Optimization method of rolling wing blade connecting rod mechanism

Abstract

The invention discloses an optimization method of a rolling wing blade connecting rod mechanism, and belongs to the field of rolling wing aircrafts, the method comprises the following steps: representing a blade swing angle through structural parameters of the blade connecting rod mechanism, and defining the structural parameters of the blade connecting rod mechanism including the length of each connecting rod, coordinates of a key hinge point and configuration offset as variable vectors; for various key working conditions such as hovering, forward flight and transition, a multi-objective function is constructed, and the multi-objective function not only quantifies the performance of the connecting rod mechanism, such as kinematics fitting errors, dynamic stability indexes and force transmission efficiency indexes of a target motion law, but also covers aerodynamic efficiency indexes and aerodynamic acoustic indexes directly related to the performance of the whole machine. The optimal structural parameters of the blade connecting rod mechanism are obtained by solving the multi-objective function under the constraint set, and global optimization of the structural parameters of the rolling wing blade connecting rod mechanism under the multi-working-condition, multi-objective and multi-constraint conditions can be achieved.

Description

Technical Field

[0001] This invention belongs to the field of rolling wing aircraft, and more specifically, relates to an optimization method for a rolling wing blade linkage mechanism. Background Technology

[0002] Roller wing aircraft (also known as cycloidal gyroplanes or cycloidal propeller aircraft) generate desired thrust magnitude and direction by arranging several blades along the circumference of a rotor and performing periodic pitching. They possess hovering, forward flight, and rapid attitude adjustment capabilities, making them suitable for applications such as urban air mobility, special operations, and short-range transportation. Unlike traditional propellers or helicopter rotors, the variation of blade pitch angle with rotation angle significantly affects the thrust vector, propulsion efficiency, and vibration and noise levels of the roller wing. The geometry and kinematic characteristics of this mechanism determine the blade pitch angle pattern within one cycle, thus influencing the thrust direction and magnitude, and are closely coupled with the overall aerodynamic efficiency, vibration, and aeroacoustic levels of the aircraft.

[0003] In engineering implementation, the geometric-kinematic-dynamic characteristics of multi-link mechanisms exhibit multi-scale and strongly coupled features with the overall aerodynamic / acoustic performance. For example, indicators such as the fitting accuracy of the blade target pitch trajectory, the peak and variability of angular velocity and angular acceleration, and the lower limit of the transmission angle constrain the mobility and reliability of the mechanism, and also affect key performance parameters such as power requirement, hovering quality factor, and overall sound pressure level under different operating conditions (hovering, forward flight, and transition). These indicators are used to characterize and constrain the kinematic feasibility and structural reliability of the mechanism, and further affect the comprehensive performance such as power requirement, hovering quality factor, and overall sound pressure level under hovering, forward flight, and transition conditions. There are often trade-offs among these indicators, requiring systematic integration in the early stages of design.

[0004] Existing design methods primarily rely on geometric approximations, graphical methods, or empirical trial-and-error to determine mechanism parameters. The process typically involves first determining the mechanism parameters, followed by performance evaluation. When facing high-dimensional, nonlinear, and strongly constrained design spaces, this process often requires multiple rounds of iterative iteration. Meanwhile, common design methods tend to use a single objective (such as pitch angle fitting) as the primary criterion, neglecting dynamic stability (angular velocity / angle plus peak value and fluctuation), force transmission efficiency (transmission angle), and aerodynamic / acoustic indicators. This results in limited feasible solution domains and representative compromise solutions under interdisciplinary constraints. Summary of the Invention

[0005] In view of the above-mentioned defects or improvement needs of the existing technology, the present invention provides an optimization method for the rolling blade linkage mechanism, thereby solving the problems of existing optimization methods that are difficult to achieve cross-disciplinary index coordination, low optimization iteration efficiency, insufficient handling of engineering constraints, and disconnect between design and verification processes.

[0006] To achieve the above objectives, according to a first aspect of the present invention, an optimization method for a rolling rotor blade linkage mechanism is provided. The rolling rotor blade linkage mechanism includes a driving link L1, an intermediate link L2, a driven link L3, and an eccentric link L4 connected sequentially to form a closed loop. The two endpoints of L1 are P1 and P2, the two endpoints of L2 are P2 and P3, the two endpoints of L3 are P3 and P4, and the two endpoints of L4 are P4 and P1. The method includes: With the objectives of minimizing the kinematic fitting error, average transmission angle, and aeroacoustic noise of the roller blade linkage mechanism, and maximizing the dynamic stability, average transmission angle, and aerodynamic efficiency of the roller blade linkage mechanism, an objective function is established and solved under preset constraints to obtain the optimal structural parameters of the roller blade linkage mechanism. The structural parameters of the roller blade linkage mechanism include the coordinates of P1 to P4. and the angle between L4 and the negative half of the y-axis ; .

