-value; used to quantitatively evaluate communication quality, and to normalize the key performance indicators by continuously monitoring the key performance indicators of the communication link;

[0078] Communication strategy dynamic adjustment module: According to the preset Q value threshold interval, it is used to automatically execute the corresponding adjustment distribution strategy. The adjustment content includes: adjusting the data publishing rate, adjusting the Agent internal message cache size, and dynamically modifying the DDS QoS service quality.

[0079] Compared with the prior art, the present invention has the following beneficial effects:

[0080] 1. The test platform constructed based on digital twins in the present invention enables testers to intuitively and comprehensively monitor the integration status and dynamic behavior of aircraft wing sections through real-time mapping and a visual interface, and to promptly detect problems through visual alarms.

[0081] 2. The present invention integrates an intelligent fault diagnosis module, uses machine learning to automatically track fault chains and locate root causes, and visualizes the results on a digital twin model, greatly shortening troubleshooting time and improving diagnostic accuracy.

[0082] 3. The present invention innovatively introduces the Q index on the DDS Agent side to evaluate network quality in real time, and dynamically adjusts the communication strategy based on the Q value. It can proactively respond to network fluctuations, effectively avoid the loss or delay of key test data, and ensure the reliability of test results.

[0083] 4. The present invention integrates the system architecture to achieve unified, efficient and intelligent management of distributed hardware test modules, thereby improving the level of test automation and overall efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] Other features, objects and advantages of the present invention will become more apparent upon reading the detailed description of non-limiting embodiments with reference to the following drawings:

[0085] Figure 1 Schematic diagram of a flow chart of a DDS-based aircraft wing section testing method of the present invention;

[0086] Figure 2 Schematic diagram of the composition of a DDS-based aircraft wing section testing system according to Example 1 of the present invention;

[0087] Figure 3 This is a block diagram of a test platform for an aircraft wing section test system according to embodiment 1 of the present invention;

[0088] Figure 4 This is a block diagram of the communication protocol format of Example 1 of the present invention. DETAILED DESCRIPTION

[0089] The present invention will be described in detail below with reference to specific embodiments. The following examples will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those skilled in the art, several changes and improvements can be made without departing from the scope of the present invention. These all fall within the scope of protection of the present invention.

[0090] The purpose of the present invention is to provide a DDS-based aircraft wing section testing method and system to solve the problems of complicated functional integrity and state consistency testing steps, low testing efficiency, and lack of efficient and unified test management in existing wing testing scenarios.

[0091] A DDS-based aircraft wing section testing method, comprising:

[0092] Step S1: Establishing a data communication link between the test platform, the DDS Agent, and one or more hardware test modules through DDS;

[0093] Step S2: The test platform sends a test instruction, which is forwarded to one or more hardware test modules via the DDS Agent. Simultaneously, the hardware test modules are instructed to perform wing test operations to obtain real-time test data.

[0094] Step S3: The hardware test module feeds back real-time test data to the test platform through the DDS communication environment;

[0095] Step S4: During the test instruction transmission and real-time test data feedback process, the DDS Agent calculates the communication quality Q value based on the DDS communication environment and dynamically adjusts the data distribution strategy of the data distribution service according to the Q value;

[0096] Step S5: Based on the adjusted distribution strategy, the test platform receives and analyzes real-time test data and performs machine learning-based fault diagnosis. The fault diagnosis processes the timing test data and static test data separately, and integrates the processing results to determine the cause of the fault and locate the fault location on the digital twin model in the test platform.

[0097] Specifically, step S1 includes:

[0098] Step S1.1: Configure the test platform, DDS Agent, and all hardware test modules in the same DDS domain. There may be one or more hardware test modules, which are configured at specific locations on the wing as needed.

[0099] Step S1.2: Establish data communication links between the test platform, the DDS Agent, and the hardware test modules in sequence through DDS. Specifically, establish data communication links with publish-subscribe relationships between the test platform and the DDS Agent, and between the DDS Agent and one or more hardware test modules based on predefined DDS topics to forward test instructions and real-time test data. The real-time test data includes sequential test data and static test data.

[0100] Specifically, step S4 includes:

[0101] Step S4.1: The DDS Agent monitors the key performance indicators (KPIs) of the data communication link in real time;

[0102] Step S4.2: normalize the monitored key performance indicators (KPIs) to obtain normalized key performance indicator scores;

[0103] Step S4.3: Preset weight coefficients, perform weighted summation on the normalized key performance indicator scores, and calculate the communication quality Q value;

[0104] Step S4.4: Set high quality threshold Qhigh and a low quality threshold Q low , determine the threshold interval of the Q value;

[0105] Step S4.5: Based on the Q value interval, adjust the distribution strategy of the preset hierarchical data; the adjustment action includes: adjusting the data publishing rate, adjusting the internal message buffer size of the DDS Agent, and modifying at least one of the DDS Quality of Service (QoS) strategies.

[0106] Among them, the key performance indicators KPIs include: end-to-end delay L, packet loss rate P, delay jitter J, effective throughput T eff and the proxy buffer occupancy rate B occ .

[0107] Specifically, step S4 includes:

[0108] Step S4.2 includes: normalizing the key performance indicators KPIs and calculating the normalized scores of the key performance indicators respectively; specifically,

[0109] Calculating latency scores : Set the maximum delay and ideal delay ; The scoring function is designed as the time delay score When the delay score is 1, the delay score is When it is 0, it decreases linearly in the middle:

[0110]

[0111] Calculating packet loss score : Set the maximum tolerable packet loss rate ; The scoring function is designed to score packet loss in When it is 1, the packet loss score is When 0:

[0112]

[0113] Calculating Jitter Score : Set the maximum jitter and ideal jitter , the scoring function is designed to score jitter in When the jitter score is 1, When it is 0, it decreases linearly in the middle:

[0114]

[0115] Calculating Throughput Score : Set the minimum throughput required for the current test task , the scoring function is designed to score the throughput in When the throughput score is 1, it decreases proportionally when it is lower than the minimum throughput:

[0116]

[0117] Calculating buffer scores : Set the buffer occupancy warning threshold and critical thresholds , the scoring function is designed as the number of buffer scores in 1, the buffer score is When it is 0, it decreases linearly in the middle:

[0118]

[0119] In the above function, min(·) means calculating the minimum value, and max(·) means calculating the maximum value;

[0120] Step S4.3 includes: calculating the Q value; specifically,

[0121] Set weight coefficient , let the weight coefficient satisfy ; then the Q value is

[0122]

[0123] in, They correspond to the weight coefficients of KPIs respectively; the weight coefficients are selected according to the importance of the key performance indicators.

