Aircraft oxygen load distribution method based on dynamic database

Abstract

The invention discloses an aircraft oxygen load distribution method based on a dynamic database, and relates to the field of aviation system engineering and control engineering, and the method comprises the steps: building a digital twin model through simulation software, obtaining the current state of aircraft oxygen, presetting the oxygen demands at different flight stages, and obtaining a digital twin model; performing multi-physics field coupling optimization on the digital twin model through an optimization algorithm, feeding back an optimized result to a dynamic database to obtain an oxygen distribution scheme, training an oxygen demand model by utilizing a federated learning technology, predicting oxygen demand in a future flight stage, and designing and deploying a self-repairing mechanism according to a prediction result; according to the method, the oxygen consumption condition is monitored in real time through the dynamic reconstruction algorithm, and the oxygen distribution scheme is intervened and optimized when it is detected that the oxygen distribution scheme is not matched with the oxygen consumption condition, so that the requirements of emergencies are met, and the error-tolerant rate and the reliability are greatly improved.

Description

Technical Field

[0001] The present invention relates to the fields of aviation systems engineering and control engineering, and particularly to an aircraft oxygen load distribution method based on a dynamic database. Background Art

[0002] With the rapid development of the aviation industry, the safety and comfort of aircraft have become key considerations in design and operation. Against this backdrop, the optimization of oxygen supply systems has gradually received extensive attention. Traditional oxygen load distribution methods mainly rely on static configurations and preset parameters, which are typically based on fixed design standards and historical experience data, aiming to ensure the safety and health of passengers and crew under various flight conditions during flight.

[0003] Although existing oxygen load distribution technologies have improved flexibility and response speed to a certain extent, they still face some challenges in practical applications. Firstly, there is a lack of sufficient intelligence level and it is unable to automatically adjust the oxygen distribution strategy according to the real-time changing flight environment. In addition, although existing optimization algorithms can improve the oxygen distribution efficiency to a certain extent, they are overwhelmed when dealing with complex multi-physical field coupling problems and it is difficult to achieve a global optimal solution. Summary of the Invention

[0004] In view of the existing problems described above, the present invention is proposed.

[0005] Therefore, the present invention provides an aircraft oxygen load distribution method based on a dynamic database to solve the problem that there is a lack of sufficient intelligence level and it is unable to automatically adjust the oxygen distribution strategy according to the real-time changing flight environment.

[0006] To solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides an aircraft oxygen load distribution method based on a dynamic database, which includes collecting historical flight data and real-time data of the aircraft, preprocessing the real-time data and uploading it to the dynamic database;

[0008] Using simulation software to establish a digital twin model, obtaining the current state of the aircraft oxygen, and presetting the oxygen requirements under different flight phases;

[0009] Performing multi-physical field coupling optimization on the digital twin model through an optimization algorithm, and feeding the optimized result back to the dynamic database to obtain an oxygen distribution plan;

[0010] Using federated learning technology to train an oxygen demand model to predict the oxygen demand in future flight phases;

[0011] Designing and deploying a self-healing mechanism according to the prediction result;

[0012] Real-time detect the operating status of oxygen distribution, and optimize it using a dynamic reconstruction algorithm according to the operating status, and adjust the oxygen distribution plan.

[0013] As a preferred solution of the aircraft oxygen load distribution method based on a dynamic database according to the present invention, wherein: collect historical flight data and real-time data of the aircraft, and upload the preprocessed real-time data to the dynamic database, specifically including the following steps,

[0014] The real-time data of the aircraft includes oxygen concentration, oxygen consumption rate, remaining oxygen amount, cabin pressure, oxygen tank pressure, cabin temperature, oxygen tank temperature, flight phase, and external atmospheric pressure and temperature;

[0015] The historical flight data is the real-time data collected and stored in the dynamic database in the past;

[0016] Transmit the collected real-time data to the central processing unit through Ethernet, and use the TLS encryption protocol during transmission;

[0017] After preprocessing the real-time data by the central processing unit, store it in the dynamic database;

[0018] The preprocessing includes preliminary cleaning and outlier removal.

[0019] As a preferred solution of the aircraft oxygen load distribution method based on a dynamic database according to the present invention, wherein: use simulation software to establish a digital twin model using historical flight data and real-time data to obtain the current state of aircraft oxygen, specifically including the following steps,

[0020] Use MATLAB as a simulation platform to create a multi-physics field coupling model covering gas dynamics, thermodynamics, and fluid mechanics, and define the oxygen consumption rate and the remaining oxygen amount as core variables;

[0021] Input the real-time data of the aircraft into the multi-physics field coupling model to calculate the current state of aircraft oxygen.

