An aerodynamic simulation analysis method based on a butterfly flapping-wing aircraft

Abstract

This invention relates to the field of aerodynamic simulation technology for flapping-wing aircraft, and discloses an aerodynamic simulation analysis method based on a butterfly flapping-wing aircraft. The method includes: automatically generating dynamic simulation scene configuration instructions based on a pre-built biomimetic flight state library; parsing the instructions to obtain scene parameters; obtaining the real-time geometric configuration sequence of the aircraft through a geometric model solver; extracting configuration data for a complete flapping-wing cycle and generating an initial flow field distribution; initiating flow field evolution iterations and simultaneously recording intermediate flow field data; and performing convergence judgment through independent processes, outputting the final simulation result when conditions are met. This method achieves automated configuration of simulation parameters and ensures strict consistency between flow field initialization and the actual motion cycle, thereby improving simulation efficiency and the physical realism of the results.

Description

Technical Field

[0001] This invention relates to the field of aerodynamic simulation technology for flapping-wing aircraft, specifically to an aerodynamic simulation analysis method based on butterfly flapping-wing aircraft. Background Technology

[0002] In the aerodynamic research of biomimetic flapping-wing aircraft, existing simulation methods typically rely on manually setting the butterfly's motion trajectory and simulation parameters. Since butterfly flapping involves complex motions encompassing flapping, twisting, and deformation, and varies significantly across different flight modes, this manual configuration method is inefficient, highly dependent on experience, and makes it difficult to ensure the consistency and comparability of parameters across different simulation tasks.

[0003] A further drawback lies in the setting of the initial conditions for flow field calculations. To simplify the process, conventional methods often use static models or simplified periodic motions as the starting point for flow field iterations. However, the actual aerodynamic effects depend on the dynamically changing geometric configuration of the wings over a complete cycle. Starting calculations from atypical initial states not only prolongs the convergence time but may also cause the results to deviate from the true physical laws, affecting the accurate analysis of vortex structures and aerodynamic mechanisms.

[0004] There is a need for a method that can automatically construct simulation scenarios of different biomimetic flight states and accurately initialize the flow field based on the geometric features of a real and complete motion cycle, in order to solve the core problems of low configuration efficiency and insufficient physical realism of the existing technology. Summary of the Invention

[0005] The purpose of this invention is to provide an aerodynamic simulation analysis method based on a butterfly flapping-wing aircraft to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, this invention provides an aerodynamic simulation analysis method based on a butterfly flapping-wing aircraft, the method comprising:

[0007] Construct a biomimetic flight state library corresponding to a butterfly flapping-wing aircraft, and generate configuration instructions for a dynamic simulation scene based on several predefined biomimetic flight states in the biomimetic flight state library.

[0008] The configuration instructions are parsed to obtain the scene parameters of the dynamic simulation scene in the time series. The scene parameters are then input into a preset geometric model solver to obtain the real-time geometric configuration sequence of the butterfly flapping-wing aircraft in the dynamic simulation scene.

[0009] The configuration data of a complete flapping cycle is extracted from the real-time geometric configuration sequence and transmitted to the flow field initialization unit. In the flow field initialization unit, an initial flow field distribution is generated based on the configuration data.

[0010] A flow field evolution process is initiated. In the flow field evolution process, the initial flow field distribution is used as the starting condition to drive the flow field of the butterfly flapping-wing aircraft to be updated iteratively, and the intermediate flow field data generated by each iteration update is recorded synchronously.

[0011] A result filtering process is established. In the result filtering process, intermediate flow field data from the flow field evolution process are continuously received. The convergence of the intermediate flow field data is judged. When the judgment result meets the preset conditions, the flow field data at the current time is output as the final aerodynamic simulation result.

[0012] Preferably, the step of generating configuration instructions for a dynamic simulation scenario based on a plurality of predefined biomimetic flight states in the biomimetic flight state library further includes:

[0013] The system identifies the user's flight intention command for the butterfly flapping-wing aircraft and matches the target bionic flight state in the bionic flight state library based on the flight intention command.

[0014] Extract all kinematic constraints contained in the target biomimetic flight state, and encode the kinematic constraints into time-driven commands;

[0015] The time-driven commands are fused with the inherent physical constraints of the butterfly flapping-wing aircraft to generate a flight control sequence that includes conflict resolution strategies. This flight control sequence constitutes the configuration commands for the dynamic simulation scenario.

[0016] Preferably, the step of inputting the scene parameters into a preset geometric model solver to obtain the real-time geometric configuration sequence of the butterfly flapping-wing aircraft in the dynamic simulation scene further includes:

[0017] The scene parameters are decomposed into two independent data streams: spatial transformation parameters and flexible deformation parameters.

[0018] The spatial transformation parameters are input into the rigid body kinematics calculation module to calculate the basic displacement and rotation data of the butterfly flapping-wing aircraft in the dynamic simulation scenario.

[0019] The flexible deformation parameters are input into the elasticity calculation module to calculate the wing surface deformation data of the butterfly flapping-wing aircraft.

[0020] A spatiotemporal mapping coordinator is established, in which the basic displacement and rotation data, the airfoil deformation data, and the physical timestamps of the dynamic simulation scene are synchronized and aligned to synthesize the real-time geometric configuration sequence.

[0021] Preferably, the step of extracting configuration data for a complete flapping cycle from the real-time geometric configuration sequence and transmitting the configuration data to the flow field initialization unit further includes:

[0022] A periodic detection operation is performed on the real-time geometric configuration sequence to identify the boundary time points where configuration features recur, and the complete flapping wing cycle is defined by the boundary time points.

[0023] Within the complete flapping wing cycle, multiple time slices are extracted according to a preset sampling density, and each time slice corresponds to a set of configuration data;

[0024] For each set of configuration data, perform data desensitization to remove sensitive geometric feature information, and simultaneously calculate the geometric complexity confidence level of that set of configuration data;

[0025] Only multiple sets of configuration data with a geometric complexity confidence level exceeding a preset threshold are packaged and transmitted to the flow field initialization unit in chronological order.

