An aviation equipment maintenance management system based on real-time data

Abstract

The present invention discloses an aviation equipment maintenance management system based on real-time data, which belongs to the technical field of aviation equipment maintenance. By adopting a multi-feature extraction mechanism and a weight redistribution mechanism, an aviation equipment fault identification model is constructed, and the aviation equipment fault identification model is used to predict the faults of the aviation equipment, so that potential faults can be discovered in advance, maintenance costs can be reduced, real-time monitoring of the status of aviation equipment can be achieved, operational safety can be improved, and staff can formulate reasonable maintenance plans, improve maintenance efficiency, reduce resource waste, and improve the intelligence level of aviation equipment maintenance management.

Description

Technical Field

[0001] The present invention belongs to the technical field of aviation equipment maintenance, and in particular relates to an aviation equipment maintenance management system based on real-time data. Background Art

[0002] Aviation equipment refers to the various systems and components used in aircraft (such as airplanes, helicopters, and drones). Together, they ensure safe flight, navigation, operation, and mission execution. With the rapid development of the aviation industry, the operational safety of aviation equipment is receiving increasing attention. Traditional aviation equipment maintenance management relies primarily on manual experience, which presents the following problems: delayed fault detection can lead to safety accidents; irrational maintenance plans result in wasted resources; and the maintenance process lacks data support, making precise maintenance difficult. Summary of the Invention

[0003] The present invention provides an aviation equipment maintenance management system based on real-time data, which is used to solve the technical problems of untimely fault discovery and unreasonable maintenance in the prior art.

[0004] An aviation equipment maintenance management system based on real-time data, comprising: a sample data acquisition module, a model building module, a model training module, a real-time data acquisition module, an equipment data identification module, and a maintenance management module;

[0005] The sample data acquisition module is used to obtain a training set and a test set; wherein the training set and the test set both include training samples and training labels corresponding to the training samples;

[0006] The model building module is used to build an aviation equipment fault identification model using a multi-feature extraction mechanism and a weight redistribution mechanism;

[0007] The model training module is used to train the aviation equipment fault identification model based on the training set and the test set to obtain the trained aviation equipment fault identification model;

[0008] The real-time data acquisition module is used to collect real-time data corresponding to the aviation equipment; wherein the data structure of the real-time data is the same as the data structure of the training sample;

[0009] The equipment data identification module is used to call the trained aviation equipment fault identification model to identify the real-time data corresponding to the aviation equipment and determine the aviation equipment fault status; wherein the aviation equipment fault status includes the specific fault type status and the normal status;

[0010] The maintenance management module is used to generate aviation equipment maintenance suggestions and abnormal situation warning information based on the aviation equipment failure situation, and perform aviation equipment maintenance management according to the aviation equipment maintenance suggestions and abnormal situation warning information.

[0011] Furthermore, the training samples include sample currents and sample voltages at multiple consecutive time points.

[0012] Furthermore, a multi-feature extraction mechanism and a weight redistribution mechanism are used to construct an aviation equipment fault identification model, including:

[0013] Constructing a first convolutional neural network, a second convolutional neural network, and a third convolutional neural network with different network layer depths, and performing multi-scale feature fusion on the features output by the fully connected layers of the first convolutional neural network, the second convolutional neural network, and the third convolutional neural network to obtain fused features;

[0014] Construct an ECA-Net model to assign weights to the fused features and output the weighted features;

[0015] Construct a BiLSTM model to perform data reinforcement on the time dimension of the features after weight assignment through forward and backpropagation bidirectional operations, and output the final vector to be identified;

[0016] Build a CatBoost model to identify the final vector to be identified and output the aviation equipment fault type;

[0017] According to the first convolutional neural network, the second convolutional neural network, the third convolutional neural network, the ECA-Net model, the BiLSTM model and the CatBoost model, an aviation equipment fault identification model is obtained.

[0018] Furthermore, based on the training set and the test set, an aviation equipment fault identification model is trained to obtain a trained aviation equipment fault identification model, including:

[0019] After initializing the model parameters to be optimized of the aviation equipment fault identification model, generating a population to be trained, and training the population to be trained using a training set to obtain an optimal parameter combination in the population to be trained;

[0020] Apply the optimal parameter combination in the training population to the aviation equipment fault identification model, and use the test set to test the aviation equipment fault identification model to obtain test results;

[0021] When the test result satisfies the conditions, the trained aviation fault identification model is obtained; otherwise, the process returns to the step of obtaining the optimal parameter combination in the to-be-trained population.