[0022] As a preferred solution of the aircraft oxygen load distribution method based on a dynamic database according to the present invention, wherein: preset the oxygen requirements under different flight phases, specifically including the following steps,

[0023] According to the flight plan and the actual flight status, divide the flight process into different phases, and set a specific oxygen requirement change pattern for each phase;

[0024] Create a calculation module through the Simulink environment of MATLAB. Input the real-time data of the aircraft and the oxygen demand change patterns at each stage into the calculation module, output the oxygen demand under different flight stages, and use the mean square error to compare the error with the historical flight data until the error is less than the requirement.

[0025] The requirement refers to a predefined error threshold.

[0026] Feed back the current oxygen state of the aircraft and the oxygen demand under different flight stages to the dynamic database.

[0027] According to the flight plan and the actual flight status, divide the flight process into different stages and set specific oxygen demand change patterns for each stage.

[0028] Create a calculation module through the Simulink environment of MATLAB. Input the real-time data of the aircraft and the oxygen demand change patterns at each stage into the calculation module, output the oxygen demand under different flight stages, and use the mean square error to compare the error with the historical flight data until the error is less than the requirement.

[0029] Feed back the current oxygen state of the aircraft and the oxygen demand under different flight stages to the dynamic database.

[0030] As a preferred scheme of the aircraft oxygen load distribution method based on the dynamic database according to the present invention, wherein: perform multi-physical field coupling optimization on the digital twin model through an optimization algorithm, and feed back the optimized result to the dynamic database to obtain an oxygen distribution scheme, which specifically includes the following steps.

[0031] Define the optimization objective and constraint conditions, and introduce the particle swarm optimization algorithm.

[0032] Initialize the particle swarm, and set the initial position and velocity of each particle.

[0033] Calculate the fitness function value of each particle according to the optimization objective and constraint conditions.

[0034] Update the position and velocity of the particles through the fitness function values of each particle, repeat the iteration until the stop condition is met, obtain the oxygen distribution scheme and transmit it to the dynamic database.

[0035] As a preferred scheme of the aircraft oxygen load distribution method based on the dynamic database according to the present invention, wherein: use the federated learning technology to train the oxygen demand model to predict the oxygen demand in future flight stages, which specifically includes the following steps.

[0036] Select the participants of federated learning, and all participants prepare their local flight data sets.

[0037] Deploy the same machine learning framework at each participating party;

[0038] Initialize the local models of all participating parties with the same initial model parameters and define the global model structure;

[0039] Each participating party independently trains its local model using the local flight dataset and updates the model parameters using the standard gradient descent method;

[0040] After each participating party completes one round of local training, it uploads the updated model parameters to the central server. The central server aggregates the model parameters of all participating parties using the weighted average method, updates the global model parameters, and distributes the updated global model parameters back to each participating party for the next round of training until the construction of the oxygen demand model is completed;

[0041] Input the oxygen demand model into the dynamic database and output the predicted oxygen demand in the future flight phase.

[0042] As a preferred solution of the aircraft oxygen load distribution method based on the dynamic database according to the present invention, wherein: design and deploy a self-healing mechanism according to the prediction result, which specifically includes the following steps,

[0043] Real-time detect the pressure and oxygen flow rate in the aircraft oxygen supply pipeline, and analyze whether the current situation meets the expected working conditions in combination with the predicted oxygen demand in the future flight phase;

[0044] When it is detected that the current situation cannot meet the oxygen demand in the future flight phase, select the optimal alternative path through calculation and send an instruction to the intelligent valve to switch to the selected alternative path.

[0045] As a preferred solution of the aircraft oxygen load distribution method based on the dynamic database according to the present invention, wherein: real-time detect the operating status of the oxygen distribution situation, optimize it using the dynamic reconstruction algorithm according to the operating status, and adjust the oxygen distribution plan, which specifically includes the following steps,

[0046] Set the oxygen threshold according to the historical flight data;

[0047] Real-time detect the actual oxygen consumption rate in the cabin and calculate the optimal oxygen consumption rate of the oxygen distribution plan. When the gap between the optimal oxygen consumption rate and the actual oxygen consumption rate is greater than the oxygen threshold, start the reconstruction;

[0048] Define the oxygen tank, pipeline segments, valves, and cabin interfaces as nodes of a weighted directed graph based on graph theory, define the pipeline connection paths as edges, use the physical constraints and real-time flow requirements as edge weights, execute the Dijkstra algorithm to calculate the shortest path from the oxygen tank node to the cabin interface, output the optimal path, and update the oxygen distribution plan.