[0007] According to a second aspect of the present invention, an electronic device is provided, comprising: a computer-readable storage medium and a processor; The computer-readable storage medium is used to store executable instructions; The processor is configured to read executable instructions stored in the computer-readable storage medium and execute the method as described in the first aspect.

[0008] According to a third aspect of the invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to perform the method as described in the first aspect.

[0009] According to a fourth aspect of the invention, a computer program product is provided, comprising a computer program or instructions that, when executed by a processor, implement the method described in the first aspect.

[0010] In summary, compared with the prior art, the above-described technical solutions conceived by this invention can achieve the following beneficial effects: The method provided by this invention establishes a unified parametric model of the blade linkage mechanism. The blade swing angle is characterized by the structural parameters of the blade linkage mechanism, including the length of each link, coordinates of key hinge points, and configuration offset. The structural parameters of the blade connecting rod mechanism, including those included, are defined as a variable vector. This invention constructs a multi-objective function for various key operating conditions such as hovering, forward flight, and transition. This multi-objective function not only quantifies the performance of the linkage mechanism itself, such as the kinematic fitting error with the target motion law, dynamic stability index, and force transmission efficiency index, but also covers aerodynamic efficiency indexes (such as hovering quality factor) and aeroacoustic indexes (such as overall sound pressure level) directly related to the overall aircraft performance. By solving this multi-objective function under the constraint set C, the optimal structural parameters of the blade linkage mechanism are obtained. The method provided by this invention couples the geometric / kinematic / dynamic design of the blade linkage mechanism layer with the aerodynamic and aeroacoustic performance evaluation of the entire rolling wing aircraft layer through a unified framework, realizing global optimization under multiple operating conditions, multiple objectives, and multiple constraints.

[0011] As a further preferred embodiment, the method provided by this invention constructs and trains a surrogate model for computationally expensive aerodynamic and aeroacoustic indices. This surrogate model is trained based on initial samples generated within a design space through experimental design, using variable vectors. The system takes aerodynamic efficiency and overall sound pressure level as inputs and outputs performance indicators such as aerodynamic efficiency and overall sound pressure level under corresponding operating conditions. This allows for improved optimization iteration efficiency and result reliability through surrogate computation and adaptive verification, even under conditions where high-fidelity aerodynamic / acoustic evaluation is costly. Furthermore, in subsequent optimization processes, an online sampling mechanism is triggered based on model uncertainty, expected improvement (EI), or other sampling criteria. This adaptively selects high-value sample points in the design space for high-fidelity simulation evaluation and updates the surrogate model, thereby significantly reducing the overall computational cost while maintaining accuracy.

[0012] In summary, the method provided by this invention has the following advantages: (1) Significantly improve design efficiency and globality: By introducing a proxy calculation and adaptive sampling verification mechanism for high-cost aero-acoustic indicators, the computational cost of high-cost indicators is reduced by using a proxy model: For high-cost indicators such as efficiency FM and overall sound pressure level OASPL that require high-fidelity simulation, this invention effectively reduces the number of high-fidelity simulations through proxy prediction and adaptive sampling mechanism, taking into account both computational efficiency and result reliability, making it possible to perform global multi-objective optimization of complex design space within an acceptable engineering time, avoiding local optima caused by computational resource limitations in traditional methods.

[0013] (2) Achieve synergistic trade-offs in multidisciplinary performance: Unify the performance indicators and constraints of multiple disciplines such as mechanism, dynamics, aerodynamics and aeroacoustics under the same optimization framework, and intuitively weigh the synergistic relationship between various objectives in the early stage of design.

[0014] (3) Enhance the reliability and engineering practicality of the design: Unlike conventional methods that only focus on the fitting of the target pitch angle curve, key constraints such as transfer angle, dynamic peak value, and geometric interference are directly incorporated into the optimization process to ensure that the output solution is feasible and reliable in terms of structure and dynamics.