[0124] Step S4.5 includes: selecting a distribution strategy for hierarchical data according to the interval of the Q value, and correspondingly adjusting at least one of the data publishing rate, the internal message cache of the DDS Agent, and modifying the service quality of the DDS; specifically,

[0125] when When , it indicates that the current network link status is excellent, then the adjustment strategy is: remove or relax the restrictions on data publishing rate, reduce cache occupancy, set the maximum data transmission throughput, and minimize the end-to-end communication delay;

[0126] when When , it indicates that the current network link quality has deteriorated to a certain extent. The adjustment strategy is: take preventive measures to reduce the frequency of publishing non-critical or delay-tolerant data, appropriately increase cache, improve QoS service quality, actively alleviate network pressure, and give priority to ensuring the transmission quality of important data;

[0127] when When , it indicates that the current network link quality has seriously deteriorated. The adjustment strategy is: adopt the most conservative transmission strategy, reserve the transmission channel for the highest priority data and ensure priority transmission, expand its internal message cache capacity to the preset maximum allowable value, and ensure the reliable transmission of the highest priority core instructions and key safety status data.

[0128] Specifically, step S5 includes:

[0129] Step S5.1: The test platform receives and analyzes real-time test data. If abnormal real-time test data is analyzed, the process proceeds to step S5.2. If normal real-time test data is analyzed, the process continues to receive data.

[0130] Step S5.2: The test platform extracts time series data features and static data features from the abnormal real-time test data;

[0131] Step S5.3: Input the time series data features into the time series processing branch (including the long short-term memory network LSTM) in the pre-trained machine learning model for processing to obtain the time series processing results;

[0132] Step S5.4: Input the static data features into the static data processing branch (including at least one fully connected layer) in the pre-trained machine learning model for processing to obtain a static processing result;

[0133] Step S5.5: Fusing the temporal processing results with the static processing results;

[0134] Step S5.6: Based on the fusion results, a prediction of the cause of the fault is output. Combined with real-time test data, the test platform simulates the actual wing condition to form a digital twin model. The fault location is located on the digital twin model and highlighted and annotated. In a preferred embodiment, a normal wing parameter range can be set, and the digital twin model is compared with the normal wing parameter range in real time. Locations in the digital twin model that fall outside the normal parameter range are marked as fault locations. Simultaneously, a fault library is established. Based on the analysis of the portions that fall outside the parameter range, the corresponding faults are matched in the fault library and the cause of the fault is inferred.

[0135] The present invention also provides a DDS-based aircraft wing section testing system. The DDS-based aircraft wing section testing system can be implemented by executing the process steps of the DDS-based aircraft wing section testing method. That is, those skilled in the art can understand the DDS-based aircraft wing section testing method as a preferred implementation of the DDS-based aircraft wing section testing system.

[0136] A DDS-based aircraft wing section test system includes: a test platform, a DDS Agent agent, and one or more hardware test modules. The test platform, the DDS Agent agent, and the hardware test modules are sequentially connected via a data communication link established by DDS; the test platform: the test platform is built based on digital twins and is used to load and run the digital twin model of the wing section. It receives and parses the real-time test data fed back by the hardware test module through the internal test platform DDS communication module, thereby realizing real-time mapping between the physical hardware status of the wing and the virtual model status; it is used to monitor test data and issue visual alarms; it integrates intelligent fault diagnosis capabilities and is used to automatically track and visually locate fault chains using machine learning; it is used to manage real-time test data; and it is also used to send test instructions; DDS Agent: connected to the test platform and hardware test module, based on the DDS topic, used for bidirectional forwarding of test instructions and real-time data; sets the communication quality evaluation Q value to evaluate the network communication quality in real time, and dynamically adjusts the data distribution strategy according to the Q value to ensure data transmission reliability; Hardware test module: configured at a specific position on the wing, used to actually measure or control the specific parameters of the aircraft wing according to the test instructions issued by the test platform, and feeds back real-time test data to the test platform through DDS.

[0137] Specifically, the test platform includes: Test platform DDS communication module: used to create DDS publishers and subscribers, and DDS The agent side is associated through DDS topics and sends, receives and parses data based on a predefined standardized instruction frame format; the wing twin model module is used to load, manage and run the three-dimensional digital twin model of the wing segment, and update the model status according to the real-time test data received by the test platform DDS communication module, so as to synchronize the wing physical test with the virtual wing segment test status; the monitoring and alarm module is used to verify the real-time test data and data range based on the hardware test module operation feedback, and trigger a visual alarm when an anomaly is found; the intelligent fault diagnosis module is based on machine learning technology and is used to establish a mapping relationship between fault mode and data features by cleaning historical data and training models; when the monitoring and alarm module detects an anomaly, it automatically extracts the abnormal data features, inputs the pre-trained model for calculation, predicts the cause of the fault, and traces its propagation path, highlights the fault source and related components on the visual interface of the wing twin model module, and outputs diagnostic suggestions; the test management module is used for the configuration, import, storage, and execution control of test cases, and is responsible for collecting test process data, alarm records, and diagnostic results.

[0138] Specifically, the intelligent fault diagnosis module includes: a pre-trained fault diagnosis model sub-module: a fault classification model trained based on historical time series test data and static test data and their corresponding known fault cause labels through a hybrid input neural network model architecture; the core function of the fault classification model is to establish a mapping relationship between input data features and specific fault cause categories, and give a predicted fault cause for abnormal test data; a test data extraction sub-module: used to obtain the currently monitored abnormal test data from the monitoring and alarm module in real time during the test, and perform feature engineering processing, including calculating statistical features and normalization processing; a fault location sub-module: used to receive the fault category prediction result output by the pre-trained fault diagnosis model for the current test data feature vector; when the prediction result indicates a specific fault, it is used to parse and map this fault category information to a specific component or three-dimensional spatial position in the digital twin model of the aircraft wing section, triggering the wing twin model module to highlight and mark the located components or areas on the visual interface.