[0022] Furthermore, the training set is used to train the population to be trained to obtain the optimal parameter combination in the population to be trained, including:

[0023] The training set is used to obtain the objective function value corresponding to each parameter combination in the population to be trained, and the optimal parameter combination is determined based on the objective function value;

[0024] Based on the optimal parameter combination, a multi-directional guided optimization is performed on each parameter combination in the training population to obtain a parameter combination after the multi-directional guided optimization;

[0025] For the parameter combination after the multi-directional guided optimization search, a fractional order memory search is performed on the parameter combination after the multi-directional guided optimization search to obtain the parameter combination after the fractional order memory search;

[0026] For the parameter combination after the fractional-order memory search, a normal distribution search is performed on the parameter combination after the fractional-order memory search to obtain the parameter combination after the normal distribution search;

[0027] Repeat multi-directional guided optimization, fractional-order memory search, and normal distribution search until the training end conditions are met and the optimal parameter combination is output.

[0028] Furthermore, based on the optimal parameter combination, a multi-directional guided optimization is performed on each parameter combination in the training population to obtain a parameter combination after the multi-directional guided optimization, including:

[0029] For each parameter combination in the training population, randomly match another parameter combination for each parameter combination, and the matched other parameter combination is different from the optimal parameter combination;

[0030] Fusing the data of other matched parameter combinations with the optimal parameter combination, and obtaining the difference information combination between the parameter combination and the fused data;

[0031] After the difference information combination is adjusted by using a random adjustment step factor, the parameter combination is subjected to multi-directional guided optimization using the adjusted difference information combination to obtain the parameter combination after the multi-directional guided optimization.

[0032] Furthermore, for the parameter combination after the multi-directional guided optimization search, a fractional order memory search is performed on the parameter combination after the multi-directional guided optimization search to obtain the parameter combination after the fractional order memory search, including:

[0033] For the parameter combination after multi-directional guided optimization, the historical position information of the parameter combination is memorized multiple times to obtain a memory information combination;

[0034] The overall information difference combination between the parameter combination and the overall position is obtained, and a fractional order memory search is performed on the parameter combination according to the memory information combination and the overall information difference combination to obtain the parameter combination after the fractional order memory search.

[0035] Furthermore, for the parameter combination after the fractional-order memory search, a normal distribution search is performed on the parameter combination after the fractional-order memory search to obtain the parameter combination after the normal distribution search, including:

[0036] For the parameter combination after the fractional-order memory search, a normal distribution is used to generate a random adjustment factor, and the random adjustment factor is adjusted using a scaling factor to obtain the adjusted random adjustment factor;

[0037] The parameter combination after the fractional-order memory search is adjusted using the adjusted random adjustment factor to obtain the parameter combination after the normal distribution search.

[0038] Furthermore, for aviation equipment, real-time data corresponding to the aviation equipment is collected, including:

[0039] For aviation equipment, based on the data sampling frequency corresponding to the training samples, the real-time current and real-time voltage corresponding to the aviation equipment are collected to obtain the real-time data corresponding to the aviation equipment.

[0040] Furthermore, based on the aviation equipment failure situation, aviation equipment maintenance suggestions and abnormal situation warning information are generated, and aviation equipment maintenance management is performed according to the aviation equipment maintenance suggestions and abnormal situation warning information, including:

[0041] When the aviation equipment failure condition is a specific failure type state, determining the aviation equipment maintenance suggestion corresponding to the preset specific failure type state;

[0042] When the aviation equipment fault condition is a specific fault type state, an audible and visual alarm message as well as a text alarm message is generated to obtain abnormal situation warning information;

[0043] The text alarm information in the aviation equipment maintenance suggestions and abnormal situation warning information is transmitted to the corresponding equipment of the staff, and the sound and light alarm information is transmitted to the sound and light alarm equipment for alarm, so that the staff can perform timely maintenance on the aviation equipment.

[0044] The present invention provides an aviation equipment maintenance management system based on real-time data. By adopting a multi-feature extraction mechanism and a weight redistribution mechanism, an aviation equipment fault identification model is constructed, and the aviation equipment fault identification model is used to predict aviation equipment faults. This can detect potential faults in advance, reduce maintenance costs, achieve real-time monitoring of the aviation equipment status, improve operational safety, enable staff to formulate reasonable maintenance plans, improve maintenance efficiency, reduce resource waste, and improve the intelligence level of aviation equipment maintenance management. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0046] Figure 1 A schematic structural diagram of an aviation equipment maintenance management system based on real-time data provided by an embodiment of the present invention.