[0015] (4) Improve the transfer from design to verification: After obtaining the Pareto solution set, the recommended scheme is output through decision support. The final output of the scheme is a complete scheme package containing geometric parameters, performance data and CAD / CAE interface, which provides clear support for subsequent detailed design, simulation verification and prototype manufacturing. Attached Figure Description

[0016] Figure 1 This is a basic structural diagram of a roll wing aircraft.

[0017] Figure 2 This is a parametric geometric schematic diagram of the roller linkage mechanism provided in an embodiment of the present invention.

[0018] Figure 3 This is a schematic diagram illustrating the process of training and online sampling of the surrogate model provided in an embodiment of the present invention.

[0019] Figure 4 This is a schematic diagram of the export process of the CAD / CAE interface provided in an embodiment of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0021] In the cyclic motion problem of a rolling vane, replacing the actual kinematics of the linkage mechanism with an idealized sine wave or a preset swing angle curve can reduce modeling complexity, but may lead to deviations in performance evaluations such as aerodynamic loads, power, and noise. While existing technologies provide a segmented approach to designing mechanisms (geometry / motion / dynamics) and demand indicators (aerodynamic efficiency, acoustic parameters) within the engineering process, mechanism parameters are often matched with aerodynamic / acoustic results later as check items. The lack of methods for co-optimizing indicators across disciplines in the early design stages makes it difficult for final parameters to remain stable under different operating conditions, and the design process cannot be reliably reproduced.

[0022] It is evident that existing technologies still have significant shortcomings in interdisciplinary collaborative modeling, optimization iteration efficiency, engineering constraint handling, and design verification. Based on this, this invention provides a method for parametric collaborative optimization of a blade linkage mechanism. This method, based on the determination of mechanism parameters, considers kinematic and dynamic constraints, as well as aerodynamic performance and noise indicators. Through a more efficient evaluation and optimization process, it outputs mechanism parameters and pitch laws that can be directly used for engineering implementation.

[0023] To facilitate understanding, before introducing the method provided by this invention, the structure of a rolling wing aircraft with four blades will be described first.

[0024] like Figures 1-2 As shown, the rolling wing aircraft includes first to fourth blades 1 to 4 evenly distributed along the circumference, and the eccentric control point 5 (i.e., P1) has an eccentricity relative to the rotor center point P4. Each blade linkage is a planar four-bar linkage, consisting of a driving link L1, an intermediate link L2, a driven link L3, and an eccentric link L4 connected sequentially to form a closed loop. The two endpoints of L1 are P1 and P2, the two endpoints of L2 are P2 and P3, the two endpoints of L3 are P3 and P4, and the two endpoints of L4 are P4 and P1. The linkage mechanism corresponding to each blade adopts the same four-bar structure and the same structural parameter configuration (including the length of each link and the eccentric angle ε).

[0025] The rotor refers to the rotating component in the roll blade that rotates around the rotor rotation axis 7. Its rotation center is located at point P4. It is used to support and drive each blade to make circular motion with the rotor center point P4. While revolving with the rotor, each blade can realize periodic pitch motion (i.e., rotation) around its own pitch axis.

[0026] corner The rotor motion angle describes the angular change of the linkage mechanism within one rotation cycle; it is the angle of rotation of the blade relative to the positive half-axis of the x-axis (the origin of the xoy coordinate system where the x-axis is located is P4). ∈[0,2π), that is The reference zero position direction is the positive x-axis; For the blade swing angle Independent variable.

[0027] That is, the blades move under the action of the eccentric control point 5 and the rotor rotation shaft 7, and each blade moves in accordance with the linkage mechanism during the rotor rotation. Changing pitch angle This results in a differentiated aerodynamic distribution within a cycle.

[0028] Blade angle This refers to the blade pitch angle, which is the angle of rotation of the rotor relative to the positive x-axis. At that time, the rotation angle of the blade about its pitch axis (the pitch axis refers to the axis around which the blade rotates during its periodic motion, i.e., the blade pitch angle) The pitch angle of the blades (the axis of rotation) directly affects the magnitude and direction of the aerodynamic force generated by the blades, thus influencing the thrust vector, aerodynamic efficiency, and noise level. By rationally designing the blade pitch angle, the aerodynamic efficiency of the roll rotor can be improved while ensuring thrust or lift requirements are met. On the other hand, the rate of change of the blade pitch angle affects unsteady aerodynamic forces and pressure pulsation characteristics, thereby impacting the noise of the roll rotor.