[0026] Preferably, the step of iteratively updating the flow field driving the butterfly flapping-wing aircraft further includes:

[0027] The current computational energy state of the flow field evolution process is monitored, and the current computational energy state is determined by the iteration step size and grid resolution of the flow field evolution process.

[0028] Based on the current computational energy state, an energy allocation strategy is dynamically scheduled. According to the energy allocation strategy, computational resources are allocated proportionally from the available computational resource pool of the butterfly flapping-wing aircraft to the flow field evolution process.

[0029] Before each iteration update begins, it is verified whether the allocated computing resources meet the estimated resource requirements for this iteration update. If they do, the iteration update operation is executed, and the actual amount of computing resources consumed by this iteration update operation is recorded in the resource consumption log.

[0030] Preferably, the step of establishing a result filtering process, in which intermediate flow field data is continuously received from the flow field evolution process, further includes:

[0031] Configure a dedicated data buffer for the result filtering process to temporarily store intermediate flow field data flowing out from the flow field evolution process;

[0032] Within the data buffer, a timestamp sorting operation is performed on the intermediate flow field data to ensure that the intermediate flow field data is arranged in iteration order;

[0033] A separate data credibility assessment subprocess is started. In the data credibility assessment subprocess, the physical rationality of the sorted intermediate flow field data is checked, and abnormal data points that do not conform to physical laws are removed.

[0034] The intermediate flow field data queue, after being processed by the data credibility assessment subprocess, is submitted to the convergence discrimination step for use.

[0035] Preferably, the step of establishing a spatiotemporal mapping coordinator, in which the basic displacement and rotation data, the wing surface deformation data, and the physical timestamps of the dynamic simulation scene are synchronized and aligned, further includes:

[0036] Establish a global timeline and mark the physical timestamp as a reference on the global timeline;

[0037] The data generation times corresponding to the basic displacement and rotation data and the airfoil deformation data are respectively mapped onto the global time axis;

[0038] The alignment deviation between the data points of the basic displacement and rotation data and the data points of the airfoil deformation data on the global time axis after detection and mapping is determined.

[0039] When the alignment deviation exceeds the allowable tolerance range, a data interpolation procedure is triggered. In the data interpolation procedure, new data points are generated to compensate for the alignment deviation based on the data change trends of the basic displacement and rotation data and the airfoil deformation data, until data synchronization is completed.

[0040] Preferably, the step of performing data desensitization operation on each group of configuration data to remove sensitive geometric feature information further includes:

[0041] Load a predefined sensitive feature recognition rule base, and use the sensitive feature recognition rule base to scan the configuration data;

[0042] Identify geometric details in the configuration data that are unrelated to the core aerodynamic performance of the butterfly flapping-wing aircraft, and mark these geometric details as sensitive geometric features to be removed;

[0043] A geometric smoothing algorithm is used to process the marked region, replacing the sensitive geometric feature information with a smooth geometric surface to generate desensitized configuration data;

[0044] The anonymized configuration data is compared with the original configuration data to generate a data modification report, which is then associated with and stored in the metadata of the anonymized configuration data.

[0045] Preferably, the step of extracting all kinematic constraints contained in the target biomimetic flight state and encoding the kinematic constraints into time-driven commands further includes:

[0046] The kinematic constraints are classified into three types: trajectory constraints, attitude constraints, and phase constraints.

[0047] A trajectory planner is assigned to the trajectory constraint, in which the continuous flight trajectory is discretized into a series of waypoints;

[0048] An attitude solver is assigned to the attitude constraint, and the attitude solver calculates the spatial orientation of the butterfly flapping-wing aircraft at the corresponding position based on the path points.

[0049] A phase synchronizer is assigned to the phase constraint, in which the timing relationship of motion between multiple flapping wing components of the butterfly flapping wing aircraft is coordinated;

[0050] The path points, spatial orientation, and motion timing relationships are encapsulated according to the timeline of the dynamic simulation scene to form the time-driven instructions.

[0051] Preferably, the step of monitoring the current calculated energy state of the flow field evolution process further includes:

[0052] Define a computational energy metric, which is obtained by weighting a computation time factor and a computational accuracy factor.

[0053] During the flow field evolution process, the specific values ​​of the iteration step size and the grid resolution are periodically collected;

[0054] The collected iteration step size and grid resolution values ​​are substituted into the computational energy metric standard for calculation, and the quantized value of the current computational energy state is output in real time.

[0055] The quantized value of the current calculated energy state is compared with the historical energy state record. Based on the comparison result, the energy change trend of the flow field evolution process in subsequent iterations is predicted, and the energy allocation strategy is adjusted in advance according to the energy change trend.

[0056] Compared with the prior art, the beneficial effects of the present invention are:

[0057] A pre-built biomimetic flight state library is used to automatically generate configuration instructions for dynamic simulation scenarios based on the predefined flight states in the library. This approach transforms the different flight modes of butterflies into callable standardized configurations, replacing the method of relying entirely on manual input of motion parameters. This achieves the proceduralization and batch processing of simulation scenario settings, enabling the rapid and accurate construction of high-fidelity dynamic simulation environments that closely match specific flight missions, thus improving the efficiency of simulation research and the consistency of comparative analysis between different states.

[0058] The geometric model solver processes the scene parameters over time to obtain a real-time geometric configuration sequence, from which configuration data for a complete flapping wing cycle is extracted for initializing the flow field. This scheme ensures that the starting point of the flow field calculation is strictly aligned with the periodic physical nature of the flapping wing motion. Using the real, continuously changing airfoil geometry during the complete cycle as the initial condition for the flow field calculation allows the simulation iteration to evolve from a point that highly approximates the real physical state. This avoids the lengthy and potentially deviating convergence process caused by starting calculations from a zero flow field or an arbitrary guessed field, allowing the flow field to enter a stable periodic solution more quickly, improving computational efficiency and the reliability of the results, and laying the foundation for accurate analysis of periodic average aerodynamic forces and transient vortex structures. Attached Figure Description

[0059] Figure 1 This is a schematic diagram illustrating the working principle of the aerodynamic simulation analysis method for butterfly flapping-wing aircraft described in this invention.