[0139] Specifically, the hardware test module includes: a hardware test DDS communication module: used to handle the connection with DDS and data transmission and reception, and send data in a predefined standard format; according to the test instructions, it is used to identify whether a specific field is related to itself, and if so, the test is performed, and if not, the message is discarded; a physical measurement and control module: including sensors and actuators, used to directly interact with the aircraft wing sections, perform measurement or control operations, and feed back real-time measurement data; a data preprocessing module: filters and calibrates the raw data from the physical measurement and control module, and then sends it through the hardware test DDS communication module; it also includes multiple test channels, each test channel corresponds to a different control word, and is used to perform different test operations.

[0140] Specifically, the DDS Agent includes: a data forwarding module: used to bidirectionally forward test platform and hardware test module instructions and real-time test data, and is in the same DDS domain as the test platform and hardware test module; hardware test modules of the same category are associated with the same topic of the DDS Agent. When the DDS Agent receives the test instructions from the test platform, it is used to forward data to all hardware test modules of the associated topics at the same time; a Q-value real-time calculation module: used to quantitatively evaluate the communication quality, by continuously monitoring the key performance indicators of the communication link, including end-to-end delay L, packet loss rate P, delay jitter J, etc., to normalize the key performance indicators; and to calculate the communication quality evaluation index Q value; a communication strategy dynamic adjustment module: based on the preset Q-value threshold range, it is used to automatically execute the corresponding adjustment distribution strategy. The adjustment content includes: adjusting the data publishing rate, adjusting the Agent internal message cache size, and dynamically modifying the DDS QoS service quality.

[0141] Example 1

[0142] Combine Figure 2-Figure 3 As shown in FIG, a DDS-based aircraft wing test system 1 is provided, comprising: a test platform 10 deployed on an industrial computer, a DDS Agent, and a hardware test module 30. The test platform 10 is the core control and analysis unit of the system, and its specific composition is also referred to Figure 3 As shown, it is used for sending and receiving instruction data for parsing, updating the virtual wing twin model, monitoring test data, intelligent diagnosis of test fault causes and test management, which correspond to the test platform DDS communication module 101, the wing twin model module 102, the monitoring and alarm module 103, the intelligent fault diagnosis module 104 and the test management module 105 respectively. Among them, the test management module 105 includes use case configuration reading, which is used to serialize and deserialize the use case configuration. It can flexibly modify the original use case according to specific needs, improve the reusability of the test case, and further reduce the test time consumption; in a preferred embodiment, a test report can also be generated. The test report mainly includes the test time, the test executor, the test items performed, the test conclusions, and the specific problems found in the test, etc., and specifically displays the test report; test case import and export functions.

[0143] According to the above embodiment, relying on the test management module 105, the corresponding test case can be selected according to the test request, thereby improving the test efficiency. Moreover, through visual operation, the test case can be added and modified, which improves the scalability and maintainability of the test case.

[0144] The DDS Agent agent and the test platform are in the same DDS Domain to ensure the establishment of a communication connection. In the test process, the data forwarding module 201 is used to bidirectionally forward the test platform and hardware test module instructions and data. The Q value real-time calculation module 203 and the communication strategy dynamic adjustment module 202 are combined to adjust the data distribution strategy in real time according to the network communication status. The hardware test module 30 establishes a communication connection with the DDS Agent agent through the internal hardware test DDS communication module, receives and executes relevant instructions of the test platform, and uploads the test measurement results through the DDS Agent agent. The hardware test module 30 also includes a data preprocessing module 302, which is used to filter and calibrate the raw data from the internal physical measurement and control module after local processing, and then send it through the hardware test DDS communication module. Further, in a preferred embodiment, the communication format 40 between the test platform and the hardware test module meets the specified string format, such as Figure 4As shown, a specific embodiment is as follows: the control protocol frame header is 55, the frame tail is AA, the first 2 bits after removing the frame header and tail are the test category, the test category field is offset to the right by 2 bits to be the hexadecimal test module number, and a maximum of 256 test modules are supported; the test module number is offset to the right by 2 bits to be the specific operation control word.

[0145] The heartbeat protocol frame header is 66 and the frame trailer is AA. After removing the header and trailer, the first two bits are the test category. The test category field is offset 2 bits to the right for the hexadecimal test module number, which supports up to 256 test modules. The test module number field is offset 2 bits to the right for the online status.

[0146] The functions of the specific modules of the system are as follows: an industrial computer, which deploys a test platform. The test platform is used to send and receive hardware test instructions, synchronize the status of the wing twin model, perform intelligent fault diagnosis, monitor test data and alarms, and manage test cases and test reports; a hardware test module, which contains a physical measurement and control module for precise measurement and control of wing rotation angles, cable connectivity, fuel tank oil temperature, etc.; a DDS Agent agent, which is in the same DDS domain as the test platform DDS communication module and is used to bidirectionally forward instructions and data from the test platform and hardware test module.

[0147] The DDS-based aircraft wing section test system provided in accordance with the above technical points includes an industrial computer, a DDSAgent agent, and a hardware test module. The test platform is deployed on the industrial computer, which includes the test platform DDS communication module, the wing twin model module, the monitoring and alarm module, the intelligent fault diagnosis module, and the test management module. The DDSAgent agent, deployed on the industrial computer or other supporting hardware devices, proxies all test data from the hardware test module and associates it with a corresponding topic on the test platform, ensuring real-time command and data exchange between the test platform and the hardware test module. The hardware test module monitors or measures the aircraft wing status, uploads the status and values of the test items to the test platform in a specified format, and executes the test platform's commands. These commands can include abnormal operating conditions and edge cases, addressing the functional integrity and state consistency testing requirements in different scenarios and improving the reliability of test results. Furthermore, by integrating the test platform's DDS communication module within the test platform, a visual interface automatically completes various data operation logic operations, from command generation and transmission to parsing and verification, significantly reducing testing time. Furthermore, by saving and importing the combined use of test cases, testers can avoid manual repetitive operations, thereby improving testing efficiency.

