Task-oriented helicopter guarantee resource scheduling method

Abstract

The invention provides a task-oriented helicopter guarantee resource scheduling method. The method comprises the following steps of 1, acquiring historical maintenance record data; 2, training a guarantee resource dynamic scheduling model by using historical maintenance record data to obtain an optimized guarantee resource dynamic scheduling model; step 3, obtaining guarantee resource scheduling and allocation requirements, and obtaining guarantee resource real-time state data from a guarantee resource library; 4, obtaining a guarantee resource configuration optimization scheme by utilizing the optimized guarantee resource dynamic scheduling model according to guarantee resource scheduling and allocation requirements and guarantee resource real-time state data; and 5, continuously training, optimizing and guaranteeing a resource dynamic scheduling model by using newly added historical maintenance record data. According to the invention, a neural network technology is adopted, data can be automatically collected, analyzed and processed, a guarantee demand scheme is intelligently generated, informatization and intelligentization of guarantee resource scheduling are realized, and the helicopter guarantee efficiency can be effectively improved.

Description

Technical Field

[0001] This application belongs to the field of helicopter integrated support technology, and in particular relates to a mission-oriented helicopter support resource scheduling method. Background Technology

[0002] To improve the efficiency of helicopter support resource utilization and to coordinate and allocate helicopter support resources, it is necessary to establish a mission-centric support resource scheduling technology, and to rationally allocate and dynamically use resources according to different missions and maintenance support processes.

[0003] The current resource allocation system relies entirely on manual processes, without the application of information and intelligent technologies, which presents the main technical problems. (1) Difficulty in confirming support needs. Support needs are the input and foundation of resource scheduling. Only by accurately and quickly identifying support needs can support resource scheduling be effectively completed and the efficiency of helicopter maintenance support be improved. Current technology relies on manual confirmation of needs, which can no longer meet the requirements of multiple departments, multiple fields and multiple professions, and there are significant gaps in real-time performance and accuracy.

[0004] (2) The formation of a resource allocation plan is time-consuming. Resource allocation involves various elements such as personnel, machinery, materials, methods, and environment. Current technology relies on manual checks of the inventory and status of aviation materials, tools, and equipment to form a support plan based on whether the support needs are met, which is time-consuming.

[0005] (3) Low efficiency in resource approval. The current technology relies on manual approval to complete the approval process for the entry, exit and use of aircraft materials, spare parts, tools and equipment, and equipment at the four stations, which is too inefficient and urgently needs to be improved with electronic approval process. Summary of the Invention

[0006] The purpose of this invention is to address the above-mentioned problems by providing a task-oriented resource scheduling technology that dynamically allocates resources required for the maintenance and support process based on different tasks, generates resource scheduling schemes, and optimizes the schemes based on the evaluation results.

[0007] This application provides a mission-oriented helicopter support resource scheduling method, which includes the following steps: Step 1: Obtain historical maintenance record data; Step 2: Using historical maintenance record data, train the dynamic scheduling model for support resources to obtain the optimized dynamic scheduling model for support resources; Step 3: Obtain the requirements for scheduling and allocation of support resources, and obtain real-time status data of support resources from the support resource database; Step 4: Based on the needs for resource scheduling and allocation, and the real-time status data of resources, use the optimized dynamic scheduling model for resources to obtain an optimized resource allocation scheme. Step 5: Utilize newly added historical maintenance record data to continuously train and optimize the dynamic scheduling model for ensuring resources.

[0008] Preferably, step 1 specifically includes: Historical maintenance records are retrieved from the helicopter fleet's support information database via scheduled requests; these historical maintenance records refer to the maintenance records that have been archived in the support information system.

[0009] Preferably, the historical maintenance record data includes helicopter flight mission data, helicopter status data, environmental data, and maintenance data, and the descriptions of each type of data are as follows: Helicopter mission data includes the name, type, number of sorties, and flight duration of the helicopter flight mission; Helicopter status data includes helicopter maintenance fault list, pilot fault list, flight parameter interpretation conclusions, fault diagnosis results, trend analysis results, status monitoring conclusions, and fault prediction results; Environmental data includes meteorological data, time data, and maintenance site data during helicopter missions; Maintenance data includes maintenance project data and support resource usage data.