[0029] To describe the blade tilt angle The dynamic change characteristics are introduced by taking into account the first derivative of each component with respect to the rotation position. and The second derivative is used to characterize the angular velocity and angular acceleration characteristics of the blade's sway angle.

[0030] Figure 1 The smaller dashed circle represents the trajectory of the eccentric control point 5 relative to the rotor center point P4 within one rotor cycle. During rotor rotation, the pitch axis position of each blade moves in a circle (i.e., revolves) around the rotor rotation axis 7. Simultaneously, under the constraint of the eccentric control point 5 and the linkage mechanism, each blade rotates periodically around its respective blade pitch axis (i.e., the blade rotates relative to the rotor), thereby achieving the blade pitch angle. Rotation angle with rotor Periodic changes, Figure 1 The lift and drag shown are the instantaneous aerodynamic components of the blade at the corresponding rotor angle position.

[0031] All blades revolve periodically around the rotor's rotation axis 7. P1, P2, P3, and P4 are also called key hinge points in the linkage mechanism; their spatial positions uniquely determine the mechanism's geometry. The leading edge refers to the tip of the blade. Figure 2 In this text, LE represents the leading edge of the propeller blade, and the distance between LE and point P3 is defined as the pitch axis length. The trailing edge refers to the point at the very end of the blade. Figure 2 In this context, TE is used as the denoting element, and the line connecting the leading and trailing edges is called a chord. Phase offset is employed. The angle between the eccentric rod L4 and the negative half of the y-axis is represented by... The reference zero position direction is the negative half-axis of the y-axis. The length of the eccentric rod is... , representing eccentricity, is used to characterize the configurational distance offset of a linkage mechanism. Tangential vector. (The direction of the tangent to the trajectory along the pitch axis) and the chord vector (Along the blade chord direction, the blade chord) is used to describe the periodic change of the blade as the rotation angle changes; the instantaneous angle (taking the acute angle) between the intermediate link L2 and the driven link L3 is defined as the transmission angle of the linkage mechanism. It is used to characterize the effectiveness of force and motion transmission, that is, the force and motion transmission quality of the linkage mechanism under the corresponding configuration.

[0032] Working condition set This set is used to represent various different flight or operational conditions, and contains a total of [number missing]. Three operating conditions are considered. To reflect the relative importance of each operating condition in the overall performance evaluation, a weighting coefficient is introduced. ,in Furthermore, each weight coefficient satisfies the normalization condition. The weighting coefficients are used to weight and summarize the performance indicators calculated under each working condition during the multi-working-condition evaluation process, thereby obtaining a comprehensive performance evaluation result. Each working condition... This corresponds to a defined set of operating condition parameters, including the incoming flow velocity v and the blade rotational speed n. It can be understood that during the flight of a roll wing aircraft, all blades are under the same incoming flow conditions, and their incoming flow velocities can be considered consistent.

[0033] This invention provides an optimization method for a rolling rotor blade linkage mechanism. The rolling rotor blade linkage mechanism is a planar four-bar linkage, comprising a driving link L1, an intermediate link L2, a driven link L3, and an eccentric link L4 connected sequentially to form a closed loop. The two endpoints of L1 are P1 and P2, the two endpoints of L2 are P2 and P3, the two endpoints of L3 are P3 and P4, and the two endpoints of L4 are P4 and P1. The method includes: With the objectives of minimizing the kinematic fitting error, average transmission angle, and aeroacoustic noise of the roller blade linkage mechanism, and maximizing the dynamic stability, average transmission angle, and aerodynamic efficiency of the roller blade linkage mechanism, an objective function is established and solved under preset constraints to obtain the optimal structural parameters of the roller blade linkage mechanism. The structural parameters of the roller blade linkage mechanism include the coordinates of P1 to P4. and the angle between L4 and the negative half of the y-axis ; .

[0034] Specifically, the lengths of the driving rod L1, intermediate connecting rod L2, driven rod L3, and eccentric rod L4 They are respectively:

[0035]

[0036]

[0037]

[0038] To facilitate subsequent angular representation, the relative positions of the members are described using vector form:

[0039]

[0040]

[0041]

[0042] according to Figure 2 The geometric relationship shown indicates that the length of one diagonal of the linkage mechanism (i.e., the length of the line connecting P1 and P3) .