[0060] Figure 2 A flowchart for configuring instructions to generate dynamic simulation scenarios;

[0061] Figure 3 Flowchart for generating real-time geometry configuration sequences;

[0062] Figure 4 A graph showing the confidence level of geometric complexity in the aerodynamic simulation of a butterfly flapping-wing aircraft;

[0063] Figure 5 Convergence analysis diagram for flow field simulation of a butterfly flapping-wing aircraft. Detailed Implementation

[0064] 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, and 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.

[0065] Please see Figure 1This invention provides an aerodynamic simulation analysis method based on a butterfly flapping-wing aircraft. The method includes: constructing a biomimetic flight state library corresponding to the butterfly flapping-wing aircraft; generating configuration instructions for a dynamic simulation scene based on several predefined biomimetic flight states in the library; parsing the configuration instructions to obtain scene parameters of the dynamic simulation scene in a time series; inputting the scene parameters into a preset geometric model solver to obtain a real-time geometric configuration sequence of the butterfly flapping-wing aircraft in the dynamic simulation scene; extracting configuration data for a complete flapping cycle from the real-time geometric configuration sequence; transmitting the configuration data to a flow field initialization unit; generating an initial flow field distribution based on the configuration data in the flow field initialization unit; and initiating a flow field evolution process, using the initial flow field distribution as the starting condition to drive iterative updates of the butterfly flapping-wing aircraft's flow field, synchronously recording the intermediate flow field data generated by each iteration. Establish a result filtering process. In the result filtering process, intermediate flow field data from the flow field evolution process are continuously received. Convergence judgment is performed on the intermediate flow field data. When the judgment result meets the preset conditions, the flow field data at the current moment is output as the final aerodynamic simulation result.

[0066] Example 1: See Figure 2 The system identifies the user's flight intention commands for the butterfly flapping-wing aircraft and matches the target bionic flight state in a bionic flight state library based on these commands. It extracts all kinematic constraints contained in the target bionic flight state and categorizes them into three types: trajectory constraints, attitude constraints, and phase constraints. A trajectory planner is assigned to the trajectory constraints, discretizing the continuous flight trajectory into a series of path points. An attitude solver is assigned to the attitude constraints, determining the spatial orientation of the butterfly flapping-wing aircraft at the corresponding location based on the path points. A phase synchronizer is assigned to the phase constraints, coordinating the temporal relationships of motion between multiple flapping-wing components of the butterfly flapping-wing aircraft. The path points, spatial orientation, and temporal relationships are encapsulated according to the timeline of the dynamic simulation scene to form time-driven commands. These time-driven commands are then fused with the inherent physical constraints of the butterfly flapping-wing aircraft to generate a flight control sequence containing conflict resolution strategies. This flight control sequence constitutes the configuration commands for the dynamic simulation scene.

[0067] In practical implementation, the flight intention command can be a specific flight mode command, such as "rapid climb" or "smooth hovering." In a specific implementation, after receiving the "rapid climb" command, the system matches the target bionic flight state in a bionic flight state database based on the flight intention command. This database pre-stores a set of calibrated motion pattern data associated with various flight intentions. In some embodiments, the matching process is accomplished by querying a mapping table between flight intentions and bionic flight state numbers. The target bionic flight state is associated with a data structure containing specific kinematic parameters.

[0068] All kinematic constraints of the target biomimetic flight state are extracted and encoded into time-driven commands. In practice, the encoding process first categorizes kinematic constraints into three types: trajectory constraints, attitude constraints, and phase constraints. Trajectory constraints define the spatial path of the butterfly flapping-wing aircraft's center of mass; attitude constraints define the spatial orientation of the aircraft's body; and phase constraints define the time difference relationships between the motions of multiple flapping-wing components. A trajectory planner is assigned to the trajectory constraints. In this planner, the continuous flight trajectory is discretized into a series of path points. The coordinate sequence of these path points and their corresponding timestamps are the core output of the trajectory planner. In practice, for the "rapid climb" command, the coordinate increment of the path point set generated by the trajectory planner in the vertical direction is significantly greater than that in the horizontal direction.

[0069] An attitude solver is assigned to the attitude constraints. In this solver, the spatial orientation of the butterfly flapping-wing aircraft at corresponding locations is determined based on path points. Optionally, the attitude solver calculates the roll, pitch, and yaw angles of the butterfly flapping-wing aircraft based on the trajectory tangent direction at each path point and a preset flight angle of attack model. The flight angle of attack model, as a key component of the kinematic constraints in the biomimetic flight state library, is specifically implemented by coupling the aerodynamic characteristics of the butterfly flapping-wing aircraft with trajectory kinematics. The attitude solver first parses the geometric information in the path point sequence, calculating the trajectory tangent direction at each path point as a reference vector. Then, it queries the flight angle of attack model library for the angle of attack parameters corresponding to the target biomimetic flight state. These parameters are pre-calibrated using kinematic data from typical butterfly flight patterns, defining the ideal angle relationship between the wing plane and the relative incoming flow. In some embodiments, the attitude solver employs an interpolation algorithm to smoothly transition between preset key attitudes based on timestamps. A phase synchronizer is assigned to the phase constraint. In the phase synchronizer, the motion timing relationship between multiple flapping wing components of the butterfly flapping wing aircraft is coordinated. The phase synchronizer will assign periodic motion commands with a fixed phase difference to each flapping wing component based on the signal of a master oscillator.