[0010] Preferably, the maintenance project data includes parts replacement, operation testing, visual inspection, disassembly and assembly testing, off-site inspection, off-site repair, in-situ repair, modification, service support, maintenance procedures / manuals related to the maintenance project, as well as the duration of each maintenance project and the personnel required; the support resource usage data includes the type and quantity of aviation materials, tools and equipment, and airfield resources used to perform each maintenance project.

[0011] Preferably, step 2 specifically includes: Historical maintenance record data is preprocessed to construct training and testing sets; an RBF neural network model is constructed, and the training dataset is used to train the dynamic scheduling model for support resources, resulting in an optimized dynamic scheduling model for support resources.

[0012] Preferably, step 3 specifically includes: Real-time helicopter flight mission data, helicopter status data, environmental data, and maintenance data are read through external data interfaces to ensure resource scheduling and allocation needs; real-time status data of support resources are obtained from the support resource database, and input vector P (P1, P2, P3...P...) is generated according to step 2. nThe external data interfaces include an interface with the ground health management system to obtain helicopter status data; and an interface with the support information system to obtain helicopter mission data, environmental data, and maintenance data.

[0013] Preferably, step 4 specifically includes: Read the input vector generated in step 3, call the optimized helicopter support resource dynamic scheduling model, and output the support resource demand vector Y (y1, y2, y3...y... n Then, the data in the resource demand vector is used to call the corresponding dimensionality-upgrading algorithm to restore the business data and generate a resource configuration optimization plan; wherein, the dimensionality-upgrading algorithm is the reverse calculation of the dimensionality-reduction algorithm in step 1.

[0014] Preferably, step 5 specifically includes: Set a training sample threshold for the helicopter support resource scheduling model, read and store newly added historical maintenance record data. When the number of newly added data records reaches the set training sample threshold, proceed to steps 1-3 again to continuously optimize the helicopter support resource scheduling model.

[0015] The beneficial technical effects of this application are as follows: This invention employs neural network technology, which can automatically collect, analyze, and process data, and intelligently generate support requirements solutions. It realizes the informatization and intelligentization of support resource scheduling, and can effectively improve the efficiency of helicopter support.

[0016] (1) It can obtain the protection needs relatively accurately. In the design of this invention, the influencing factors of resource scheduling were fully considered. The input parameters were designed to be complete, reasonable and scientific. The invention also realized the automatic collection, analysis and processing of relevant data, thus ensuring the accuracy of the guarantee requirements.

[0017] (2) Reduce the time required to form a scheduling plan This invention utilizes network information technology to transform the manual querying of aircraft materials, spare parts, tools, equipment inventory, and status into an automatic querying system. It can quickly ascertain information such as the quantity, quality, and status of resources, which is conducive to the rapid formulation of resource scheduling plans.

[0018] (3) Achieve electronic approval for guarantee resources Based on this invention, the approval process for the entry, exit, and use of aircraft materials, spare parts, tools, equipment, and four-station equipment can be digitized, greatly improving approval efficiency. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. The drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart of the task-oriented resource scheduling process of the present invention; Figure 2 This is a diagram illustrating the implementation process of the neural network algorithm of this invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] The features and illustrative embodiments of various aspects of the present invention will now be described in detail. Numerous specific details are set forth in the following detailed description to provide a thorough understanding of the invention. However, it will be apparent to those skilled in the art that the invention may be practiced without requiring some of these specific details. The following description of embodiments is merely intended to provide a better understanding of the invention by illustrating examples of the invention. The invention is by no means limited to any specific setups and methods set forth below, but covers any improvements, substitutions, and modifications to structures, methods, and devices without departing from the spirit of the invention. Well-known structures and techniques are not shown in the drawings and the following description to avoid unnecessarily obscuring the invention.