[0043] According to the Law of Cosines, in the triangle formed by P1, P2, and P3... There is:

[0044] According to the law of sines, we can obtain:

[0045] exist By the Law of Cosines, we have:

[0046] in , It is the intermediate angle used to assist in calculating the final blade swing angle. It is possible and Represented as: .

[0047] Then, define the variable vector. Structural parameters used to characterize linkage mechanisms:

[0048] To ensure the engineering feasibility of the obtained structural parameter scheme, the constraints are preset as a set of constraints. ,include: (1) Geometric assembly and mobility constraints of linkage mechanisms The length of the four-bar linkage is positive: and

[0049] The distances between each hinge point satisfy the geometric relationships required for mechanism assembly. To ensure the mechanism can complete assembly and form an effective closed loop, the lengths of the four links must meet geometric conditions to avoid any link being too long, which would prevent the closed loop from being formed. That is:

[0050] To ensure continuous mobility of the mechanism in its assembled state, the length of the four-bar linkage must satisfy the Grashof criterion, namely:

[0051] in , These represent the shortest and longest link among the four links, respectively. , This indicates the connecting rods of the other two intermediate lengths.

[0052] The above constraints are used to ensure that the linkage mechanism can move normally.

[0053] (2) Transmission angle constraint of linkage mechanism:

[0054] in, , This is the minimum transmission angle.

[0055] (3) Dynamic stability constraints of linkage mechanisms Within one motion cycle, the peak values ​​of angular velocity and angular acceleration must not exceed preset upper limits to ensure the smoothness of the mechanism's motion. That is:

[0056]

[0057] in , These are the maximum values ​​of the angular velocity and angular acceleration of the blade swing angle, respectively.

[0058] (4) Capacity constraints of linkage mechanisms Based on the aerodynamic forces of the propeller blades and the inertial forces of the mechanism, calculations are performed under the following operating conditions. The lower linkage mechanism at the corner position Instantaneous driving torque Define the rotor angular velocity as (Unit: rad / s), under operating conditions As the rotor angular position changes with time, the blade pitch angle is determined by the rotor angular position. Therefore, according to the chain rule:

[0059] The blade pitch angle is determined by the rotor angular position (i.e. Therefore, according to the chain rule:

[0060] Similarly,

[0061] Simplified calculation of driving torque:

[0062] in, It is the equivalent moment of inertia of the linkage mechanism about the pitch axis.

[0063] Then the linkage mechanism in working condition The periodic average power is:

[0064] To ensure reliable operation of the linkage mechanism, the following conditions must be met:

[0065]

[0066] in, and These are the rated torque and rated power of the linkage mechanism, respectively.

[0067] Construct a set of objective functions for multi-objective optimization:

[0068] (1) Kinematic fitting error target

[0069] in, This is a reference curve for a given blade sway angle.

[0070] (2) Dynamic stationarity index

[0071]

[0072] in For angular velocity stability, For angular acceleration stability, , The smaller the value, the higher the dynamic stability.

[0073] (3) Average transmission angle index Minimum transfer angle This has already been introduced as a constraint condition for the transmission angle, and the average transmission angle can also be used as an auxiliary optimization objective to improve the overall force transmission quality.

[0074] (4) Multi-condition aerodynamic and aeroacoustic targets Calculate separately in the set of operating conditions Various working conditions aerodynamic efficiency index Aeroacoustic Indicators And according to the weighting coefficients Perform weighted summation:

[0075]

[0076] Among them, the weighting coefficient Used to reflect the relative importance of different operating conditions in comprehensive performance evaluation. And satisfy .

[0077] In actual calculations, the rotor period is... Discretize the data into at least 360 sampling points to avoid insufficient sampling and ensure the accuracy of numerical calculations. .

[0078] Under working conditions Below, the blade motion is determined by the design variable vector. The determined blade swing angle Decision. Based on the blade tilt angle, under operating conditions... The periodic average thrust was obtained through simulation. With periodic average power .

[0079] 1) Aerodynamic efficiency index (FM) Under working conditions Below, the thrust coefficient and power coefficient characterize thrust output capability and power consumption level, respectively, and are defined as follows:

[0080]

[0081] in, and They are respectively determined by the blade tilt angle The generated periodic average thrust and periodic average power; air density; The rotor reference area; Angular velocity; Where is the rotor radius.