[0148] Furthermore, the test platform also includes a wing twin model module. The wing twin is synchronized with the actual hardware test module based on parsed hardware data. After receiving test module data via DDS, it is parsed and the twin state in the rendered scene is updated in real time based on the parsed status or measured values, allowing testers to intuitively identify anomalies during testing. The test platform's embedded DDS communication module needs to be integrated with the data processing method for easy access.

[0149] Using this technical approach, hardware test modules of the same category are associated with the same topic. When the DDS agent receives a command from the test platform, it simultaneously forwards the data to all test modules associated with that topic. The hardware test module identifies whether the command data is relevant to itself based on the test module field in the command data. If so, it proceeds to the next step; otherwise, it discards the message. When the hardware test module generates test data that needs to be forwarded to the test platform, it still follows the above protocol format.

[0150] Based on the above technical means, the aircraft wing section test system can realize test case configuration and automatic execution, data analysis and real-time update of the aircraft twin status in the scenario according to the specific test scenario, thereby improving test efficiency.

[0151] A DDS-based aircraft wing segment testing method is also provided. The method includes: first, establishing a data communication link via DDS between a test platform 10 (via its test platform DDS communication module 101), a DDS agent 20, and one or more hardware test modules 30 (via their hardware test DDS communication modules 301). During testing, the test platform 10 (initiated by the test management module 105 and constructed via the internal test platform DDS communication module) sends a test command. The DDS agent 20 (primarily via its data forwarding module 201) efficiently forwards the command to the hardware test module 30. The hardware test module 30 (with its physical measurement and control module 303) then performs wing testing operations and acquires real-time test data. The acquired real-time data is then reliably fed back to the test platform 10 by the hardware test module 30 via the DDS agent 20. Throughout the entire command and data exchange process, the DDS Agent 20 (through its real-time Q-value calculation module 203 and dynamic communication strategy adjustment module 202) continuously evaluates and optimizes DDS communication quality and data distribution strategies. Finally, the test platform 10 receives and analyzes the real-time test data feedback (performed by the monitoring and alarm module 103, etc.), performs machine learning-based fault diagnosis provided by the intelligent fault diagnosis module 104 to determine the cause of the fault, and synchronously updates the diagnostic results and real-time status to the digital twin model managed by the wing twin model module 102, enabling visual fault location.

[0152] Furthermore, in a preferred embodiment, the test platform constructs the specified format instruction according to the tester's visual operation, including the following steps: determining the first 4 bits according to the selected test category and test module; splicing the specified operation control words according to the specific test requirements and operations, and calling the test platform DDS communication module interface to send messages; finally, saving or deleting the test case of this operation according to the test needs.

[0153] Furthermore, the saved test cases include normal scenario test cases and abnormal scenario test cases, which are used to more comprehensively cover the test requirements and improve the test coverage and test reliability.

[0154] Example 2

[0155] In a preferred embodiment, the data distribution process of the dynamic adjustment data distribution service is as follows:

[0156] Data forwarding: Bidirectionally forwards commands and test data between the test platform and hardware test modules. Commands and data are forwarded bidirectionally between the test platform and hardware test modules, all within the same DDS domain. This forwarding is based on DDS topics. Hardware test modules of the same category are associated with the same topic as the DDS agent. When the DDS agent receives a command from the test platform, it simultaneously forwards the data to all hardware test modules associated with that topic. The hardware test modules identify the command based on specific fields in the command. If so, they execute the test; otherwise, they discard the message.

[0157] Real-time calculation of Q value: This is used to quantitatively evaluate the communication quality. By continuously monitoring the key performance indicators (KPIs) of the communication link, including end-to-end delay (L), packet loss rate (P), delay jitter (J), etc., and normalizing these indicators, according to the preset weights Perform weighted summation to calculate the real-time communication quality evaluation index ,in is the normalized key performance indicator score;

[0158] Dynamic communication strategy adjustment: Another innovative feature of this invention is the dynamic optimization of communication by receiving the Q-value output from the real-time Q-value calculation module. Based on preset Q-value thresholds, the system automatically executes corresponding adjustment strategies, including adjusting the data publishing rate, adjusting the Agent's internal message buffer size, and dynamically modifying the DDS QoS policy, to ensure data transmission reliability during wing section testing.

[0159] The main steps are as follows:

[0160] 1. Select and measure key performance indicators

[0161] To comprehensively evaluate communication quality, the following KPIs that have a significant impact on aircraft wing test data interaction were selected for real-time monitoring:

[0162] End-to-end latency (L): This reflects the time it takes for a command or data to be transmitted from the sender to the receiver, measured in milliseconds. Low latency is crucial for wing control commands or high-frequency state feedback. L is measured using the built-in timestamp mechanism of DDS.

[0163] Packet Loss Rate (P): This indicates the proportion of packets lost during transmission and is a key indicator of connection reliability. It is accurately calculated using the confirmation mechanism provided by DDS's RELIABLE QoS, or by comparing the test platform and hardware test module serial numbers under BEST_EFFORT QoS. P is a dimensionless ratio, P ∈ [0,1].

[0164] Delay jitter (J): The variation in end-to-end delay, measured by the standard deviation of the end-to-end delay L, is expressed in milliseconds (ms).

[0165] Effective throughput ( T eff ): It indicates the amount of effective data successfully transmitted per unit time, in Mbps. It can be used to determine whether the current test transmission rate meets the test requirements by monitoring the actual test data traffic.

[0166] Agent buffer occupancy rate ( B occ ): Indicates the current fill level of the DDS Agent's internal buffer queue for sending and receiving DDS messages, expressed as a ratio (0-1). High occupancy is a direct indicator of network congestion and is monitored in real time by the Agent's internal monitoring mechanism.