[0070] The path points, spatial orientation, and motion timing relationships are encapsulated according to the timeline of the dynamic simulation scene to form time-driven commands. In specific implementation, the encapsulation operation packages time, position, attitude, and flapping phase into data frames according to a unified timestamp format. It can be understood that the time-driven command is a data stream strictly arranged in a time sequence. The time-driven command is then fused with the inherent physical constraints of the butterfly flapping-wing aircraft to generate a flight control sequence containing conflict resolution strategies. These physical constraints include, but are not limited to, the structural strength limits, joint range of motion limits, and drive power limits of the butterfly flapping-wing aircraft. In specific implementation, if a flapping joint angle calculated in the time-driven command exceeds the upper limit of the joint range of motion defined in the physical constraints, the conflict resolution strategy will correct the angle value to the upper limit and recalculate the relevant motion parameters at adjacent time points to ensure continuity. The flight control sequence constitutes the configuration commands for the dynamic simulation scene.

[0071] Example 2: See Figure 3 The scene parameters are decomposed into two independent data streams: spatial transformation parameters and flexible deformation parameters. The spatial transformation parameters are input into the rigid body kinematics calculation module to calculate the basic displacement and rotation data of the butterfly flapping-wing aircraft in the dynamic simulation scene. The flexible deformation parameters are input into the elasticity calculation module to calculate the wing deformation data of the butterfly flapping-wing aircraft. A spatiotemporal mapping coordinator is established, and a global time axis is created within it, with physical timestamps used as reference markers on the global time axis. The generation times of the basic displacement and rotation data and the wing deformation data are mapped onto the global time axis respectively. The alignment deviation between the data points of the basic displacement and rotation data and the data points of the wing deformation data on the global time axis after mapping is detected. When the alignment deviation exceeds the allowable tolerance range, a data interpolation program is triggered. In the data interpolation program, new data points are generated according to the data change trends of the basic displacement and rotation data and the wing deformation data to compensate for the alignment deviation until data synchronization is completed. The basic displacement and rotation data, airfoil deformation data, and physical timestamps of the dynamic simulation scene are synchronized and aligned to synthesize a real-time geometric configuration sequence.

[0072] In practical implementation, an aerodynamic simulation analysis method based on a butterfly flapping-wing aircraft involves decomposing scene parameters into two independent data streams: spatial transformation parameters and flexible deformation parameters. The scene parameters originate from the configuration instructions of the dynamic simulation scene. The spatial transformation parameters describe the overall translational and rotational motion of the butterfly flapping-wing aircraft in space, while the flexible deformation parameters describe the shape changes of the elastic components of the butterfly flapping-wing aircraft's wing surface. In practice, the spatial transformation parameters can be a dataset containing time-series position coordinates and Euler angles, while the flexible deformation parameters can be a series of function coefficients describing the changes in wing surface curvature and twist angle over time.

[0073] The spatial transformation parameters are input into the rigid body kinematics solution module to calculate the basic displacement and rotation data of the butterfly flapping-wing aircraft in the dynamic simulation scenario. In specific implementations, the rigid body kinematics solution module, based on the position and attitude information in the spatial transformation parameters, solves for the spatial coordinates of each preset rigid connection point on the butterfly flapping-wing aircraft's body frame at each moment. These connection point coordinates can be understood as forming the basic skeleton of the butterfly flapping-wing aircraft's motion. In some embodiments, the rigid body kinematics solution module applies a homogeneous coordinate transformation matrix to handle continuous motion.

[0074] The flexible deformation parameters are input into the elasticity calculation module to calculate the wing deformation data of the butterfly flapping-wing aircraft. In practice, the elasticity calculation module, based on the pre-established finite element model of the wing and the input flexible deformation parameters, solves for the deformation displacement of each node on the wing mesh relative to its reference position. It can be understood that the wing deformation data supplements the flexible details missing in rigid body motion.

[0075] A spatiotemporal mapping coordinator is established, and a global time axis is created within it. The physical timestamp is used as the reference mark on the global time axis, which serves as the unified reference system for all time-related data. The generation times of the basic displacement and rotation data, as well as the wing surface deformation data, are mapped onto the global time axis. The data generation times are the timestamps carried when these two types of data are generated in their respective solution modules.

[0076] The alignment deviation between the data points of the base displacement and rotation data and the airfoil deformation data on the global time axis after mapping is detected. This alignment deviation reflects whether the two types of data are synchronously available at the same physical moment. The alignment deviation can be calculated by comparing the timestamp intervals of adjacent data points. A specific deviation metric formula is as follows:

[0077]

[0078] Where: Δ represents the alignment deviation calculated within the k-th sampling interval of the global time axis. This represents the precise time corresponding to the k-th data point extracted from the basic displacement and rotation data. This represents the precise time corresponding to data points within the same time window, extracted from the wing surface deformation data.

[0079] When the alignment deviation exceeds the allowable tolerance range, a data interpolation procedure is triggered. In this procedure, new data points are generated to compensate for the alignment deviation based on the respective data change trends of the basic displacement and rotation data and the airfoil deformation data. In practice, the data interpolation procedure can employ linear interpolation or spline interpolation algorithms, estimating missing or out-of-sync data values ​​at the target time based on known, timestamped data points. Optionally, the tolerance range can be set to a fixed value, such as 1 millisecond, according to simulation accuracy requirements. The goal of the data interpolation procedure is to ensure strict temporal alignment between the basic displacement and rotation data and the airfoil deformation data. This continues until data synchronization is complete, synchronizing and aligning the basic displacement and rotation data, the airfoil deformation data, and the physical timestamps of the dynamic simulation scene to synthesize a real-time geometric configuration sequence.