[0047] Among them, 11-sample data acquisition module, 12-model construction module, 13-model training module, 14-real-time data acquisition module, 15-equipment data identification module, and 16-maintenance management module.

[0048] The above drawings illustrate specific embodiments of the present invention, which will be described in more detail below. These drawings and the accompanying description are not intended to limit the scope of the present invention in any way, but rather to illustrate the concept of the present invention to those skilled in the art by reference to specific embodiments. DETAILED DESCRIPTION

[0049] Exemplary embodiments will be described in detail herein, examples of which are illustrated in the accompanying drawings. In the following description, when referring to the drawings, like numbers in different figures represent like or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all possible embodiments consistent with the present invention. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present invention, as detailed in the appended claims.

[0050] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0051] like Figure 1 As shown, an embodiment of the present invention provides an aviation equipment maintenance management system based on real-time data, comprising: a sample data acquisition module 11, a model construction module 12, a model training module 13, a real-time data acquisition module 14, an equipment data identification module 15, and a maintenance management module 16;

[0052] The sample data acquisition module 11 is used to obtain a training set and a test set; wherein the training set and the test set both include training samples and training labels corresponding to the training samples;

[0053] The training samples are data collected under various specific fault types or data at the early stage of a fault. The training labels corresponding to the training samples are the specific fault types. However, there may be no fault during real-time monitoring. Therefore, the training labels should also include normal distributions, so that the fault characteristics and normal characteristics can be learned through the training set and combined with deep learning technology, and the feasibility of the learning results can be determined through the test set.

[0054] The model building module 12 is used to build an aviation equipment fault identification model using a multi-feature extraction mechanism and a weight redistribution mechanism;

[0055] In existing technologies, a single data classification model is often used for data learning and recognition, resulting in low accuracy and ineffective fault prediction. Therefore, embodiments of the present invention utilize a multi-feature extraction and weight redistribution mechanism to construct an aviation equipment fault recognition model, enhancing feature analysis capabilities and ultimately improving fault prediction capabilities.

[0056] The model training module 13 is used to train the aviation equipment fault identification model based on the training set and the test set to obtain a trained aviation equipment fault identification model;

[0057] Training the aviation equipment fault identification model may include: using a gradient descent method, a particle swarm algorithm, or other intelligent optimization algorithms to train the aviation equipment fault identification model to obtain a trained aviation equipment fault identification model.

[0058] The real-time data acquisition module 14 is used to collect real-time data corresponding to the aviation equipment; wherein the data structure of the real-time data is the same as the data structure of the training sample;

[0059] By ensuring that the data structure of real-time data is the same as that of training samples, the trained aviation equipment fault identification model can accurately identify data, thereby achieving fault prediction.

[0060] The equipment data identification module 15 is used to call the trained aviation equipment fault identification model to identify the real-time data corresponding to the aviation equipment and determine the aviation equipment fault status; wherein the aviation equipment fault status includes the specific fault type status and the normal status;

[0061] The maintenance management module 16 is configured to generate aviation equipment maintenance suggestions and abnormality warning information based on the aviation equipment failure situation, and perform aviation equipment maintenance management according to the aviation equipment maintenance suggestions and abnormality warning information.

[0062] A database can be pre-configured to link aviation equipment failures with maintenance recommendations. These recommendations can be obtained from various historical data. After predicting aircraft failures, recommendations can be quickly provided, improving maintenance efficiency. Furthermore, abnormal situation warning information can be generated and used to provide early warnings, enabling real-time monitoring of aircraft status.

[0063] The present invention provides an aviation equipment maintenance management system based on real-time data. By adopting a multi-feature extraction mechanism and a weight redistribution mechanism, an aviation equipment fault identification model is constructed, and the aviation equipment fault identification model is used to predict aviation equipment faults. This can detect potential faults in advance, reduce maintenance costs, achieve real-time monitoring of the aviation equipment status, improve operational safety, enable staff to formulate reasonable maintenance plans, improve maintenance efficiency, reduce resource waste, and improve the intelligence level of aviation equipment maintenance management.