[0049] In a second aspect, the present invention provides a computer device, including a memory and a processor, where the memory stores a computer program, and: when the computer program is executed by the processor, any step of the method for aircraft oxygen load distribution based on a dynamic database as described in the first aspect of the present invention is implemented.

[0050] In a third aspect, the present invention provides a computer-readable storage medium, on which a computer program is stored, and: when the computer program is executed by the processor, any step of the method for aircraft oxygen load distribution based on a dynamic database as described in the first aspect of the present invention is implemented.

[0051] The beneficial effects of the present invention are as follows: By creating a highly accurate digital twin model, the real flight environment is simulated, and the oxygen demand at different flight stages of the aircraft is obtained. Then, through the application of the particle swarm optimization algorithm, the global optimal solution can be found in a complex multi-physical field environment, so as to generate an oxygen distribution plan, rather than a local optimal solution. This significantly improves the efficiency of oxygen distribution and resource utilization rate, reduces the operating cost, and through the dynamic reconstruction algorithm, the oxygen consumption is monitored in real time, and when it is detected that the oxygen distribution plan does not match the oxygen consumption, it intervenes to optimize the oxygen distribution plan, so as to meet the needs of emergencies, greatly increasing the fault tolerance rate and reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0053] Figure 1 It is a flowchart of the method for aircraft oxygen load distribution based on a dynamic database in Embodiment 1.

[0054] Figure 2 It is a schematic diagram for optimizing the oxygen distribution plan in Embodiment 1. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0055] In order to make the above objects, features, and advantages of the present invention more obvious and understandable, the specific embodiments of the present invention will be described in detail below with reference to the drawings in the specification.

[0056] Many specific details are set forth in the following description in order to provide a thorough understanding of the present invention, but the present invention may be implemented in other ways different from those described herein. Those skilled in the art can make similar extensions without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0057] Second, the "one embodiment" or "embodiment" referred to herein means a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The appearances of "in one embodiment" in different places in this specification do not all refer to the same embodiment, nor are they separate or alternative embodiments that exclude each other with other embodiments.

[0058] Embodiment 1, referring to Figure 1 and Figure 2 , is the first embodiment of the present invention. This embodiment provides a method for aircraft oxygen load distribution based on a dynamic database, including the following steps:

[0059] S1. Collect historical flight data and real-time data of the aircraft, and upload the preprocessed real-time data to the dynamic database.

[0060] Specifically, it includes the following steps.

[0061] Historical flight data is real-time data collected and stored in the dynamic database in the past. Historical flight data is not only the basis for building models but also can be used as a standard for validating and calibrating new models.

[0062] Collect parameters such as oxygen concentration, oxygen consumption rate, remaining oxygen amount, cabin pressure, oxygen tank pressure, cabin temperature, oxygen tank temperature, flight phase, and external atmospheric pressure and temperature of the aircraft through sensors, providing a comprehensive view of the current state of the aircraft to help make accurate judgments and adjustments.

[0063] Transmit the collected real-time data to the central processing unit through a secure and reliable communication protocol (such as TCP / IP). During this process, the TLS encryption protocol is used to protect the data from unauthorized access or tampering.

[0064] Further explanation, this step ensures the integrity and confidentiality of real-time data during transmission, preventing potential security threats, which is particularly important for aviation applications involving sensitive information.

[0065] After receiving the data, the central processing unit removes obviously incorrect data points, such as readings outside the reasonable range or inconsistent timestamps. Subsequently, statistical methods are used to identify and remove outliers to ensure the quality of the data set. High-quality data is the basis for building accurate models and helps improve the accuracy of subsequent simulations and optimization results.

[0066] Further explanation, high-quality data can significantly improve the accuracy of simulation and optimization models, reducing misjudgments and decision-making mistakes caused by data quality problems, thus making decision-making more scientific.

[0067] S2. Establish a digital twin model using simulation software to obtain the current state of the aircraft's oxygen and preset the oxygen requirements under different flight phases.