[0167] 2. Normalize key performance indicators

[0168] Since the dimensions, value ranges, and evaluation directions (larger values are better / smaller values are better) of each KPI are different, they need to be uniformly mapped to a dimensionless scoring interval of [0, 1], where 1 represents the best communication quality and 0 represents the worst communication quality. The following normalization function is used: , and set the threshold value in combination with the specific requirements of aircraft wing testing:

[0169] Latency score ( ) : Set the maximum acceptable delay and ideal delay The scoring function is designed to be When 1, When it is 0, it decreases linearly in the middle:

[0170]

[0171] Packet Loss Score ( ) : Set the maximum tolerable packet loss rate The scoring function is designed to be When 1, When 0:

[0172]

[0173] Jitter score ( ) : Similar to delay, set the maximum acceptable jitter and ideal jitter :

[0174]

[0175] Throughput score ( ): Set the minimum throughput required for the current test task The scoring function is When it is 1, it decreases proportionally when it is lower than the demand:

[0176]

[0177] Buffer score ( ) : Set the buffer occupancy warning threshold and critical thresholds The scoring function is When 1, When it is 0, it decreases linearly in the middle:

[0178]

[0179] In the above function, min(·) means calculating the minimum value, and max(·) means calculating the maximum value. L, P, J, T, B OCC They represent the current measured delay, packet loss rate, jitter, throughput, and buffer size respectively.

[0180] 3. Weighted fusion calculation of comprehensive Q value

[0181] The above normalized KPI scores are weighted and summed according to their relative importance to communication reliability in the aircraft wing section test scenario to obtain the final comprehensive communication quality evaluation index. . Weight coefficient ( ) is set by domain experts and satisfies , It is the normalized score of the corresponding KPI.

[0182]

[0183] In aircraft wing testing, low latency of control commands ( ) and zero packet loss ( ) has the highest priority, so and Assign higher weights (e.g. ); For sensor data that requires precise time synchronization, jitter ( ) Also need to consider (e.g. Throughput ( The importance of ) depends on the amount of data currently being transmitted. In the data-intensive transmission phase should be increased (e.g. ); Buffer occupancy ( ) as an early indicator of congestion and also has a certain weight (for example, ).

[0184] Should value( )Depend on The value real-time calculation module calculates and updates and provides it to the communication strategy dynamic adjustment module.

[0185] 4. Dynamic adjustment of communication strategies

[0186] Real-time reception of communication quality evaluation indicators output from the Q value real-time calculation module , and the Values and pre-set high quality thresholds and low quality threshold To compare, according to The range of values automatically triggers and executes corresponding hierarchical communication resource management and service quality assurance strategies. These strategies are designed to dynamically optimize data publishing rates, internal resource allocation on the DDSAgent agent, and the configuration of Quality of Service (QoS) parameters for the DDS communication protocol. The ultimate goal is to prioritize the real-time, integrity, and high reliability of critical data transmission during aircraft wing section testing under various network conditions. Specific adjustment strategies implemented include, but are not limited to, the following three main levels:

[0187] Level 1: High-quality communication strategy ( )

[0188] When the Q value output by the Q value real-time calculation module reaches or exceeds the preset high quality threshold When [number of connections are available] is reached, the current network link is in excellent condition, characterized by high bandwidth, low latency, low packet loss, and low jitter. The strategic goal at this point is to maximize data throughput, minimize end-to-end communication latency, and fully utilize excellent network resources while ensuring basic communication reliability.

[0189] Specific adjustment actions:

[0190] Data release rate optimization: Remove or relax restrictions on data release rates. Allow the test platform (or when forwarded by an agent) to release test instructions and data at the maximum nominal rate specified by its application layer requirements or hardware capabilities.

[0191] Internal cache resource management: Maintain the cache space within the DDS Agent, which is used to temporarily store DDS messages to be forwarded or processed, at a preset standard or minimum baseline level to reduce unnecessary memory usage and avoid additional processing delays that may be introduced by an overly large cache.

[0192] DDS QoS Policy Configuration: For most non-critical data flows, the best-effort QoS (RELIABILITY QoS = BEST EFFORT) is preferred. This configuration minimizes protocol overhead (no confirmation messages or retransmissions), helping to reduce end-to-end latency. Only core control commands or security-related data flows, where the application layer explicitly requires guaranteed delivery, are retained or configured with the reliable delivery QoS (RELIABILITY QoS = RELIABLE).

[0193] Level 2: Medium quality communication strategy ( )

[0194] When the Q value output by the Q value real-time calculation module is between the low quality threshold With high quality threshold If the link quality is between 0 and 100, it indicates that the current network link quality has deteriorated to a certain extent. The strategic goal at this time is to take preventive measures to proactively alleviate network pressure, enhance the system's tolerance to network fluctuations, prioritize the transmission quality of important data, and prevent communication quality from further deteriorating to a low level.

[0195] Specific adjustment actions:

[0196] Selective Data Rate Adjustment: Enables rate control based on data priority or type. Based on the pre-set importance of data (identified by DDS topics), the rate of publishing of non-critical or latency-sensitive data is reduced, freeing up network bandwidth and processing resources for critical data.

[0197] Dynamic expansion of internal cache: Appropriately increase the internal message cache capacity of the DDS Agent. The increased cache space can more effectively absorb the disordered arrival timing or temporary backlog of data packets caused by network jitter or instantaneous congestion, thereby reducing the risk of actively discarding data due to internal cache overflow of the Agent.

[0198] Moderately enhance the DDS QoS policy: Based on data criticality assessment, the service quality of some important data flows is selectively improved. For example, for the more important test instruction topic, the RELIABILITY QoS policy on the publisher is dynamically adjusted from BEST_EFFORT to RELIABLE, enabling the data confirmation and retransmission mechanism at the DDS level to improve delivery assurance.

[0199] Level 3: Low-quality communication strategy ( )

[0200] when The output of the real-time calculation module Values below the preset low quality threshold , indicating that the current network link quality has seriously deteriorated, with significant increases in latency, high packet loss rates, high jitter, or signs of severe congestion. The strategy adjustment goal at this time is to adopt the most conservative transmission strategy in extremely harsh network environments, sacrificing overall system throughput and the real-time performance of non-critical data in order to ensure the reliable transmission of the highest-priority core instructions and critical safety status data, thereby ensuring the basic controllability and security of the test system.

[0201] Specific adjustment actions:

[0202] Strict Rate Control and Priority Guarantee: Implements enhanced, priority-based traffic scheduling and rate limiting. This prioritizes and reserves transmission channels for the highest-priority data. Simultaneously, it significantly reduces the rate of medium-priority data and temporarily suspends all low-priority data transmission until network quality returns.