[0080] Example 3: A periodicity detection operation is performed on the real-time geometric configuration sequence to identify the boundary time points where configuration features recur, defining the complete flapping wing cycle. Within the complete flapping wing cycle, multiple time slices are extracted according to a preset sampling density, with each time slice corresponding to a set of configuration data. Data anonymization is performed on each set of configuration data, loading a predefined sensitive feature recognition rule base and scanning the configuration data using this rule base. Geometric details unrelated to the core aerodynamic performance of the butterfly flapping wing aircraft are identified and marked as sensitive geometric features to be removed. A geometric smoothing algorithm is used to process the marked regions, replacing the sensitive geometric features with smooth geometric surfaces to generate anonymized configuration data. The anonymized configuration data is compared with the original configuration data to generate a data modification report, which is then associated and stored in the metadata of the anonymized configuration data. Simultaneously, the geometric complexity confidence score of this set of configuration data is calculated. Only multiple sets of configuration data with geometric complexity confidence exceeding a preset threshold are packaged and transmitted to the flow field initialization unit in chronological order.

[0081] In a specific implementation, an aerodynamic simulation analysis method based on a butterfly flapping-wing aircraft involves performing a periodic detection operation on a real-time geometric configuration sequence. The real-time geometric configuration sequence is time-continuous data synthesized by a spatiotemporal mapping coordinator. The periodic detection operation identifies the boundary time points where configuration features recur. It can be understood that the configuration features can be selected as the spatial position trajectory of the left wingtip or the pitch angle change curve of the butterfly flapping-wing aircraft body. In a specific implementation, the boundary time points are determined by calculating the autocorrelation function or identifying the recurrence of key attitudes. The complete flapping-wing cycle is defined by the boundary time points. The complete flapping-wing cycle is the time interval from the start of a characteristic motion state to its next complete recurrence.

[0082] Within a complete flapping wing cycle, multiple time slices are extracted according to a preset sampling density. Each time slice corresponds to a set of configuration data. The sampling density determines how many instantaneous geometric snapshots are collected within a complete flapping wing cycle. In some embodiments, the sampling density can be set to collect one time slice every one-hundredth of a flapping wing cycle. The configuration data includes the three-dimensional coordinate information of all surface mesh nodes of the butterfly flapping wing aircraft under that time slice. Data desensitization is performed on each set of configuration data, and a predefined sensitive feature recognition rule base is loaded. The sensitive feature recognition rule base contains a series of rules for identifying non-aerodynamic key geometric features, such as identifying micron-level surface roughness, biomimetic scale texture, or the geometry of small structural connectors unrelated to aerodynamic analysis. The configuration data is scanned using the sensitive feature recognition rule base.

[0083] Geometric details unrelated to the core aerodynamic performance of the butterfly flapping-wing aircraft are identified in the configuration data. These details are marked as sensitive geometric features to be removed. These unrelated features may include minor protrusions or grooves used for decoration or structural reinforcement but not significantly affecting the macroscopic flow field. In practice, the identification process is performed by comparing the local curvature of the configuration data with thresholds in a rule base. A geometric smoothing algorithm is then used to process the marked regions, replacing the sensitive geometric features with smoothed geometric surfaces to generate desensitized configuration data. The geometric smoothing algorithm can be a grid-based Laplacian smoothing or a surface fitting algorithm. In practice, the smoothing process aims to eliminate microscopic geometric undulations while maintaining the continuity of the airfoil's macroscopic curvature.

[0084] The anonymized configuration data is compared with the original configuration data to generate a data modification report. This report records the location of the smoothed regions, the original geometric feature description, and the parameters of the smoothed geometric features. In some embodiments, the difference comparison is achieved by calculating the Euclidean distance between the corresponding grid node coordinates, and a summary function is applied to assess the overall degree of geometric change.

[0085]

[0086] Where: Ψ represents the average geometric deviation, N represents the total number of grid nodes participating in the comparison, and p i Let p′ represent the coordinate vector of the i-th node in the original configuration data. iThis represents the coordinate vector corresponding to the i-th node in the anonymized configuration data. The data modification report is then associated and stored in the metadata of the anonymized configuration data. Simultaneously, the geometric complexity confidence score of this set of configuration data is calculated. The geometric complexity confidence score is an indicator used to quantify the reliability of the configuration data in retaining the main aerodynamic shape features after anonymization. Optionally, the geometric complexity confidence score can be calculated by analyzing the standard deviation of the curvature distribution of the anonymized surface and its fit with the reference model. Only multiple sets of configuration data with geometric complexity confidence scores exceeding a preset threshold are packaged and transmitted to the flow field initialization unit in chronological order. The preset threshold is a value pre-set based on simulation accuracy requirements.

[0087] See Figure 4 This is a graph analyzing the confidence level of geometric complexity in the aerodynamic simulation of a butterfly flapping-wing aircraft. Within one flapping cycle (0-1 second), the confidence level first decreases (0-0.25 seconds) and then fluctuates upwards, generally fluctuating slightly around the threshold. Only two time slices (approximately 0.25 seconds and 0.8 seconds) have confidence levels close to the threshold, while the remaining slices meet the requirements, indicating that most of the anonymized configuration data at this stage can be used for flow field simulation. This graph is used in the configuration data screening stage of the aerodynamic simulation of the butterfly flapping-wing aircraft. By using the confidence level index, it ensures that the configuration data of the input flow field not only removes non-critical geometric details (anonymization) but also retains the core features affecting aerodynamic performance, which is a key preliminary analysis for improving simulation accuracy and efficiency.

[0088] Example 4: Monitoring the current computational energy state of the flow field evolution process. The current computational energy state is jointly determined by the iteration step size and grid resolution of the flow field evolution process. A computational energy metric is defined, which is obtained by weighting a computation duration factor and a computational accuracy factor. During the flow field evolution process, the specific values ​​of the iteration step size and grid resolution are periodically collected. The collected values ​​of the iteration step size and grid resolution are substituted into the computational energy metric for calculation, and the quantized value of the current computational energy state is output in real time. The quantized value of the current computational energy state is compared with the historical energy state records. Based on the comparison results, the energy change trend of the flow field evolution process in subsequent iterations is predicted, and the energy allocation strategy is adjusted in advance according to the energy change trend. Based on the current computational energy state, an energy allocation strategy is dynamically scheduled. According to the energy allocation strategy, computational resources are allocated proportionally from the available computational resource pool of the butterfly flapping-wing aircraft to the flow field evolution process. Before each iteration update begins, it is verified whether the allocated computing resources meet the estimated resource requirements for this iteration update. If they do, the iteration update operation is executed, and the actual amount of computing resources consumed by this iteration update operation is recorded in the resource consumption log.