[0068] Specifically, it includes the following steps:

[0069] S2.1. With MATLAB as the core, build a multi-physics field coupling model integrating gas dynamics, thermodynamics, and fluid mechanics. Through coupling oxygen flow path analysis (gas dynamics), pipeline temperature rise and heat dissipation calculation (thermodynamics), and pressure loss modeling (fluid mechanics), achieve multi-dimensional simulation of oxygen delivery. MATLAB is a high-level technical computing language and interactive environment for algorithm development, data visualization, data analysis, and numerical calculation. It provides a rich library of mathematical functions and toolboxes, supporting complex mathematical modeling and high-precision numerical calculation. This is crucial for establishing a multi-physics field coupling model covering gas dynamics, thermodynamics, and fluid mechanics.

[0070] Oxygen flow path analysis refers to calculating the pressure and flow velocity distributions at each point along the pipeline through the ideal gas state equation, Bernoulli equation, and continuity equation. Simulate the oxygen flow path from the oxygen storage tank to the passenger cabin and consider the influence of air pressure changes on the flow velocity.

[0071] Pipeline temperature rise and heat dissipation calculation refers to collecting the properties (such as thermal conductivity, specific heat capacity, etc.) of relevant materials (such as pipeline materials, insulation materials, etc.) and environmental conditions (such as ambient temperature, wind speed, etc.), using the finite difference method to solve the heat conduction equation to obtain the pipeline wall temperature distribution, and combining Newton's cooling law to calculate the heat exchange amount between oxygen and the pipeline, and then determine the oxygen temperature distribution.

[0072] Pressure loss modeling refers to for each section of the pipeline and each pipeline connection, determining the corresponding geometric dimensions and local resistance coefficients, applying the Darcy-Weisbach formula and the local resistance coefficient method to calculate the total pressure loss, and adjusting the design parameters (such as increasing the pipeline diameter, selecting low-resistance valve parts, etc.) according to the calculation results to reduce the overall pressure loss.

[0073] Train the model parameters based on historical flight data. For example, compare the oxygen consumption rate curve recorded in history with the simulation results and use polynomial fitting to optimize the multi-physics field coupling model.

[0074] Preferably, compared with the traditional single-physics field approximation model, this solution combines multi-physics field coupling simulation, which can comprehensively and accurately simulate the complex physical processes inside the aircraft, ensuring that the simulation results are highly consistent with the actual situation.

[0075] The reason for taking the oxygen consumption rate and the remaining oxygen amount as the core variables is that they are directly related to the goals of safety, real-time performance, and resource optimization.

[0076] Input the real-time data of the aircraft from the dynamic database into the multi-physical field coupling model. The multi-physical field coupling model performs real-time calculations based on the real-time data of the aircraft to obtain the current state of oxygen (such as oxygen concentration, flow rate, remaining amount, etc.).

[0077] S2.2. Divide the flight process into different stages (such as takeoff, cruise, descent, landing, etc.) according to the flight plan and the actual flight status.

[0078] Set a specific oxygen demand change pattern for each flight stage. For example, the oxygen demand may be constant during the cruise stage, while the oxygen demand may gradually decrease during the descent stage. The following is an example. For the cruise stage, assume the oxygen consumption rate is 5 L / min; for the descent stage, the oxygen consumption rate gradually decreases to 3 L / min.

[0079] Preferably, compared with the traditional fixed oxygen supply method, the present invention can set a specific oxygen demand pattern according to the characteristics of different flight stages, better adapt to different flight conditions and requirements, avoid unnecessary resource waste, and improve the utilization rate of oxygen.

[0080] Build a calculation module including an input interface, an algorithm core, and an output interface in Simulink. Simulink is a graphical programming environment based on MATLAB. The input layer inputs the real-time data of the aircraft and the oxygen demand change pattern of each stage. The algorithm core uses a multivariate quadratic polynomial to fit the real-time data. The output layer outputs the oxygen demand under different flight stages, and uses the mean square error (MSE) to iteratively calculate the module parameters. Compare the demand of each stage output by the simulation with the historical data. If the error exceeds the set value (which can be customized according to the actual situation), continue to iterate, and stop iterating at most 1000 times or when the error converges within 0.01.

[0081] Upload the calculated current oxygen state of the aircraft and the oxygen demand under different flight stages to the dynamic database through a secure communication protocol. All relevant parties can obtain the latest oxygen state information in a timely manner, supporting efficient management and decision-making.

[0082] S3. Optimize the multi-physical field coupling of the digital twin model through an optimization algorithm, and feedback the optimized result to the dynamic database to obtain an oxygen distribution plan.

[0083] Specifically, it includes the following steps

[0084] S3.1. Take maximizing the oxygen utilization rate as the optimization goal to ensure that the oxygen supply in each flight stage meets the demand. The constraint conditions include physical constraints, safety constraints, and economic constraints.