[0023] In the description of this invention, it should be noted that the directions or positional relationships indicated by terms such as "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer" are based on the directions or positional relationships shown in the accompanying drawings and are only for the convenience of describing and simplifying the invention, and should not be construed as limiting the invention. Furthermore, the use of ordinal numbers (e.g., "first and second," etc.) is for distinguishing objects and is not limited to this order, and should not be construed as indicating or implying relative importance.

[0024] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly, encompassing both direct connection and indirect connection via an intermediate medium. Those skilled in the art can understand the specific meaning of these terms in this invention based on the specific circumstances.

[0025] It should be noted that, unless otherwise specified, the embodiments of the present invention and the features thereof can be combined with each other, and the various embodiments can be referenced and cited in each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0026] The present invention will be further described in detail below with reference to the embodiments and accompanying drawings, but the embodiments of the present invention are not limited thereto.

[0027] This technology conducts comprehensive data analysis on support information such as aviation material spare parts resources, human resources, and material resources, establishes a dynamic scheduling model for support resources, and generates support resource scheduling plans.

[0028] The mission-oriented dynamic scheduling model for support resources addresses the varying resource requirements of different missions. Based on maintenance work plans and resource needs, it utilizes techniques such as linear programming and knowledge learning to optimize resource scheduling, formulate resource scheduling schemes, and recommend reasonable resource reserve scales and procurement plans. This transforms the past passive resource support into knowledge-based autonomous support.

[0029] This invention adopts the following technical solution: a task-oriented maintenance and support resource scheduling process as follows: Figure 1 As shown. The design process is described below: The mission-oriented maintenance support resource scheduling model takes into account scheduling requirements (project, time, content, personnel, etc.), support resource status information (personnel, aircraft materials, facilities, tools and equipment, etc.), and historical resource allocation information (record number, specialty, system, phenomenon description, scheduling reason, handling method, etc.), and outputs a dynamic resource scheduling optimization scheme. Factors influencing the model include support time, support priority, number of resource allocation points, and personnel proficiency.

[0030] The scheduling demand information is issued through intelligent maintenance decision-making; the support resource database, located within the agile support information database, categorizes and stores information related to support resources such as personnel, tools and equipment, aviation materials, and support facilities, while also recording the allocation, consumption, and occupancy of these resources. By querying the support resource database, one can understand the resource usage of the entire system. Furthermore, through knowledge accumulation and learning, the relationship between support activities and support resources can be analyzed.

[0031] Meanwhile, it incorporates the idea of ​​optimizing machine learning algorithms: First, it identifies multiple alternative resource allocation schemes and the optimal scheme for each support task in historical data; then, it calculates the impact of the priority of demand points, support time, and number of resource allocation points of these alternative and optimal schemes on the completion of the support task; next, it uses the impact of these factors to calculate the cost value of the alternative scheme (final scheme); finally, taking the same support task as the smallest unit, it establishes an RBF neural network to analyze the mapping relationship between the cost value of each alternative scheme and the cost value of the optimal scheme. For a support task, when the cost value of its alternative resource allocation schemes is input, the algorithm can determine the optimal support resource allocation scheme.

[0032] This application provides a mission-oriented helicopter support resource scheduling method, which includes the following steps: Step 1: Obtain historical maintenance record data; Step 2: Using historical maintenance record data, train the dynamic scheduling model for support resources to obtain the optimized dynamic scheduling model for support resources; Step 3: Obtain the requirements for scheduling and allocation of support resources, and obtain real-time status data of support resources from the support resource database; Step 4: Based on the needs for resource scheduling and allocation, and the real-time status data of resources, use the optimized dynamic scheduling model for resources to obtain an optimized resource allocation scheme. Step 5: Utilize newly added historical maintenance record data to continuously train and optimize the dynamic scheduling model for ensuring resources.

[0033] Step 1 specifically involves: Historical maintenance records are retrieved from the helicopter fleet's support information database via scheduled requests. These historical maintenance records refer to archived maintenance records within the support information system. Specifically, the maintenance records include helicopter flight mission data, helicopter status data, environmental data, and maintenance data. Descriptions of each data type are as follows: Helicopter mission data includes information such as the name, type, number of sorties, and flight duration of the helicopter flight mission; Helicopter status data includes helicopter MFL (Maintenance Fault List), PFL (Pilot Fault List), flight parameter interpretation conclusions, fault diagnosis results, trend analysis results, status monitoring conclusions, fault prediction results, and other data.