[0064] In this embodiment of the present invention, the training samples include sample currents and sample voltages at multiple consecutive time points. Current and voltage can effectively reflect equipment operating conditions, such as short circuits, open circuits, poor contact, camera damage, or abnormal motor operation. Therefore, this embodiment of the present invention uses current and voltage to detect abnormal operation of aviation equipment.

[0065] It is worth noting that in addition to the above features, other features can be used to construct training samples for different aviation equipment. For example, when it comes to motor equipment, vibration frequency can also be collected to improve the accuracy of fault prediction.

[0066] The sample currents and sample voltages at multiple consecutive time points are data collected using a preset data sampling frequency. When subsequently collecting real-time data, data should also be collected using the data sampling frequency to ensure that the data can be accurately identified.

[0067] In an embodiment of the present invention, a multi-feature extraction mechanism and a weight redistribution mechanism are used to construct an aviation equipment fault identification model, including:

[0068] Constructing a first convolutional neural network, a second convolutional neural network, and a third convolutional neural network with different network layer depths, and performing multi-scale feature fusion on the features output by the fully connected layers of the first convolutional neural network, the second convolutional neural network, and the third convolutional neural network to obtain fused features;

[0069] Multi-scale feature fusion can be achieved by performing an element-by-element product operation on the features output by the first, second, and third convolutional neural networks to obtain fused features. Therefore, collected training samples or real-time data can be constructed as a data matrix as input, effectively achieving feature extraction.

[0070] Build an ECA-Net (Efficient Channel Attention Network, lightweight attention mechanism) model to weight the fused features and output the weighted features;

[0071] Build a Bidirectional Long Short-Term Memory (BiLSTM) model to perform data reinforcement on the time dimension of the weighted features through forward and backpropagation operations, and output the final vector to be identified.

[0072] Build a CatBoost (regression classification) model to identify the final vector to be identified and output the aviation equipment fault type;

[0073] According to the first convolutional neural network, the second convolutional neural network, the third convolutional neural network, the ECA-Net model, the BiLSTM model and the CatBoost model, an aviation equipment fault identification model is obtained.

[0074] The aviation equipment fault identification model provided by the embodiment of the present invention can first use the first convolutional neural network, the second convolutional neural network and the third convolutional neural network with different network layer depths to more comprehensively extract the characteristics of the data and enhance the overall performance of the network. Then, the weights are redistributed through the ECA-Net model to effectively capture the information of cross-channel interactions. Then, the data features are further enhanced in the time dimension through the BiLSTM model. Finally, the fault classification is performed through the CatBoost model. Compared with the single classification model, the data recognition accuracy can be effectively improved, thereby improving the fault prediction accuracy.

[0075] In an embodiment of the present invention, an aviation equipment fault identification model is trained based on the training set and the test set to obtain a trained aviation equipment fault identification model, including:

[0076] After initializing the model parameters to be optimized of the aviation equipment fault identification model, generating a population to be trained, and training the population to be trained using a training set to obtain an optimal parameter combination in the population to be trained;

[0077] The model parameters to be optimized of the aviation equipment fault identification model can be the connection weights between each network layer. By optimizing these connection weights, the aviation equipment fault identification model can effectively identify real-time data in the subsequent process.

[0078] Apply the optimal parameter combination in the training population to the aviation equipment fault identification model, and use the test set to test the aviation equipment fault identification model to obtain test results;

[0079] The test result may be an accuracy rate obtained using a test set. When the accuracy rate is higher than a preset threshold, it may be considered that the condition is met; otherwise, it may be determined that the condition is not met.

[0080] When the test result satisfies the conditions, the trained aviation fault identification model is obtained; otherwise, the process returns to the step of obtaining the optimal parameter combination in the to-be-trained population.

[0081] In the process of training deep learning models with existing technologies, there is a technical problem of easily falling into local optimality, resulting in a low final fault prediction accuracy. Therefore, the embodiments of the present invention improve the global search capability of the algorithm through multi-directional guided optimization, fractional-order memory search, and normal distribution search, thereby making the acquired parameters more accurate, improving the performance of the deep learning model, and ultimately improving the fault prediction accuracy.

[0082] In an embodiment of the present invention, the training set is used to train the to-be-trained population to obtain the optimal parameter combination in the to-be-trained population, including:

[0083] The training set is used to obtain the objective function value corresponding to each parameter combination in the population to be trained, and the optimal parameter combination is determined based on the objective function value; the parameter combination can be a vector to facilitate data calculation.