[0085] Physical constraints refer to considering conditions such as the maximum capacity of the oxygen tank and the maximum flow rate of the pipeline;

[0086] The safety constraint means ensuring that the oxygen concentration in the cabin is not lower than the safety standard under any circumstances;

[0087] The economic constraint means considering cost factors and minimizing unnecessary resource waste.

[0088] S3.2. Optimize the digital twin model using the particle swarm optimization algorithm. Particle swarm optimization (PSO) is a stochastic search algorithm based on swarm intelligence that simulates the foraging behavior of birds. It searches for the global optimal solution by continuously updating the position and velocity of particles.

[0089] First, initialize the particle swarm. Each particle represents a possible solution (usually 50 - 200, the more, the wider the solution space covered), that is, an oxygen distribution plan. Second, distribute each particle evenly within the feasible region according to historical flight data to avoid local optima as the initial position; the initial velocity of each particle is randomly generated, and the initial velocity represents the moving direction and step size of the particle in the solution space.

[0090] Verify through the digital twin model integrating multiple physical fields to simulate the oxygen distribution results in the real scenario. If all constraint conditions are met, score according to the objective function, that is, the fitness function value of the particle. If a constraint condition is violated, it is determined as an infeasible solution.

[0091] For each particle, record the best position found since the start (i.e., the position with the highest fitness). This helps guide the particle towards a better solution. In the entire particle swarm, record the best positions found by all particles. This provides a common target direction for the entire group and helps the group as a whole converge towards the optimal solution. Update the position and velocity of each particle in this way until the maximum number of iterations is reached or the fitness change is less than a certain threshold (which can be customized according to the usage scenario and requirements). After the iteration stops, obtain the oxygen distribution plan and transmit the oxygen distribution plan to the dynamic database for subsequent use.

[0092] Further explanation, through the particle swarm optimization algorithm, the utilization of oxygen resources is made more efficient. This not only reduces unnecessary resource waste but also ensures sufficient oxygen supply under any circumstances. It significantly improves the oxygen utilization rate and reduces operating costs. And it considers various physical and safety constraint conditions to ensure that the obtained oxygen distribution plan is still effective under various complex conditions, greatly enhancing the reliability.

[0093] S4. Use federated learning technology to train the oxygen demand model to predict the oxygen demand in future flight phases.

[0094] Specifically, it includes the following steps,

[0095] S4.1. Select airlines, aircraft manufacturers, airport operators, third-party maintenance service providers, etc. as the participants in federated learning.

[0096] Each participant has its own independent flight dataset. For example, the flight logs (altitude, speed, cabin pressure, temperature), number of passengers, and oxygen consumption records of airlines; the aircraft design parameters (duct layout, oxygen tank capacity) of aircraft manufacturers; the historical failure reports (such as leakage incidents) and environmental meteorological data (air pressure, temperature gradient) of third-party maintenance service providers.

[0097] Furthermore, since the data does not need to leave the local environment, this greatly enhances the security and privacy of the data. And different participants come from different geographical locations or operating environments, providing diverse data, which helps to improve the generalization ability of the oxygen demand model.

[0098] S4.2. Deploy the TensorFlow Federated framework on each participant because it is suitable for large-scale distributed training and supports dynamic model aggregation. The unified framework enables each participant to follow the same standard for model training, reducing compatibility issues.

[0099] Provide the same initial model parameters for all participants. This can be achieved by generating a set of randomly initialized weights by the central server and distributing them to each participant. And determine the overall architecture of the local model, such as the number of layers, the number of neurons in each layer, etc. All participants build local models according to this architecture.

[0100] Furthermore, it ensures that all participants start training from the same starting point, avoiding performance fluctuations caused by initialization differences. The local model is the model that runs on a single participant's device in the federated learning framework, and its role is to aggregate the data of each participant's local model to form a final oxygen demand model.

[0101] S4.3. Each participant uses its local dataset to train the local model and updates the parameters of the local model using the Adam optimizer. After each iteration, the participant will obtain a new set of model parameters, which reflect the best fit on the current local dataset.

[0102] Furthermore, each participant can perform training in parallel on its own computing resources, making full use of the advantages of distributed computing.

[0103] After each participant completes a round of local training, it uses the RSA algorithm to encrypt the updated model parameters and uploads them to the central server to prevent eavesdropping or tampering during transmission.