[0089] In practical implementation, an aerodynamic simulation analysis method based on a butterfly flapping-wing aircraft involves monitoring the current computational energy state of the flow field evolution process. The flow field evolution process is a computational task involving iterative flow field solution, and the current computational energy state is jointly determined by the iteration step size and mesh resolution of the flow field evolution process. A computational energy metric is defined to quantify the computational resource demand of the flow field evolution process; it is a dimensionless index that integrates the effects of iteration step size and mesh resolution. In practical implementation, the computational energy metric can be formally expressed as:

[0090]

[0091] Where: E represents the quantized value of the calculated energy state, which is a dimensionless number; ω represents the configuration weight coefficient; λ represents the reference time step; ρ represents the current iteration step size; γ represents the current grid resolution; and γ represents the reference grid resolution. The formula uses the logarithm of the ratio of the iteration step size to the reference step size, and the logarithm of the ratio of the grid resolution to the reference resolution, to jointly characterize the relative intensity of the computational load. During the flow field evolution process, the specific values ​​of the iteration step size and grid resolution are periodically collected, and the collection action is triggered by an independent monitoring thread at set time intervals.

[0092] The collected iteration step size and grid resolution values ​​are substituted into the computational energy metric for calculation, and the quantized value of the current computational energy state is output in real time. This quantized value reflects the relative demand for computational resources during the flow field evolution process at a specific moment. The quantized value of the current computational energy state is compared with historical energy state records. Based on the comparison results, the energy change trend of the flow field evolution process in subsequent iterations is predicted. Historical energy state records are stored in a circular buffer, recording the sequence of quantized computational energy state values ​​from the most recent collection cycles. In practice, the prediction can be based on simple extrapolation or a time series analysis model. Energy allocation strategies are adjusted in advance based on energy change trends; for example, when an increase in energy demand is predicted, more memory allocation is requested in advance.

[0093] Based on the current computational energy state, an energy allocation strategy is dynamically scheduled. This strategy is a set of rules defining how computational resources are allocated to the flow field evolution process. According to the energy allocation strategy, computational resources are proportionally allocated to the flow field evolution process from the available computational resource pool of the butterfly flapping-wing aircraft. The computational resource pool represents the total available computational resources of the simulation system. In some embodiments, referring to Table 1, the energy allocation strategy is implemented as a lookup table, mapping the quantified computational energy state value to a specific resource allocation scheme.

[0094] Table 1: A mapping relationship table

[0095] Calculate the range of energy state quantization values ​​(E) CPU core allocation Memory allocation (GB) Process priority E≤0.5 2 8 Low 0.5<E≤1.5 4 16 middle E>1.5 8 32 high

[0096] Before each iteration, the allocated computing resources are verified to meet the estimated resource requirements for the current iteration. These resource estimates are derived by the flow field solver based on the current mesh size and the complexity of the physical model. The flow field solver first collects specific metrics of the current mesh size, including the total number of mesh cells, the mesh density in key flow field regions, and the mesh quality factor. These metrics directly affect memory usage and parallel computing efficiency. Simultaneously, the flow field solver assesses the complexity of the physical model, which involves the type of currently active physical equations, the degree of nonlinearity of the equations, and the iterative requirements of the coupled solution. Based on these parameters, the flow field solver calls its built-in resource prediction module. This module integrates historical resource consumption logs and predefined resource mapping rules, generating estimated values ​​for the number of CPU cores, memory size, and computation time required for the current iteration through weighted calculations. This provides a quantitative basis for resource verification. Verification is a necessary step to prevent computational failures or drastic performance degradation due to insufficient resources. If the conditions are met, the current iteration update operation will be executed, and the actual amount of computing resources consumed in this iteration update operation will be recorded in the resource consumption log. The resource consumption log will be used for subsequent analysis and optimization of energy allocation strategies.

[0097] Example 5: Establish a result filtering process. Within this process, a dedicated data buffer is configured to temporarily store intermediate flow field data flowing from the flow field evolution process. Within the data buffer, timestamp sorting is performed on the intermediate flow field data to ensure it is arranged in iteration order. An independent data credibility assessment subprocess is started. This subprocess performs a physical rationality check on the sorted intermediate flow field data, removing outlier data points that do not conform to physical laws. The queue of intermediate flow field data processed by the data credibility assessment subprocess is then submitted to the convergence discrimination step. Intermediate flow field data from the flow field evolution process is continuously received, and convergence discrimination is performed on the data. When the discrimination result meets preset conditions, the flow field data at the current moment is output as the final aerodynamic simulation result.

[0098] In practice, the result filtering process is a software module that runs independently of the flow field evolution process. Within the result filtering process, a dedicated data buffer is configured. This data buffer is a contiguous storage area allocated in the computing system's memory, used to temporarily store intermediate flow field data flowing from the flow field evolution process. This intermediate flow field data contains numerical solutions of the flow field variables at specific iteration steps. In practice, the data buffer is managed using a first-in, first-out (FIFO) queue data structure, and its capacity must be able to hold at least several hundred time steps of intermediate flow field data to prevent overflow.

[0099] Within the data buffer, a timestamp sorting operation is performed on the intermediate flow field data. This operation rearranges the logical positions of the data in the buffer in ascending order based on the iteration step number or physical time stamp carried by each intermediate flow field data point, ensuring that the intermediate flow field data is arranged according to the iteration order. In some embodiments, the timestamp sorting operation is particularly important when the flow field evolution process outputs data in parallel, causing slight temporal disturbances. A separate data reliability assessment subprocess is initiated, executing in parallel with the sorting operation. Within the data reliability assessment subprocess, the sorted intermediate flow field data undergoes a physical rationality check, which examines the data based on the fundamental conservation laws of fluid mechanics and physical constraints.