[0203] Maximize internal cache utilization: Within the scope of the memory hardware resources of the DDS Agent, expand its internal message cache capacity to the preset maximum allowable value, maximize buffering of data backlogs caused by severe network congestion, and provide space for subsequent DDS retransmission or data transmission after network recovery.

[0204] Enforce high-reliability DDS QoS: Forcibly set the RELIABILITY QoS policy to RELIABLE for all publishers of critical or important data flows. Also, set the HISTORY QoS policy to KEEP_ALL to ensure that all historical samples are stored for retransmission. Adjust the RESOURCE_LIMITS QoS policy to increase the memory resources used for sample storage. These adjustments sacrifice transmission efficiency and increase latency in exchange for the highest probability of data delivery.

[0205] Example 3

[0206] In a preferred embodiment, the fault cause determination and fault location process is as follows:

[0207] S100: Build a data set based on historical test data and fault data analysis; load the digital twin model of the aircraft wing section into the test platform and configure test cases.

[0208] S101: Divide the historical test data of the wing section into sequential test data and static test data.

[0209] Timing test data represents the preset time window W (Include T A specified physical quantity that is continuously measured and changed within a time step. For example, the sequence of wing rotation angles , fuel tank oil temperature series Etc. These data constitute a time series feature matrix ,in is the number of time series features, that is, the number of test types. Represents the set of real numbers.

[0210] Static test values refer to discrete states or values that remain unchanged within the same time window W or are measured only at a specific moment. For example, the on / off status of a preset channel cable , Initial oil temperature at the start of the test Etc. These data constitute a static feature vector ,in is the number of static features.

[0211] S102: Extracting data segments from historical test logs and fault analysis data.

[0212] Each data fragment By the corresponding time window Time series data within and static data By analyzing the failure data for each historical sample Mark the confirmation tag , represents the system status within the time window (normal or specific fault cause category), and there are K categories in total (1 normal category + K-1 fault categories).

[0213] S103: Perform data set construction and preprocessing.

[0214] Combine all labeled data fragment samples to form a training data set ,in N is the total number of samples. Then normalize it to X ts and The values in are mapped to the interval [0,1] using min-max scaling, i.e. If there are missing values in the data, the adjacent bit means are used to fill them.

[0215] S200: Building a Hybrid Input Neural Network Model Architecture

[0216] According to the characteristics of timing test data and static test data, a neural network architecture with two independent input branches is adopted.

[0217] S201: Building a sequential processing branch

[0218] The input is a normalized time series data matrix: The core layer uses the existing Long Short-Term Memory (LSTM) network to build temporal dependencies in the data, which includes a forget gate , input gate and output gate , cell state , the core calculation process of state update is as follows:

[0219] Forget Gate : Determine the cell state from the previous moment The information discarded.

[0220]

[0221] Input Gate : Decide what new information will be stored in the cell state It consists of two parts: one is to decide which values to update , the other is to create a candidate value vector .

[0222]

[0223]

[0224] Cell state : Combine the results of the forget gate and the input gate to update the cell state .

[0225]

[0226] Output Gate : Determine the current cell state Which parts of Output.

[0227]

[0228]

[0229] in, is the time step The input feature vector of is the hidden state at the current moment, is the hidden state vector at the previous time step ( initialized to zero vector), is the cell state vector at the previous time step ( initialized to zero vector), is the size of the LSTM hidden layer. and They are the weight matrix and bias vector corresponding to the forget gate, input gate, cell state, and output gate respectively. Represents the Sigmoid activation function (output range (0,1), ⊙ represents element-wise multiplication, and tanh is the hyperbolic tangent activation function.

[0230] Output: After After processing time steps, the time series processing branch outputs a sequence that can represent the entire input sequence. Vector of Information . That is, the hidden state of the last time step ,Right now .vector Encapsulates key dynamic features extracted from time series data.

[0231] S202: Building a static data processing branch

[0232] The input is a normalized static feature vector In order to transform the feature space and extract more abstract feature representations, Input to two fully connected layers, using ReLU activation function ( ):

[0233]

[0234]

[0235]

[0236]

[0237] in and are the weight matrices and bias vectors of the two fully connected layers respectively. is the result of linear calculation, is the output after the first layer is activated. This branch finally outputs a vector , represents the processed static information, where is the output dimension of the second fully connected layer, is the set of real numbers.

[0238] S203: Fusion of Time Series Data Features and Static Data Features

[0239] The output vector of the timing processing branch Output vector of static data processing branch Splicing is performed to obtain a combined feature vector that integrates two types of data information ([;] indicates vector concatenation):

[0240]

[0241] S204: Build hybrid model prediction output

[0242] The fused feature vector Input to the fully connected layer for final fault classification. The fully connected layer has Output neurons ( is the total number of categories, including normal and Output Input into the Softmax function and convert it into the corresponding Probability distribution of categories , the conversion method is:

[0243]

[0244] in Indicates that the model predicts that the current input sample belongs to the category The probability of is the output of the fully connected layer j elements, and satisfy , exp represents the exponential function.

[0245] S300: Training a Hybrid Neural Network Model

[0246] By using the labeled training data set Learn and adjust all trainable parameters of the entire hybrid neural network model (Contains the weight matrices of all LSTM layers and fully connected layers and the bias vector ), so that the model's prediction results for the training samples are as consistent as possible with the actual fault labels. At the same time, the classification cross entropy is selected as the loss function. For example, for samples, K The average loss calculation formula for small batch data of categories is as follows:

[0247]

[0248] in, It is The true labels of the samples (the true categories Corresponding , the rest are 0), The model is Samples are predicted as categories The probability of Softmax output elements). Adopt Adam (AdaptiveMomentEstimation) optimizer and back propagation algorithm to calculate the loss function of model parameters. Gradient , and iteratively update the model parameters based on this gradient , to minimize the loss function The model training process will traverse the training data set for multiple cycles to monitor the training process, adjust hyperparameters and prevent the model from overfitting. After the training is completed, the model with the optimal parameters is obtained. Θ trained Hybrid neural network model h(X ts ,X static ;Θ trained ) ,Deploy it into the intelligent fault diagnosis module within the test platform.