[0104] After the central server receives the updated model parameters, it uses the weighted average method (or other aggregation strategies) to integrate the model parameters of all participants to form new global model parameters. The central server distributes the updated global model parameters back to each participant for use in the next round of training. This process continues until a termination condition is reached, which could be that the error is less than 1% after three consecutive rounds of training or a preset maximum of 100 rounds of training (specifically, it can be customized according to the usage scenario and requirements).

[0105] Integrate the finally trained oxygen demand model into the dynamic database. And predict the future oxygen demand based on the real-time data of the aircraft.

[0106] Preferably, compared with the traditional method of centralized raw data, which violates privacy regulations, relies on the data of a single institution, has weak model generalization ability, and requires each institution to independently build a data pipeline, the present invention only shares model parameters through federated learning, the raw data never leaves the local server, does not cause data leakage, and integrates the data of multiple different participants, enhancing applicability. It also shares the federated learning framework and reuses model components, not only greatly improving the accuracy of prediction, but also significantly reducing the development cost.

[0107] S5. Design and deploy a self-healing mechanism according to the prediction results.

[0108] Specifically, it includes the following steps

[0109] The pressure and oxygen flow rate in the aircraft oxygen supply pipeline are detected in real time through pressure sensors and flow sensors. Real-time monitoring enables any situation deviating from the normal operating range to be quickly detected, helping to prevent potential safety hazards.

[0110] Compare and analyze the prediction results of the oxygen demand model with the current actual oxygen flow rate and pressure situation. For example, if the prediction shows that the oxygen demand will increase significantly in a future stage, but the current flow rate does not show a corresponding increasing trend, it may indicate potential problems such as pipeline blockage or leakage.

[0111] Set a switching threshold. For example, if the actual oxygen flow rate is lower than 90% of the predicted demand, trigger the alarm mechanism. This step can identify potential problems in advance, allowing the operator to have enough time to react and avoid a crisis.

[0112] Select the optimal backup path through the weighted scoring method and send instructions to the intelligent valve (such as an electromagnetic pulse signal).

[0113] Furthermore, by detecting deviations in advance to estimate risks and quickly react, the probability of accidents can be reduced and safety can be improved.

[0114] S6. Real-time detect the operating status of oxygen distribution, and optimize it using a dynamic reconstruction algorithm according to the operating status, and adjust the oxygen distribution plan.

[0115] Specifically, it includes the following steps:

[0116] Extract the oxygen consumption records in historical flight data (such as the flow rate and time series during takeoff, cruise, and descent phases). Calculate the oxygen consumption benchmark values and their standard deviations for different flight phases, and set an oxygen threshold.

[0117] Collect the oxygen flow rate data in the cabin every second, and calculate the actual oxygen consumption rate. Based on the oxygen demand prediction model of federated learning, input the current flight parameters (such as altitude, number of passengers), and output the optimal oxygen consumption rate. If the difference between the actual oxygen consumption rate and the optimal oxygen consumption rate is greater than the oxygen threshold, start the reconstruction.

[0118] S6.1. Define the oxygen tank, pipeline segments, valves, and cabin interfaces as the nodes of a weighted directed graph; define the pipeline connection paths as edges (such as oxygen tank A → pipeline 1 → valve B → cabin interface C); use the physical constraints (for example, the maximum flow rate of the pipeline (such as 12 L / min), the minimum pressure (such as 50 psi)) and the real-time flow demand as the edge weights. Thus, complete the construction of the graph.

[0119] Apply the Dijkstra algorithm. The goal is to start from the oxygen tank node and find the shortest path (minimum total resistance or maximum flow rate) that meets the flow demand, and preferentially select the path with sufficient flow margin (if both pipelines can carry the demand, select the one with lower maintenance cost). Dynamically update the edge weights to reflect the real-time working conditions until a path combination that can meet the real-time flow demand and has the minimum energy loss is found to obtain the optimal path, and update the oxygen distribution plan to ensure that the subsequent oxygen supply is carried out according to the new path.

[0120] Further explanation: Real-time monitoring and dynamic reconstruction avoid oxygen shortage, dynamically match the flow rate and demand, reduce unnecessary consumption, and have a fast response speed, reducing the maintenance cost.

[0121] S7. Real-time detect the operating status of the aircraft, and quickly start the emergency response mechanism in case of an emergency.