[0100] In practice, physical plausibility checks include, but are not limited to, checking whether the pressure field exhibits negative absolute pressure values, whether the velocity field shows abnormally large values ​​exceeding physical limits in local regions, and whether the density field conforms to positive definiteness and monotonicity. These checks aim to identify non-physical numerical solutions arising from numerical computation instability or error accumulation. The data reliability assessment subprocess applies a set of predefined physical rules to screen the data at each grid cell or monitoring point, identifying anomalous data points that do not conform to physical laws. Optionally, physical rules can be expressed using logical judgment statements or a comprehensive physical consistency index function; for example, the following formula can be used to assess the local physical plausibility of the velocity field:

[0101]

[0102] Where: Φ represents the local physical consistency index. Φ represents the local flow velocity, c represents the local speed of sound, p represents the local static pressure, ρ represents the local density, R represents the gas constant, and T represents the local temperature. When the value of Φ exceeds a reasonable range set based on theory or experience... min ,Φ max If a data point is deemed suspicious, it is considered abnormal. Abnormal data points that do not conform to physical laws are removed. This removal process can involve marking the abnormal data point as invalid or replacing it with an interpolated value from its adjacent valid data point. In some embodiments, for removed data points, their location and the type of physical rule violated are recorded, generating a data quality log.

[0103] The intermediate flow field data queue, processed by the data credibility assessment subprocess, is submitted to the convergence discrimination step. In practice, this means that the result screening process passes a cleaned, ordered sequence of intermediate flow field data that has passed basic physical checks to the subsequent convergence criterion calculation module. This step ensures that the data basis used to determine whether the simulation has convergence has physical credibility. Intermediate flow field data from the flow field evolution process is continuously received, and convergence discrimination is performed on the intermediate flow field data. When the discrimination result meets preset conditions, the flow field data at the current moment is output as the final aerodynamic simulation result.

[0104] See Figure 5 This is a convergence analysis chart of the flow field simulation for a butterfly flapping-wing aircraft. The residuals after cleaning (blue curve) represent the residual data after physical rationality verification and outlier removal. The convergence threshold (red dashed line) is the preset convergence criterion (the simulation is considered convergent if the residual value is below this threshold). The convergence point (red pentagram) is marked as the moment the simulation converges when the residual value first falls below the convergence threshold at iteration step 265. The residuals generally show a downward trend, decreasing rapidly from the initial value (approximately 0.98), with the fluctuation range narrowing after 200 steps. At iteration step 265, the residual value reaches below the convergence threshold, and subsequent residuals stabilize in a low-fluctuation range, indicating that the flow field simulation results meet the accuracy requirements. This chart is used in the result screening stage of the flow field simulation. By comparing the residual curve with the convergence threshold, it determines whether the simulation has reached a stable state, which is a key basis for determining the final aerodynamic simulation results and can ensure the reliability and accuracy of the simulation results.

[0105] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0106] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An aerodynamic simulation analysis method based on a butterfly flapping-wing aircraft, characterized in that, Includes the following steps: Construct a biomimetic flight state library corresponding to a butterfly flapping-wing aircraft, and generate configuration instructions for a dynamic simulation scene based on several predefined biomimetic flight states in the biomimetic flight state library. The configuration instructions are parsed to obtain the scene parameters of the dynamic simulation scene in the time series. The scene parameters are then input into a preset geometric model solver to obtain the real-time geometric configuration sequence of the butterfly flapping-wing aircraft in the dynamic simulation scene. The configuration data of a complete flapping cycle is extracted from the real-time geometric configuration sequence and transmitted to the flow field initialization unit. In the flow field initialization unit, an initial flow field distribution is generated based on the configuration data. A flow field evolution process is initiated. In the flow field evolution process, the initial flow field distribution is used as the starting condition to drive the flow field of the butterfly flapping-wing aircraft to be updated iteratively, and the intermediate flow field data generated by each iteration update is recorded synchronously. A result filtering process is established. In the result filtering process, intermediate flow field data from the flow field evolution process are continuously received. The convergence of the intermediate flow field data is judged. When the judgment result meets the preset conditions, the flow field data at the current time is output as the final aerodynamic simulation result.

2. The aerodynamic simulation analysis method for butterfly flapping-wing aircraft according to claim 1, characterized in that, The step of generating configuration instructions for a dynamic simulation scenario based on several predefined biomimetic flight states in the biomimetic flight state library further includes: The system identifies the user's flight intention command for the butterfly flapping-wing aircraft and matches the target bionic flight state in the bionic flight state library based on the flight intention command. Extract all kinematic constraints contained in the target biomimetic flight state, and encode the kinematic constraints into time-driven commands; The time-driven commands are fused with the inherent physical constraints of the butterfly flapping-wing aircraft to generate a flight control sequence that includes conflict resolution strategies. This flight control sequence constitutes the configuration commands for the dynamic simulation scenario.

3. The aerodynamic simulation analysis method for butterfly flapping-wing aircraft according to claim 1, characterized in that, The step of inputting the scene parameters into a preset geometric model solver to obtain the real-time geometric configuration sequence of the butterfly flapping-wing aircraft in the dynamic simulation scene further includes: The scene parameters are decomposed into two independent data streams: spatial transformation parameters and flexible deformation parameters. The spatial transformation parameters are input into the rigid body kinematics calculation module to calculate the basic displacement and rotation data of the butterfly flapping-wing aircraft in the dynamic simulation scenario. The flexible deformation parameters are input into the elasticity calculation module to calculate the wing surface deformation data of the butterfly flapping-wing aircraft. A spatiotemporal mapping coordinator is established, in which the basic displacement and rotation data, the airfoil deformation data, and the physical timestamps of the dynamic simulation scene are synchronized and aligned to synthesize the real-time geometric configuration sequence.