[0034] Environmental data includes meteorological data, time data, and environmental data of maintenance sites (bases, civil airports, YZ Airport, etc.) during helicopter missions.

[0035] Maintenance data includes maintenance project data and support resource usage data. Maintenance project data includes data on parts replacement, operational testing, visual inspection, disassembly and assembly testing, off-site inspection, off-site repair, in-situ repair, modifications, service support, maintenance procedures / manuals related to maintenance projects, and the duration and personnel required for each maintenance project. Support resource usage data includes data on the type and quantity of aircraft materials, tools, equipment, and airfield resources used in performing each maintenance project.

[0036] Step 2 is as follows: The historical maintenance record data is preprocessed to construct training and testing sets; an RBF neural network model is constructed, and the training dataset is used to train the dynamic scheduling model of support resources to obtain the optimized dynamic scheduling model of support resources. (a) Constructing the training and test sets The acquired historical maintenance record data was preprocessed and randomly divided into a training set and a test set (70% training set and 30% test set) and stored in a database. Data preprocessing methods included data format conversion, data quantification, and data cleaning.

[0037] Specific data format conversion refers to converting the original data format of maintenance records into the data format required for subsequent processing by this system, such as converting the character set from GB3212 to UTF-8; data quantification processing refers to converting the qualitative descriptive information in the original data into quantifiable data indicators, such as converting the qualitative description of maintenance task types into INT type numerical values; data cleaning refers to using various data cleaning algorithms to modify outliers, missing values, etc. in the original data to improve data quality, such as filling in analog data in helicopter status data using the mean method.

[0038] (b) Constructing an RBF neural network model The training of the mission-oriented helicopter support resource scheduling model is achieved using the RBF neural network algorithm. The RBF neural network is a feedforward neural network whose hidden layer neurons employ radial basis functions (RBFs), and its neuron topology is similar to other feedforward networks. It consists of three layers: the first layer is the input layer, which receives the sample data; the second layer is the hidden layer, where the number of neurons is determined by the specific research object; unlike other feedforward networks, the transfer function (radial basis function) of the hidden layer neurons in the RBF network is a local response function; and the third layer is the output layer, which outputs the network response.

[0039] The RBF neural network uses a Gaussian function as its radial basis function, and its activation function is expressed as: Formula 1:

[0040] in, For the P-th input sample, which is the input parameter extracted from a structured historical maintenance record, this is an n-dimensional vector, which can be represented as (P 1(直升机编号) P 2(任务类型) P 3(任务类型) P 4(飞行架次) P 5(飞行时长) P 6(故障编号+类型，需解析) P 7(特征参数ID+数值，需解析) ...P n(在库航材件号+数量，需解析) The algorithm is used to reduce the dimensionality of vectors with higher dimensions in the samples to improve training efficiency. For example, the aviation material data includes the part number and quantity information of the aviation materials in the warehouse. By setting the part number coding standard, the part number of the aviation materials is automatically encoded. The last two digits of the code represent the quantity of aviation materials in the warehouse. After the model training is completed, the aviation material data vector is parsed in the output layer to restore the part number and quantity of the aviation materials. Center of Gaussian function Let be the variance of the Gaussian function, representing the range of influence of the center of the Gaussian function. This is the Euclidean distance, also known as the Euclidean distance.

[0041] The output of the RBF network is: Formula 2:

[0042] in, Let p be the p-th input sample, where p = 1, 2, 3, ..., P, and P is the total number of samples. The center of the hidden layer node; The connection weights between the output layer and the hidden layer; i = 1, 2, 3, ..., l represents the number of nodes in the hidden layer; Let be the output of the j-th output node of the RBF neural network. The output value is a j-dimensional vector, which is the output variable set of the mission-oriented helicopter support resource scheduling model, and can be represented as (y 1(所需航材件号+数量,需解析) y 2所需工具件号+数量,需解析) y 3(所需场站资源编号+数量,需解析) y 4(所需燃油牌号+数量,需解析) y 5(所需液压油牌号+数量,需解析) y 6(所需器材编号+数量,需解析) ...y j(所需维修人员等级+数量,需解析) ).