[0084] The objective function value can be obtained through the cross entropy loss function. The smaller the objective function value, the better the parameter combination. Therefore, the parameter combination with the smallest objective function value can be determined as the optimal parameter combination.

[0085] Based on the optimal parameter combination, a multi-directional guided optimization is performed on each parameter combination in the training population to obtain a parameter combination after the multi-directional guided optimization;

[0086] For the parameter combination after the multi-directional guided optimization search, a fractional order memory search is performed on the parameter combination after the multi-directional guided optimization search to obtain the parameter combination after the fractional order memory search;

[0087] For the parameter combination after the fractional-order memory search, a normal distribution search is performed on the parameter combination after the fractional-order memory search to obtain the parameter combination after the normal distribution search;

[0088] Repeat the multi-directional guided optimization, fractional-order memory search, and normal distribution search until the training end condition is met and the optimal parameter combination is output. The training end condition can be that the number of training times reaches the maximum number of training times.

[0089] In an embodiment of the present invention, based on the optimal parameter combination, a multi-directional guided optimization is performed on each parameter combination in the training population to obtain a parameter combination after the multi-directional guided optimization, including:

[0090] For each parameter combination in the training population, randomly match another parameter combination for each parameter combination, and the matched other parameter combination is different from the optimal parameter combination;

[0091] The other matching parameter combinations are fused with the optimal parameter combination, and based on the fused data, the difference information combination between the parameter combinations is obtained as follows: ;in, Indicates the t During the training i parameter combinations, Expressed as t During the training i parameter combinations that match other parameter combinations, represents the optimal parameter combination.

[0092] After the difference information combination is adjusted by randomly adjusting the step size factor, the parameter combination is multi-directionally guided to optimize using the adjusted difference information combination. The parameter combination after multi-directional guidance optimization is: ;in, Represents the parameter combination after multi-directional guided optimization , represents the random adjustment step size factor, , Indicates the preset maximum step size, Represents a random number between (0,1), i =1,2,…,M, where M represents the total number of parameter combinations.

[0093] The multi-directional guided optimization provided by the embodiment of the present invention can not only search the parameter combination in the optimal direction, but also be affected by other parameter combinations and fluctuate continuously during the search process, thereby ensuring the algorithm search efficiency while reducing the probability of the algorithm falling into local optimality.

[0094] In an embodiment of the present invention, for a parameter combination after the multi-directional guided optimization search, a fractional-order memory search is performed on the parameter combination after the multi-directional guided optimization search to obtain a parameter combination after the fractional-order memory search, including:

[0095] For the parameter combination after multi-directional guided optimization, the historical position information of the parameter combination is memorized multiple times, and the memory information combination is obtained as follows: ;in, represents the memory factor, and , e represents a natural constant, iter represents the maximum number of training times, Indicates the t During the training j The parameter combination after multi-directional guided optimization, Indicates parameter combination In the t -1 state during training, Indicates parameter combination In the t -2 states during training, Indicates parameter combination In the t - Status during 3 training sessions;

[0096] The overall information difference combination between the obtained parameter combination and the overall position is: ;in, Represents a random number between (0,0.5), Represents the overall position, that is, the parameter of each dimension is the mean of all parameters combined in the same dimension.

[0097] According to the memory information combination and the overall information difference combination, the parameter combination is searched by fractional order memory, and the parameter combination after the fractional order memory search is obtained as follows: ;in, Represents the parameter combination after fractional-order memory search .

[0098] The fractional-order memory search provided by the embodiment of the present invention can enable the parameter combination to memorize its own historical state, and reduce it over time, which can effectively ensure the diversity of parameter combinations and improve the global search capability of the algorithm. As the algorithm proceeds, all parameter combinations gradually converge in the solution space, and the memory position becomes better and better, which can effectively improve the local search capability of the algorithm.

[0099] In an embodiment of the present invention, for the parameter combination after the fractional-order memory search, a normal distribution search is performed on the parameter combination after the fractional-order memory search to obtain the parameter combination after the normal distribution search, including:

[0100] For the parameter combination after the fractional-order memory search, a normal distribution is used to generate a random adjustment factor, and a scaling factor is used to adjust the random adjustment factor. The random adjustment factor after adjustment is: ;in, Indicates that for d The random adjustment factor generated by the dimensional parameter is obtained through the standard normal distribution generate, Indicates that for d The scaling factor generated by the dimension parameter, and , Indicates a preset constant factor (such as 0.5, 0.55, etc.), Indicates the d The upper limit of the dimension parameter, Indicates the d The lower limit of the dimension parameter;

[0101] The parameter combination after the fractional-order memory search is adjusted using the adjusted random adjustment factor, and the parameter combination after the normal distribution search is obtained as follows:

[0102]

[0103] in, Indicates the t During the training m The first parameter combination after the fractional order memory search d dimensional parameters, Indicates the parameters after normal distribution search .