[0082] Using hovering quality factor The hovering quality factor measures the aerodynamic efficiency of an aircraft during hovering. It characterizes how close the actual power consumption is to the ideal minimum power consumption under a given thrust condition. This hovering quality factor is a dimensionless index; a higher value indicates higher energy utilization efficiency. The blade pitch angle, by adjusting the angle at different pitch positions, directly changes the thrust generation efficiency and power consumption level. In other words,

[0083] Therefore, under multiple operating conditions, aerodynamic efficiency The calculation formula is:

[0084] 2) Aeroacoustic Specifications (OASPL) The blade pitch angle affects the blade aerodynamics, which in turn directly affects the noise of the roll rotor.

[0085] Under working conditions Below, calculate or measure the A-weighted sound pressure signal at the standard monitoring location. Overall Sound Pressure Level (OASPL) is a comprehensive acoustic index characterizing the intensity of aerodynamic noise. A lower OASPL value indicates a lower level of aerodynamic noise radiation. Its overall sound pressure level is defined as:

[0086] in, For reference sound pressure level; One motion cycle; For the propeller blade angle The determined aerodynamic noise sound pressure response; It is the periodic integral of the square of the sound pressure, used to construct the aeroacoustic evaluation index OASPL, and does not represent the actual acoustic power.

[0087] Under multiple operating conditions, the objective function for aeroacoustic noise is defined as:

[0088] and All of these depend on the unsteady aerodynamic forces and pressure field distribution generated by the blades during their periodic motion.

[0089] Among them, periodic average thrust Periodic average power And the A-weighted sound pressure signal at the standard monitoring location is usually based on a given variable vector. and blade angle The data is calculated or obtained through high-fidelity aerodynamic or aeroacoustic simulation (CFD (Computational Fluid Dynamics) / CAA (Computational Aeroacoustics)) or experimental measurement methods.

[0090] Based on the above evaluation indicators, the collaborative optimization problem of the roll rotor blade linkage mechanism is formulated as the following multi-objective optimization model:

[0091]

[0092]

[0093] in, Corresponding to the number One constraint condition; and To design the upper and lower bounds of the variable vector.

[0094] It is understandable that for performance metrics where a large value is desired, introducing a negative sign into the objective function is equivalent to minimizing the objective function. Here, minimizing... It helps reduce mechanical impact, friction loss, and mechanical stress; The performance index can be maximized by taking the negative sign. This describes the effective aerodynamic angle of attack formed by the blades during their cyclic motion. Maximizing this parameter improves aerodynamic efficiency; minimizing it... It helps improve flight stability and cycle consistency; maximizes It can improve propulsion efficiency; minimize This allows for the achievement of low-noise design goals. By satisfying a preset set of constraints Under the premise of solving the above multi-objective optimization model, a set of non-dominated solutions can be obtained.

[0095] It is understandable that the structural parameters of the linkage mechanism of each blade of a rolling wing aircraft are the same. That is, the linkage mechanism corresponding to each blade of the rolling wing aircraft has the same set of specifications and geometric configuration, and the geometric dimensions are consistent, which facilitates processing and assembly and improves reliability.

[0096] The aforementioned multi-objective optimization model can be solved using any existing method, such as multi-objective evolutionary optimization algorithms with elitist strategies (e.g., NSGA-II, NSGA-III, SPEA2, MOEA / D, etc.), or multi-objective Bayesian optimization methods based on surrogate models (e.g., ParEGO, EHVI / qEHVI, etc.).

[0097] Furthermore, considering It is difficult to obtain directly through analytical methods; it usually requires based on given design variables. and the corresponding blade swing angle (Also written as) This is calculated or obtained through high-fidelity aerodynamic or aeroacoustic simulation (CFD / CAA) or experimental measurement methods. and These are performance metrics with high computational costs. In other words, aerodynamic efficiency and aeroacoustic metrics usually rely on high-cost simulations or experiments. Without proxy computation and adaptive verification mechanisms for high-cost metrics, it is difficult to conduct iterative searches for multiple objectives and operating conditions under limited computational budgets, which in turn limits the completeness and reliability of compromise solution sets.

[0098] Based on this, as a further preferred embodiment, the method provided by the present invention, when solving the objective function, uses a pre-trained surrogate model to predict and... The initial solutions and the corresponding candidate solutions and To further calculate aerodynamic efficiency Aeroacoustic noise .