[0122] Specifically, it includes the following steps:

[0123] Define situations such as oxygen leakage (abnormal flow rate), pipeline rupture, and valve failure as emergency events and classify them. For example, a first-level event directly threatens flight safety and requires immediate handling; a second-level event may affect flight safety and requires handling within a limited time (e.g., activate the emergency plan within 1 minute). Preset emergency plans, such as the oxygen failure emergency plan, which need to include trigger conditions (detect that the oxygen flow rate continuously drops > 30% or the pipeline pressure is below the safety threshold (e.g., 50 psi)), handling measures (close the oxygen valve in the faulty area, automatically switch to the standby oxygen tank (such as the crew oxygen cylinder or the emergency oxygen supply pipeline), and start cabin pressurization to prevent hypoxia). And based on the current state of the aircraft (current position, altitude, speed, remaining fuel quantity), geographical data (list of nearby airports, status of alternate landing fields, terrain obstacles), and meteorological data (real-time wind speed, visibility, etc.), use the A* algorithm to calculate the best emergency landing location;

[0124] Send notifications of emergency events to the pilot through voice alerts and ECAM (Electronic Cabin Visual Alert Function), and send encrypted messages (such as ADS-B broadcasts) to the ground control center.

[0125] Afterwards, record all relevant event information (such as event type, occurrence time, handling measures, etc.) and the subsequent handling process in detail in the dynamic database for post-event analysis and improvement.

[0126] Furthermore, by defining emergency events and a hierarchical response mechanism, not only can potential problems be effectively prevented, but also actions can be taken promptly when problems occur to ensure that emergency situations during flight are properly handled, greatly improving the safety and reliability of flight.

[0127] This embodiment also provides a computer device applicable to the case of the aircraft oxygen load distribution method based on a dynamic database, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the aircraft oxygen load distribution method based on a dynamic database as proposed in the above embodiment.

[0128] The computer device can be a terminal, which includes a processor, a memory, a communication interface, a display screen, and an input device connected via a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The communication interface of the computer device is used to communicate with external terminals in a wired or wireless manner, and the wireless manner can be achieved through WIFI, carrier networks, NFC (Near Field Communication), or other technologies. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads provided on the outer shell of the computer device, or an external keyboard, touchpad, or mouse, etc.

[0129] This embodiment also provides a storage medium, on which a computer program is stored. When the program is executed by a processor, it implements the method for aircraft oxygen load distribution based on a dynamic database as proposed in the above embodiment; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, abbreviated as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, abbreviated as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, abbreviated as EPROM), programmable read-only memory (Programmable Red-Only Memory, abbreviated as PROM), read-only memory (Read-Only Memory, abbreviated as ROM), magnetic memory, flash memory, magnetic disks, or optical discs.

[0130] In summary, through the creation of a highly accurate digital twin model, the present invention simulates the real flight environment and obtains the oxygen demand of the aircraft at different flight stages. Then, through the application of the particle swarm optimization algorithm, it can find the global optimal solution in a complex multi-physical field environment, thereby generating an oxygen distribution plan instead of a local optimal solution. This significantly improves the efficiency of oxygen distribution and resource utilization rate, reduces the operating cost, and through the dynamic reconstruction algorithm, it monitors the oxygen consumption situation in real time and intervenes when it detects that the oxygen distribution plan does not match the oxygen consumption situation to optimize the oxygen distribution plan, so as to meet the needs of emergencies, greatly increasing the fault tolerance rate and reliability.

[0131] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. A method for distributing oxygen load on an aircraft based on a dynamic database, characterized in that: include, Collect historical flight data and real-time data of aircraft, and upload the real-time data to the dynamic database after pre-processing; Use simulation software to build a digital twin model to obtain the current status of aircraft oxygen and preset oxygen requirements in different flight phases; The digital twin model is optimized through multi-physics coupling through optimization algorithms, and the optimized results are fed back to the dynamic database to obtain the oxygen distribution plan; Use federated learning technology to train an oxygen demand model to predict oxygen demand in future flight phases; Design and deploy self-healing mechanisms based on prediction results; The operating status of oxygen distribution is detected in real time, and the dynamic reconstruction algorithm is used to optimize and adjust the oxygen distribution plan according to the operating status.

2. The method for distributing oxygen load on an aircraft based on a dynamic database as claimed in claim 1, characterized in that: Collect historical flight data and real-time data of the aircraft, and upload the real-time data to the dynamic database after pre-processing. The specific steps include the following: The real-time data of the aircraft includes oxygen concentration, oxygen consumption rate, remaining oxygen amount, cabin pressure, oxygen tank pressure, cabin temperature, oxygen tank temperature, flight phase and external atmospheric pressure and temperature; The historical flight data is real-time data collected and stored in a dynamic database in the past; The collected real-time data is transmitted to the central processing unit via Ethernet, and the TLS encryption protocol is used during transmission; The central processing unit pre-processes the real-time data and stores it in a dynamic database; The preprocessing includes preliminary cleaning and removal of outliers.