[0104] Optionally, when the objective function value of the parameter combination after the normal distribution search decreases, the search is retained, otherwise the search is rejected.

[0105] Whenever a search is performed on a parameter combination, the parameter combination should be processed to make the first d The dimension parameter is always in [ , ]between.

[0106] In an embodiment of the present invention, for aviation equipment, collecting real-time data corresponding to the aviation equipment includes:

[0107] For aviation equipment, based on the data sampling frequency corresponding to the training samples, the real-time current and real-time voltage corresponding to the aviation equipment are collected to obtain the real-time data corresponding to the aviation equipment.

[0108] In an embodiment of the present invention, based on the aviation equipment failure condition, generating aviation equipment maintenance suggestions and abnormality warning information, and performing aviation equipment maintenance management according to the aviation equipment maintenance suggestions and abnormality warning information, includes:

[0109] When the aviation equipment failure condition is a specific failure type state, determining the aviation equipment maintenance suggestion corresponding to the preset specific failure type state;

[0110] When the aviation equipment fault condition is a specific fault type state, an audible and visual alarm message as well as a text alarm message is generated to obtain abnormal situation warning information;

[0111] The text alarm information in the aviation equipment maintenance suggestions and abnormal situation warning information is transmitted to the corresponding equipment of the staff, and the sound and light alarm information is transmitted to the sound and light alarm equipment for alarm, so that the staff can perform timely maintenance on the aviation equipment.

[0112] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0113] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0114] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0115] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps for the function specified in one or more boxes.

[0116] Those skilled in the art will understand that all or part of the steps in implementing the above facts and methods can be completed by instructing relevant hardware through a program, and the program involved or the program can be stored in a computer-readable storage medium. When the program is executed, it includes the following steps: the corresponding method steps are then brought out, and the storage medium can be ROM / RAM, a disk, an optical disk, etc.

[0117] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An aviation equipment maintenance management system based on real-time data, characterized in that: include: Sample data acquisition module, model building module, model training module, real-time data acquisition module, equipment data identification module and maintenance management module; The sample data acquisition module is used to obtain a training set and a test set; wherein the training set and the test set both include training samples and training labels corresponding to the training samples; The model building module is used to construct an aviation equipment fault identification model using a multi-feature extraction mechanism and a weight distribution mechanism, including: Constructing a first convolutional neural network, a second convolutional neural network, and a third convolutional neural network with different network layer depths, and performing multi-scale feature fusion on the features output by the fully connected layers of the first convolutional neural network, the second convolutional neural network, and the third convolutional neural network to obtain fused features; Construct an ECA-Net model to assign weights to the fused features and output the weighted features; Construct a BiLSTM model to perform data reinforcement on the time dimension of the features after weight assignment through forward and backpropagation bidirectional operations, and output the final vector to be identified; Build a CatBoost model to identify the final vector to be identified and output the aviation equipment fault type; According to the first convolutional neural network, the second convolutional neural network, the third convolutional neural network, the ECA-Net model, the BiLSTM model and the CatBoost model, an aviation equipment fault identification model is obtained; The model training module is used to train the aviation equipment fault identification model based on the training set and the test set to obtain the trained aviation equipment fault identification model, including: After initializing the model parameters to be optimized of the aviation equipment fault identification model, generating a population to be trained, and training the population to be trained using a training set to obtain an optimal parameter combination in the population to be trained; Apply the optimal parameter combination in the training population to the aviation equipment fault identification model, and use the test set to test the aviation equipment fault identification model to obtain test results; When the test result satisfies the conditions, the trained aviation fault identification model is obtained; otherwise, the step of obtaining the optimal parameter combination in the to-be-trained population is returned; The training set is used to train the population to be trained to obtain the optimal parameter combination in the population to be trained, including: The training set is used to obtain the objective function value corresponding to each parameter combination in the population to be trained, and the optimal parameter combination is determined based on the objective function value; Based on the optimal parameter combination, a multi-directional guided optimization is performed on each parameter combination in the training population to obtain a parameter combination after the multi-directional guided optimization; For the parameter combination after the multi-directional guided optimization search, a fractional order memory search is performed on the parameter combination after the multi-directional guided optimization search to obtain the parameter combination after the fractional order memory search; For the parameter combination after the fractional-order memory search, a normal distribution search is performed on the parameter combination after the fractional-order memory search to obtain the parameter combination after the normal distribution search; Repeat multi-directional guided optimization, fractional-order memory search, and normal distribution search until the training end conditions are met and the optimal parameter combination is output; The real-time data acquisition module is used to collect real-time data corresponding to the aviation equipment; wherein the data structure of the real-time data is the same as the data structure of the training sample; The equipment data identification module is used to call the trained aviation equipment fault identification model to identify the real-time data corresponding to the aviation equipment and determine the aviation equipment fault status; wherein the aviation equipment fault status includes the specific fault type status and the normal status; The maintenance management module is used to generate aviation equipment maintenance suggestions and abnormal situation warning information based on the aviation equipment failure situation, and perform aviation equipment maintenance management according to the aviation equipment maintenance suggestions and abnormal situation warning information.