[0099] Specifically, regarding the multi-condition aerodynamic efficiency index defined above... Aeroacoustic Indicators Since its calculation relies on high-fidelity numerical simulation, frequent calls during the optimization process would lead to excessive computational costs. Therefore, when solving the objective function, an efficient evaluation method based on a surrogate model is introduced.

[0100] Within the variable vector space, an initial sample set of variable vectors is generated using experimental design methods: The number of samples Based on design variable dimensions Select a multiple of it (e.g.) (Take a factor of 5 to 15). For each design point in the initial sample set. , Under various working conditions Next, perform high-fidelity CFD / CAA simulations to obtain the corresponding true values ​​of FM and OASPL, forming the initial dataset:

[0101] The initial dataset is divided into a training set (80% of the initial dataset) and a validation set (20% of the initial dataset). The agent model is trained based on the training set and validated using the validation set.

[0102] The surrogate model can be implemented using one or a combination of the following models: the Kriging model (also known as the Gaussian Process Regression, GPR), the Radial Basis Function (RBF) model, or the Artificial Neural Network (ANN) model. That is, the surrogate model can be implemented using a single surrogate model, or multiple surrogate models can be used to make predictions and then the results can be fused. The result fusion can employ methods such as weighted averaging, error correction, or optimal selection to improve the prediction accuracy and stability of the surrogate model.

[0103] The surrogate model's prediction accuracy and generalization ability are evaluated using K-fold cross-validation (K-fold cross-validation divides the samples into K subsets, alternately using one subset as the validation set and the remaining subsets as the training set for K training and validation iterations, and averaging the evaluation metrics as the model performance evaluation result). This ensures that it has an acceptable prediction error level within the design space. The input to the surrogate model is... The output is for each working condition. Below and .

[0104] like Figure 3 As shown, in the subsequent multi-objective optimization iteration process, the surrogate model is used to quickly evaluate the performance of candidate design points on aerodynamic efficiency and aeroacoustic objectives.

[0105] Since the optimization objective of this invention is a multi-objective, multi-condition problem, the surrogate model not only needs to maintain accuracy around a single performance index, but also needs to maintain reliable predictions in the trade-off region between multiple objectives (i.e., the potential non-dominated solution region). Therefore, during the optimization iteration process, an adaptive sampling strategy based on the uncertainty estimation or expected improvement index (EI) of the surrogate model is introduced.

[0106] Specifically, in the first In each iteration, based on the prediction results and uncertainty distribution of the current surrogate model in the multi-objective evaluation space, a new batch of samples is selected to form the sample addition set. High-fidelity CFD / CAA simulations were performed again on the augmented data points to obtain new real label data, which was then incorporated into the existing dataset. In other words, when the reliability of the prediction results output by the surrogate model is insufficient or has significant potential for improvement, a high-fidelity simulation verification is triggered. The new samples obtained from the verification are added to the sample set and the surrogate model is updated to improve the reliability of the surrogate model.

[0107] The following adopts Taking the algorithm for multi-objective collaborative optimization as an example, the process of solving the objective function is explained.

[0108] In multi-objective optimization, different computational strategies are adopted based on the varying computational costs of the evaluation metrics: Low computational cost metrics: including kinematic fitting error Dynamic stationarity index , and average transmission angle These types of indicators can be directly based on the swing angle. Numerical calculations are performed using their respective formulas, without the need to call up high-fidelity simulations.

[0109] High computational cost metrics: including aerodynamic efficiency Aeroacoustic noise That is, multi-condition aerodynamic efficiency index Aeroacoustic Indicators . and The trained surrogate model is used for rapid prediction, and CFD / CAA verification is triggered for key candidate solutions when necessary to update the surrogate model.

[0110] In other words, in each iteration, computationally inexpensive metrics (such as kinematic fitting error, dynamic stability, and transfer angle) are directly calculated based on relevant formulas. For computationally expensive aerodynamic and aeroacoustic metrics, since each evaluation requires simulation using CFD (Computational Fluid Dynamics) / CAA (Computational Aeroacoustics) techniques, the computational cost is high. Therefore, a trained surrogate model is used to quickly predict these metrics. Based on this, a multi-objective evolutionary update process is performed on the candidate solution set, iterating generation by generation through mechanisms such as non-dominated relation discrimination, solution set distribution maintenance, and elite solution retention until the preset convergence conditions are met. When necessary, high-fidelity verification is performed on key candidate solutions to improve model accuracy.