[0043] The specific training process of the RBF network is as follows: Step 1: Use the K-Means clustering algorithm to obtain the centers of the radial basis functions.

[0044] a) Initialization: Randomly select l training samples from the training set, i.e., one historical maintenance record data mentioned above, and use it as the cluster center ci (i=1,2,…,l).

[0045] b) Group the input sample set according to the nearest neighbor principle: Based on the Euclidean distance between Xp and ci, divide Xp into various cluster sets (p=1,2,…,P).

[0046] c) Reorganize cluster centers: based on each cluster set The average value of the training samples is used to obtain the new cluster center ci. If ci is stable and does not change, then it is the final basis function center; otherwise, return to b) to continue execution.

[0047] Step 2: Calculate the variance.

[0048] If the RBF network selects a Gaussian function as the basis function, then the variance It can be represented as follows:

[0049] in, The maximum distance between cluster centers is max| - |; h is the scaling factor, which is an adjustable hyperparameter. The larger h is, the stronger the function's generalization ability; the smaller h is, the more sensitive the function is to business characteristics. The initial value is set to 1.2.

[0050] Step 3: Solve for the connection weights between the network output layer and the hidden layer.

[0051] The weights between the output layer and the hidden layer are calculated using the least squares method, with the specific formula as follows:

[0052] The above three steps complete the training of the mission-oriented helicopter support resource scheduling model.

[0053] (c) Verify the model training effect Set a confidence index for the mission-oriented helicopter support resource scheduling model. Read the aforementioned test set data, call the mission-oriented helicopter support resource scheduling model, and output an optimized support resource allocation scheme. Compare the support resource allocation scheme with the actual support resource usage data and calculate the confidence index of the model data results. When the confidence index reaches or exceeds the index requirement, publish the model and record the model version; when the confidence index is lower than the index requirement, return to step (b) and retrain the model by increasing the training set, optimizing the training set data preprocessing method, adjusting the activation function hyperparameters, etc., until the confidence index requirement is met.

[0054] Step 3 specifically involves: Real-time helicopter flight mission data, helicopter status data, environmental data, and maintenance data are read through external data interfaces to ensure resource scheduling and allocation needs; real-time status data of support resources are obtained from the support resource database, and input vector P (P1, P2, P3...P...) is generated according to step 2. n The external data interfaces include an interface with the ground health management system to obtain helicopter status data; and an interface with the support information system to obtain helicopter mission data, environmental data, and maintenance data.

[0055] Step 4 is as follows: Read the input vector generated in step 3, call the optimized helicopter support resource dynamic scheduling model, and output the support resource demand vector Y (y1, y2, y3...y... n Then, the data in the resource demand vector is used to call the corresponding dimensionality-upgrading algorithm to restore it to business data, generating an optimized resource configuration scheme. The dimensionality-upgrading algorithm is the reverse calculation of the dimensionality-reduction algorithm in step 1.

[0056] Step 5 specifically involves: Set a training sample threshold for the helicopter support resource scheduling model (set to 200 records, adjustable), read and store newly added historical maintenance record data. When the number of newly added data records reaches the set training sample threshold, proceed to steps 1-3 again to continuously optimize the helicopter support resource scheduling model.

[0057] The data details required for a mission-oriented helicopter support resource scheduling method are listed in the table below.

[0058] Table 1 Aircraft Material Status Table

[0059] Table 2 Aircraft Material Application Form

[0060] Table 3 Aircraft Material Requisition Form

[0061] Table 4 Aircraft Material Demand Table

[0062] Table 5 Tool and Equipment Status Table

[0063] Table 6 Tool and Equipment Application Form

[0064] Table 7 Tool and Equipment Requisition Form

[0065] Table 8 Tool and Equipment Requirements

[0066] Table 9: Status of Helicopters Supporting the Airfield

[0067] Table 10 Application Form for Helicopter Support for Airports

[0068] Table 11. Supply of Helicopters for Airport Support

[0069] Table 12 Helicopter Requirements for Airport Support

[0070] Table 13 Personnel Information Table

[0071] Table 14 Information on Support Facilities

[0072] (b) Output items The algorithm output is a resource allocation scheme.