4. The aerodynamic simulation analysis method for butterfly flapping-wing aircraft according to claim 1, characterized in that, The step of extracting configuration data for a complete flapping cycle from the real-time geometric configuration sequence and transmitting the configuration data to the flow field initialization unit further includes: A periodic detection operation is performed on the real-time geometric configuration sequence to identify the boundary time points where configuration features recur, and the complete flapping wing cycle is defined by the boundary time points. Within the complete flapping wing cycle, multiple time slices are extracted according to a preset sampling density, and each time slice corresponds to a set of configuration data; For each set of configuration data, perform data desensitization to remove sensitive geometric feature information, and simultaneously calculate the geometric complexity confidence level of that set of configuration data; Only multiple sets of configuration data with a geometric complexity confidence level exceeding a preset threshold are packaged and transmitted to the flow field initialization unit in chronological order.

5. The aerodynamic simulation analysis method for butterfly flapping-wing aircraft according to claim 1, characterized in that, The step of iteratively updating the flow field driving the butterfly flapping-wing aircraft further includes: The current computational energy state of the flow field evolution process is monitored, and the current computational energy state is determined by the iteration step size and grid resolution of the flow field evolution process. Based on the current computational energy state, an energy allocation strategy is dynamically scheduled. According to the energy allocation strategy, computational resources are allocated proportionally from the available computational resource pool of the butterfly flapping-wing aircraft to the flow field evolution process. Before each iteration update begins, it is verified whether the allocated computing resources meet the estimated resource requirements for this iteration update. If they do, the iteration update operation is executed, and the actual amount of computing resources consumed by this iteration update operation is recorded in the resource consumption log.

6. The aerodynamic simulation analysis method for butterfly flapping-wing aircraft according to claim 1, characterized in that, The step of establishing a result filtering process, in which intermediate flow field data from the flow field evolution process are continuously received, further includes: Configure a dedicated data buffer for the result filtering process to temporarily store intermediate flow field data flowing out from the flow field evolution process; Within the data buffer, a timestamp sorting operation is performed on the intermediate flow field data to ensure that the intermediate flow field data is arranged in iteration order; A separate data credibility assessment subprocess is started. In the data credibility assessment subprocess, the physical rationality of the sorted intermediate flow field data is checked, and abnormal data points that do not conform to physical laws are removed. The intermediate flow field data queue, after being processed by the data credibility assessment subprocess, is submitted to the convergence discrimination step for use.

7. The aerodynamic simulation analysis method for butterfly flapping-wing aircraft according to claim 3, characterized in that, The step of establishing a spatiotemporal mapping coordinator, in which the basic displacement and rotation data, the wing surface deformation data, and the physical timestamps of the dynamic simulation scene are synchronized and aligned, further includes: Establish a global timeline and mark the physical timestamp as a reference on the global timeline; The data generation times corresponding to the basic displacement and rotation data and the airfoil deformation data are respectively mapped onto the global time axis; The alignment deviation between the data points of the basic displacement and rotation data and the data points of the airfoil deformation data on the global time axis after detection and mapping is determined. When the alignment deviation exceeds the allowable tolerance range, a data interpolation procedure is triggered. In the data interpolation procedure, new data points are generated to compensate for the alignment deviation based on the data change trends of the basic displacement and rotation data and the airfoil deformation data, until data synchronization is completed.

8. The aerodynamic simulation analysis method for butterfly flapping-wing aircraft according to claim 4, characterized in that, The step of performing data desensitization operation on each set of configuration data to remove sensitive geometric feature information further includes: Load a predefined sensitive feature recognition rule base, and use the sensitive feature recognition rule base to scan the configuration data; Identify geometric details in the configuration data that are unrelated to the core aerodynamic performance of the butterfly flapping-wing aircraft, and mark these geometric details as sensitive geometric features to be removed; A geometric smoothing algorithm is used to process the marked region, replacing the sensitive geometric feature information with a smooth geometric surface to generate desensitized configuration data; The anonymized configuration data is compared with the original configuration data to generate a data modification report, which is then associated with and stored in the metadata of the anonymized configuration data.

9. The aerodynamic simulation analysis method for butterfly flapping-wing aircraft according to claim 2, characterized in that, The step of extracting all kinematic constraints contained in the target biomimetic flight state and encoding the kinematic constraints into time-driven commands further includes: The kinematic constraints are classified into three types: trajectory constraints, attitude constraints, and phase constraints. A trajectory planner is assigned to the trajectory constraint, in which the continuous flight trajectory is discretized into a series of waypoints; An attitude solver is assigned to the attitude constraint, and the attitude solver calculates the spatial orientation of the butterfly flapping-wing aircraft at the corresponding position based on the path points. A phase synchronizer is assigned to the phase constraint, in which the timing relationship of motion between multiple flapping wing components of the butterfly flapping wing aircraft is coordinated; The path points, spatial orientation, and motion timing relationships are encapsulated according to the timeline of the dynamic simulation scene to form the time-driven instructions.

10. The aerodynamic simulation analysis method for butterfly flapping-wing aircraft according to claim 5, characterized in that, The step of monitoring the current calculated energy state of the flow field evolution process further includes: Define a computational energy metric, which is obtained by weighting a computation time factor and a computational accuracy factor. During the flow field evolution process, the specific values ​​of the iteration step size and the grid resolution are periodically collected; The collected iteration step size and grid resolution values ​​are substituted into the computational energy metric standard for calculation, and the quantized value of the current computational energy state is output in real time. The quantized value of the current calculated energy state is compared with the historical energy state record. Based on the comparison result, the energy change trend of the flow field evolution process in subsequent iterations is predicted, and the energy allocation strategy is adjusted in advance according to the energy change trend.

Images