[0249] Those skilled in the art will appreciate that, in addition to implementing the system and its various devices, modules, and units provided by the present invention in purely computer-readable program code, it is entirely possible to implement the same functions of the system and its various devices, modules, and units provided by the present invention in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system and its various devices, modules, and units provided by the present invention can be considered a hardware component, and the devices, modules, and units included therein for implementing various functions can also be considered as structures within the hardware component; the devices, modules, and units for implementing various functions can also be considered as both software modules implementing the method and structures within the hardware component.

[0250] The methods, devices, systems, etc. in the embodiments disclosed herein may also be implemented in other forms. For example, the module division of the device is only the division of logical functions. In actual applications, different division methods can be adopted according to needs. For example, multiple modules can be combined or integrated into another system, and some features can be omitted or not executed. In addition, the coupling or communication between modules can be direct or indirectly achieved through interfaces, devices or intermediate modules, and the connection form can be electrical, mechanical or other methods. The separation or combination of modules can be physical or logical, that is, the modules can be deployed centrally or distributed on multiple network nodes.

[0251] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functions, and operations of the systems, methods, and computer program products according to the embodiments of the present disclosure. Each box in the flowchart or block diagram may represent a module, program segment, or portion of code, which may include one or more executable instructions that implement logical functions. In some implementations, the functions in the flowchart or block diagram may be executed in an order different from that shown in the accompanying drawings, for example, two consecutive boxes may be executed in parallel, or in reverse order. Such adjustments may be determined based on actual functional requirements. Each box in the block diagram or flowchart, or a combination thereof, may be implemented by a dedicated hardware system, or by a combination of dedicated hardware and computer instructions to perform the specified function or operation.

[0252] The above describes specific embodiments of the present invention. It should be understood that the present invention is not limited to the specific embodiments described above, and that those skilled in the art may make various changes or modifications within the scope of the claims, without affecting the essence of the present invention. Unless there is a conflict, the embodiments of this application and the features in the embodiments may be combined with each other in any manner.

Claims

1. A DDS-based aircraft wing section testing method, characterized in that: include: Step S1: Establishing a data communication link between the test platform, the DDS Agent, and one or more hardware test modules through DDS; Step S2: instructing the test platform to send a test instruction, which is forwarded to one or more hardware test modules via the DDS Agent; and instructing the hardware test modules to perform wing test operations to obtain real-time test data. Step S3: the hardware test module feeds back real-time test data to the test platform via DDS; Step S4: During the test command transmission and real-time test data feedback process, the DDS Agent calculates the communication quality Q value and dynamically adjusts the distribution strategy according to the Q value; The step S4 comprises: Step S4.1: The DDS Agent monitors the key performance indicators (KPIs) of the data communication link in real time; Step S4.2: normalize the monitored key performance indicators (KPIs) to obtain normalized key performance indicator scores; Step S4.3: Preset weight coefficients, perform weighted summation on the normalized key performance indicator scores, and calculate the communication quality Q value; Step S4.4: Set high quality threshold Q high and a low quality threshold Q low , determine the threshold interval of the Q value; Step S4.5: adjusting the distribution strategy of the preset hierarchical data according to the interval of the Q value; Among them, the key performance indicators KPIs include: end-to-end delay L, packet loss rate P, delay jitter J, effective throughput T eff and the proxy buffer occupancy rate B occ ; Step S5: Based on the adjusted distribution strategy, the test platform receives and analyzes the real-time test data, and performs fault diagnosis to determine the cause of the fault and locate the fault location on the digital twin model in the test platform.

2. The aircraft wing section testing method based on DDS according to claim 1, characterized in that: The step S1 comprises: Step S1.1: Configuring the test platform, the DDS Agent, and all hardware test modules within the same DDS domain; wherein there may be one or more hardware test modules, configured at specific locations on the wing as needed; Step S1.2: Establish data communication links between the test platform, the DDS Agent and the hardware test modules in sequence through DDS. Specifically, establish data communication links with publish-subscribe relationships between the test platform and the DDS Agent, and between the DDS Agent and one or more hardware test modules based on predefined DDS topics, for forwarding test instructions and real-time test data. The real-time test data includes sequential test data and static test data.

3. The aircraft wing section testing method based on DDS according to claim 1, characterized in that: Step S4 includes: Step S4.2 includes: normalizing the key performance indicators KPIs and calculating the normalized scores of the key performance indicators respectively; specifically, Calculating latency scores : Set the maximum delay and ideal delay ; The scoring function is designed as the time delay score When the delay score is 1, the delay score is When it is 0, it decreases linearly in the middle: Calculating packet loss score : Set the maximum tolerable packet loss rate ; The scoring function is designed to score packet loss in When it is 1, the packet loss score is When 0: Calculating Jitter Score : Set the maximum jitter and ideal jitter , the scoring function is designed to score jitter in When the jitter score is 1, When it is 0, it decreases linearly in the middle: Calculating Throughput Score : Set the minimum throughput required for the current test task , the scoring function is designed to score the throughput in When the throughput score is 1, it decreases proportionally when it is lower than the minimum throughput: Calculating buffer scores : Set the buffer occupancy warning threshold and critical thresholds , the scoring function is designed as the number of buffer scores in 1, the buffer score is When it is 0, it decreases linearly in the middle: In the above function, min(·) means calculating the minimum value, and max(·) means calculating the maximum value; Step S4.3 includes: calculating the Q value; specifically, Set weight coefficient , so that the weight coefficient satisfies ; then the Q value is in, The weight coefficients corresponding to the KPIs respectively; Step S4.5 includes: selecting a distribution strategy for hierarchical data according to the interval of the Q value, and correspondingly adjusting at least one of the data publishing rate, the internal message cache of the DDS Agent, and modifying the service quality of the DDS; specifically, when When , it indicates that the current network link status is excellent, then the adjustment strategy is: remove or relax the restrictions on data publishing rate, reduce cache occupancy, set the maximum data transmission throughput, and minimize the end-to-end communication delay; when When , it indicates that the current network link quality has deteriorated to a certain extent. The adjustment strategy is: take preventive measures to reduce the frequency of publishing non-critical or delay-tolerant data, appropriately increase cache, improve QoS service quality, actively alleviate network pressure, and give priority to ensuring the transmission quality of important data; when When , it indicates that the current network link quality has seriously deteriorated. The adjustment strategy is: adopt the most conservative transmission strategy, reserve the transmission channel for the highest priority data and ensure priority transmission, expand its internal message cache capacity to the preset maximum allowable value, and ensure the reliable transmission of the highest priority core instructions and key safety status data.