3. The method for distributing oxygen load on an aircraft based on a dynamic database as claimed in claim 2, characterized in that: Use simulation software to build a digital twin model to obtain the current status of aircraft oxygen, which includes the following steps: Use MATLAB as a simulation platform to create a multi-physics coupling model covering gas dynamics, thermodynamics, and fluid mechanics, and define oxygen consumption rate and residual oxygen amount as core variables; The aircraft's real-time data is input into the multi-physics coupling model to calculate the aircraft's current oxygen status.

4. The method for distributing oxygen load on an aircraft based on a dynamic database as claimed in claim 3, characterized in that: The oxygen demand under different flight phases is preset, specifically including the following steps: According to the flight plan and actual flight status, the flight process is divided into different stages, and a specific oxygen demand change mode is set for each stage; Create a calculation module using MATLAB's Simulink environment, input the aircraft's real-time data and the oxygen demand change pattern at each stage into the calculation module, output the oxygen demand at different flight stages, and use the mean square error to compare the error with historical flight data until the error is less than the requirement; The requirement refers to a predefined error threshold; The current oxygen status of the aircraft and the oxygen demand in different flight phases are fed back to the dynamic database.

5. The method for distributing oxygen load on an aircraft based on a dynamic database as claimed in claim 4, characterized in that: The digital twin model is optimized through multi-physical field coupling through the optimization algorithm, and the optimized results are fed back to the dynamic database to obtain the oxygen distribution plan, which specifically includes the following steps: Define optimization objectives and constraints, and introduce particle swarm optimization algorithm; Initialize the particle swarm and set the initial position and velocity of each particle; According to the optimization objectives and constraints, calculate the fitness function value of each particle; The position and velocity of each particle are updated by the fitness function value, and the iteration is repeated until the stop condition is met. The oxygen distribution plan is obtained and transmitted to the dynamic database.

6. The method for distributing oxygen load on an aircraft based on a dynamic database as claimed in claim 5, characterized in that: Federated learning technology is used to train the oxygen demand model to predict the oxygen demand in future flight phases. The specific steps include: Select the participants of federated learning, and all participants prepare local flight datasets; Deploy the same machine learning framework on each participant; Initialize the local models of all participants using the same initial model parameters and define the global model structure; Each participant independently trains a local model using a local flight dataset and updates the model parameters using a standard gradient descent method; After completing a round of local training, each participant uploads the updated model parameters to the central server. The central server aggregates the model parameters of each participant using the weighted average method, updates the global model parameters, and distributes the updated global model parameters back to each participant for the next round of training until the construction of the oxygen demand model is completed. The oxygen demand model is input into a dynamic database and outputs the predicted oxygen demand for future flight phases.

7. The method for distributing oxygen load on an aircraft based on a dynamic database as claimed in claim 6, characterized in that: Design and deploy a self-repair mechanism based on the prediction results, which includes the following steps: Real-time detection of the pressure and oxygen flow rate in the aircraft's oxygen supply pipeline, combined with the predicted oxygen demand in the future flight phase, to analyze whether the current operating conditions meet the expected conditions; When it is detected that the oxygen demand in the future flight phase cannot be met at present, the optimal backup path is selected through calculation, and a command is sent to the smart valve to switch to the selected backup path.

8. The method for distributing oxygen load on an aircraft based on a dynamic database as claimed in claim 7, characterized in that: Real-time detection of the operating status of oxygen distribution, optimization using dynamic reconstruction algorithm according to the operating status, and adjustment of oxygen distribution plan, including the following steps: Setting oxygen thresholds based on historical flight data; Real-time detection of the actual oxygen consumption rate in the cabin and calculation of the optimal oxygen consumption rate for the oxygen distribution plan. When the difference between the optimal oxygen consumption rate and the actual oxygen consumption rate is greater than the oxygen threshold, the reconstruction is initiated. Based on graph theory, oxygen tanks, pipeline segments, valves and cabin interfaces are defined as nodes of a weighted directed graph, pipeline connection paths are defined as edges, physical constraints and real-time flow requirements are used as edge weights, and the Dijkstra algorithm is executed to calculate the shortest path from the oxygen tank node to the cabin interface. The optimal path is output and the oxygen distribution plan is updated.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the aircraft oxygen load distribution method based on a dynamic database according to any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the aircraft oxygen load distribution method based on a dynamic database according to any one of claims 1 to 8 are implemented.

Images