[0073] Table 15 Information on Resource Allocation

[0074] The above detailed embodiments are a description of the present invention. It should not be considered that the specific embodiments of the present invention are limited to these descriptions. For those skilled in the art, several simple deductions and substitutions can be made without departing from the concept of the present invention, and all of these should be considered to fall within the protection scope of the present invention.

Claims

1. A mission-oriented helicopter support resource scheduling method, characterized in that, The method includes the following steps: Step 1: Obtain historical maintenance record data; Step 2: Using historical maintenance record data, train the dynamic scheduling model for support resources to obtain the optimized dynamic scheduling model for support resources; Step 3: Obtain the requirements for scheduling and allocation of support resources, and obtain real-time status data of support resources from the support resource database; Step 4: Based on the needs for resource scheduling and allocation, and the real-time status data of resources, use the optimized dynamic scheduling model for resources to obtain an optimized resource allocation scheme. Step 5: Utilize newly added historical maintenance record data to continuously train and optimize the dynamic scheduling model for ensuring resources.

2. The method according to claim 1, characterized in that, Step 1 is as follows: Historical maintenance records are retrieved from the helicopter fleet's support information database via scheduled requests; these historical maintenance records refer to the maintenance records that have been archived in the support information system.

3. The method according to claim 2, characterized in that, Historical maintenance records include helicopter flight mission data, helicopter status data, environmental data, and maintenance data. The descriptions of each type of data are as follows: Helicopter mission data includes the name, type, number of sorties, and flight duration of the helicopter flight mission; Helicopter status data includes helicopter maintenance fault list, pilot fault list, flight parameter interpretation conclusions, fault diagnosis results, trend analysis results, status monitoring conclusions, and fault prediction results; Environmental data includes meteorological data, time data, and maintenance site data during helicopter missions; Maintenance data includes maintenance project data and support resource usage data.

4. The method according to claim 3, characterized in that, Maintenance project data includes parts replacement, operation testing, visual inspection, disassembly and assembly testing, off-site inspection, off-site repair, in-situ repair, modification, service support, maintenance procedures / manuals related to maintenance projects, as well as the duration of each maintenance project and the personnel required; support resource usage data includes the type and quantity of aviation materials, tools and equipment, and airfield resources used to perform each maintenance project.

5. The method according to claim 4, characterized in that, Step 2 is as follows: Historical maintenance record data is preprocessed to construct training and testing sets; an RBF neural network model is constructed, and the training dataset is used to train the dynamic scheduling model for support resources, resulting in an optimized dynamic scheduling model for support resources.

6. The method according to claim 5, characterized in that, Step 3 specifically involves: Real-time helicopter flight mission data, helicopter status data, environmental data, and maintenance data are read through external data interfaces to ensure resource scheduling and allocation needs; real-time status data of support resources are obtained from the support resource database, and input vector P (P1, P2, P3...P...) is generated according to step 2. n The external data interfaces include an interface with the ground health management system to obtain helicopter status data; and an interface with the support information system to obtain helicopter mission data, environmental data, and maintenance data.

7. The method according to claim 6, characterized in that, Step 4 is as follows: Read the input vector generated in step 3, call the optimized helicopter support resource dynamic scheduling model, and output the support resource demand vector Y (y1, y2, y3...y... n Then, the data in the resource demand vector is used to call the corresponding dimensionality-upgrading algorithm to restore the business data and generate a resource configuration optimization plan; wherein, the dimensionality-upgrading algorithm is the reverse calculation of the dimensionality-reduction algorithm in step 1.

8. The method according to claim 7, characterized in that, Step 5 specifically involves: Set a training sample threshold for the helicopter support resource scheduling model, read and store newly added historical maintenance record data. When the number of newly added data records reaches the set training sample threshold, proceed to steps 1-3 again to continuously optimize the helicopter support resource scheduling model.

Images