4. The aircraft wing section testing method based on DDS according to claim 1, characterized in that: The step S5 comprises: Step S5.1: The test platform receives and analyzes real-time test data. If abnormal real-time test data is analyzed, the process proceeds to step S5.

2. If normal real-time test data is analyzed, the process continues to receive data. Step S5.2: The test platform extracts time series data features and static data features from the abnormal real-time test data; Step S5.3: Inputting the time series data features into the time series processing branch in the pre-trained machine learning model for processing to obtain a time series processing result; Step S5.4: Inputting the static data features into the static data processing branch in the pre-trained machine learning model for processing to obtain a static processing result; Step S5.5: Fusion of the temporal processing result and the static processing result; Step S5.6: Based on the fusion results, a prediction of the cause of the fault is output. At the same time, combined with real-time test data, the test platform simulates the actual condition of the wing to form a digital twin model, locates the fault location on the digital twin model, and highlights and annotates it.

5. A DDS-based aircraft wing section testing system, characterized in that: The system is used to implement the aircraft wing section testing method based on DDS according to claim 1, comprising: A test platform, a DDS Agent, and one or more hardware test modules, wherein the test platform, the DDS Agent, and the hardware test modules are sequentially connected via a data communication link established by DDS; Test platform: Built on digital twins, the test platform is used to load and run digital twin models of wing sections. It receives and analyzes real-time test data fed back by the hardware test module through the internal test platform DDS communication module, achieving real-time mapping between the wing's physical hardware status and the virtual model's status. It is used to monitor test data and generate visual alarms. It integrates intelligent fault diagnosis capabilities, using machine learning to automatically track and visually locate fault chains. It is also used to manage real-time test data and send test instructions. DDS Agent: Connected to the test platform and hardware test modules, it forwards test instructions and real-time data bidirectionally based on DDS topics. It sets a communication quality evaluation (Q) value to assess network communication quality in real time and dynamically adjusts data distribution strategies based on the Q value to ensure data transmission reliability. Hardware test module: It is configured at a specific position on the wing and is used to actually measure or control specific parameters of the aircraft wing according to the test instructions issued by the test platform, and feed back real-time test data to the test platform through DDS.

6. The DDS-based aircraft wing section testing system according to claim 5, characterized in that: The test platform includes: The test platform's DDS communication module is used to create DDS publishers and subscribers, associate with the DDS Agent through DDS topics, and send, receive, and parse data based on a predefined standardized command frame format. Wing Twin Model Module: This module is used to load, manage, and run the 3D digital twin model of the wing segment. It updates the model status based on real-time test data received by the test platform's DDS communication module, synchronizing the wing physical test with the virtual wing segment test status. Monitoring and alarm module: used to verify the real-time test data and the range of the data according to the feedback from the hardware test module operation, and trigger a visual alarm when an abnormality is found; Intelligent Fault Diagnosis Module: Based on machine learning technology, it is used to establish a mapping relationship between fault modes and data features by cleaning historical data and training models. When the monitoring and alarm module detects an anomaly, it automatically extracts the abnormal data features, inputs them into the pre-trained model for calculation, predicts the cause of the fault, and traces its propagation path. The fault source and related components are highlighted on the visual interface of the wing twin model module, and diagnostic suggestions are output. Test management module: used for configuration, import, storage, and execution control of test cases, and responsible for collecting test process data, alarm records, and diagnostic results.

7. The DDS-based aircraft wing section testing system according to claim 6, characterized in that: The intelligent fault diagnosis module includes: Pre-trained fault diagnosis model submodule: This module trains a fault classification model using a hybrid input neural network model architecture based on historical time-series test data and static test data, along with their corresponding known fault cause labels. The core function of the fault classification model is to map input data features to specific fault cause categories, and to predict fault causes for abnormal test data. Test data extraction submodule: During the test period, it is used to obtain the currently monitored abnormal test data from the monitoring and alarm module in real time and perform feature engineering processing, including calculating statistical features and performing normalization processing; Fault location submodule: used to receive the fault category prediction result output by the pre-trained fault diagnosis model for the current test data feature vector; when the prediction result indicates a specific fault, it is used to parse and map this fault category information to the specific components or three-dimensional spatial positions in the digital twin model of the aircraft wing section, trigger the wing twin model module, and highlight and mark the located components or areas on the visual interface.

8. The DDS-based aircraft wing section testing system according to claim 5, characterized in that: The hardware testing module includes: Hardware test DDS communication module: used to handle DDS connection and data transmission and reception, sending data in a predefined standard format; according to the test instructions, it is used to identify whether a specific field is relevant to itself, and if so, perform the test; if not, discard the message; Physical measurement and control module: includes sensors and actuators for directly interacting with the aircraft wing sections, performing measurement or control operations, and feeding back real-time measurement data; Data preprocessing module: Filters and calibrates the raw data from the physical measurement and control module, and then sends it through the hardware test DDS communication module; It also includes multiple test channels, each test channel corresponds to a different control word and is used to perform different test operations.

9. The DDS-based aircraft wing section testing system according to claim 5, characterized in that: The DDSAgent agent includes: Data forwarding module: used to bidirectionally forward test platform and hardware test module instructions and real-time test data. It resides in the same DDS domain as the test platform and hardware test modules. Hardware test modules of the same category are associated with the same topic as the DDS Agent. When the DDS Agent receives a test instruction from the test platform, it forwards data to all hardware test modules associated with the topic simultaneously. Q-value real-time calculation module: used to calculate the communication quality evaluation index Q-value; used to quantitatively evaluate communication quality, and to normalize the key performance indicators by continuously monitoring the key performance indicators of the communication link; Communication strategy dynamic adjustment module: According to the preset Q value threshold interval, it is used to automatically execute the corresponding adjustment distribution strategy. The adjustment content includes: adjusting the data publishing rate, adjusting the Agent internal message cache size, and dynamically modifying the DDS QoS service quality.

Images