2. The aviation equipment maintenance management system based on real-time data according to claim 1, characterized in that: The training samples include sample currents and sample voltages at a plurality of consecutive time points.

3. The aviation equipment maintenance management system based on real-time data according to claim 1, characterized in that: Based on the optimal parameter combination, a multi-directional guided optimization is performed on each parameter combination in the training population to obtain a parameter combination after the multi-directional guided optimization, including: For each parameter combination in the training population, randomly match another parameter combination for each parameter combination, and the matched other parameter combination is different from the optimal parameter combination; Fusing the data of other matched parameter combinations with the optimal parameter combination, and obtaining the difference information combination between the parameter combination and the fused data; After the difference information combination is adjusted by using a random adjustment step factor, the parameter combination is subjected to multi-directional guided optimization using the adjusted difference information combination to obtain the parameter combination after the multi-directional guided optimization.

4. The aviation equipment maintenance management system based on real-time data according to claim 3, characterized in that: For the parameter combination after the multi-directional guided optimization search, a fractional order memory search is performed on the parameter combination after the multi-directional guided optimization search to obtain the parameter combination after the fractional order memory search, including: For the parameter combination after multi-directional guided optimization, the historical position information of the parameter combination is memorized multiple times to obtain a memory information combination; The overall information difference combination between the parameter combination and the overall position is obtained, and a fractional order memory search is performed on the parameter combination according to the memory information combination and the overall information difference combination to obtain the parameter combination after the fractional order memory search.

5. The aviation equipment maintenance management system based on real-time data according to claim 4, characterized in that: For the parameter combination after the fractional-order memory search, a normal distribution search is performed on the parameter combination after the fractional-order memory search to obtain the parameter combination after the normal distribution search, including: For the parameter combination after the fractional-order memory search, a random adjustment factor is generated using a normal distribution, and the random adjustment factor is adjusted using a scaling factor to obtain the adjusted random adjustment factor; The parameter combination after the fractional-order memory search is adjusted using the adjusted random adjustment factor to obtain the parameter combination after the normal distribution search.

6. The aviation equipment maintenance management system based on real-time data according to claim 1, characterized in that: For aviation equipment, collect real-time data corresponding to aviation equipment, including: For aviation equipment, based on the data sampling frequency corresponding to the training samples, the real-time current and real-time voltage corresponding to the aviation equipment are collected to obtain the real-time data corresponding to the aviation equipment.

7. The aviation equipment maintenance management system based on real-time data according to claim 6, characterized in that: Based on the aviation equipment failure situation, aviation equipment maintenance suggestions and abnormal situation warning information are generated, and aviation equipment maintenance management is performed according to the aviation equipment maintenance suggestions and abnormal situation warning information, including: When the aviation equipment failure condition is a specific failure type state, determining the aviation equipment maintenance suggestion corresponding to the preset specific failure type state; When the aviation equipment fault condition is a specific fault type state, an audible and visual alarm message as well as a text alarm message is generated to obtain abnormal situation warning information; The text alarm information in the aviation equipment maintenance suggestions and abnormal situation warning information is transmitted to the corresponding equipment of the staff, and the sound and light alarm information is transmitted to the sound and light alarm equipment for alarm, so that the staff can perform timely maintenance on the aviation equipment.

Images