[0111] The specific solution process includes: 1. Initialization Within the range of design variable values Under the premise of this, an initial population is randomly generated within the design space:

[0112] Population size Determined based on problem size and computing resources.

[0113] 2. Individual evaluation For the first Any candidate solution in the population Please evaluate according to the following steps: Based on the parametric mechanism model in S1, given design variables Under the conditions, calculate and Corresponding blade swing angle (Also recorded as) ), and obtain its first and second derivatives relative to the turning position. (Also recorded as) )and (Also recorded as) ), and the transmission angle (Also recorded as) For computational costs, numerical calculations are performed directly according to the calculation formula. For high computational cost indicators, a proxy model is called for prediction. A batch of "high-value points" (which may include elite solutions on non-dominated frontiers) are selected from the candidate set for CFD / CAA verification, and the verification results are fed back to update the proxy model.

[0114] 3. Non-dominated sorting and crowding calculation The population is hierarchically sorted according to the multi-objective non-dominated relationship to obtain several non-dominated fronts; the crowding distance is calculated within each front to measure the diversity of solutions.

[0115] 4. Selection, Crossover, and Mutation Based on non-dominant hierarchy and crowding distance, a tournament-style selection of parent individuals is adopted, and offspring populations are generated through crossover and mutation operators. and will and After the merger, a new round of non-dominated ranking will be conducted to build the next generation of elite population. .

[0116] 5. Termination Criteria The iteration terminates when the number of iterations reaches a preset upper limit or the change in the overall hypervolume index falls below a threshold. The resulting non-dominated solution set approximates the Pareto front of the roll rotor rod mechanism design problem.

[0117] Finally, decision support and engineering outputs are provided: such as Figure 4As shown, the obtained non-dominated solution set is analyzed, and several candidate schemes with engineering representativeness are selected according to the preset decision requirements (e.g., noise priority, efficiency priority, or comprehensive compromise). For each candidate scheme, the system outputs its mechanism geometric parameter set, key performance indicators under different working conditions, and representative performance curves. It also outputs reproducible pitch angle laws, constraint satisfaction, and interface data for CAD / CAE / simulation, and generates standardized interface data files for CAD / CAE software for subsequent detailed configuration design, high-fidelity simulation verification, and prototype manufacturing.

[0118] For the sake of consistency, the performance evaluation indexes of the roll rotor blade linkage mechanism are defined as the following set of indexes:

[0119] After obtaining the non-dominated solution set, decision analysis and engineering output are performed on the candidate design, specifically including: 1. Performance distribution and sensitivity analysis. For each candidate solution in the non-dominated solution set, statistical indicators are compiled and visualized. The distribution of the data was analyzed, and the interrelationships between different indicators were examined.

[0120] 2. Solution Selection Decision preferences are set based on engineering requirements, for example: low noise priority (prioritizing the selection of components while meeting minimum efficiency and kinematic performance requirements). Minimum candidate solution; efficiency priority (choose the solution that best meets the noise limit and dynamic stability constraints). The selection process may involve choosing the smallest candidate solution or prioritizing a comprehensive compromise (choosing a balance between noise and aerodynamic efficiency based on actual needs). Several representative candidate solutions are selected from the Pareto front.

[0121] 3. Exporting Engineering Data Generate interface data files compatible with CAD / CAE software, such as... Figure 4 As shown, the data includes key node coordinates, link lengths, offset parameters, expected swing angle curves, and multi-condition performance data, which will be used for subsequent research.

[0122] This invention provides an electronic device, including: a computer-readable storage medium and a processor; The computer-readable storage medium is used to store executable instructions; The processor is configured to read executable instructions stored in the computer-readable storage medium and execute the method as described in any of the above embodiments.

[0123] This invention provides a computer-readable storage medium storing computer instructions that cause a processor to perform the method described in any of the above embodiments.

[0124] This invention provides a computer program product, including a computer program or instructions, which, when executed by a processor, implement the method described in any of the above embodiments.

[0125] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An optimization method for a roller rotor blade linkage mechanism, wherein the roller rotor blade linkage mechanism comprises a driving link L1, an intermediate link L2, a driven link L3, and an eccentric link L4 connected sequentially to form a closed loop, wherein the two endpoints of L1 are P1 and P2, the two endpoints of L2 are P2 and P3, the two endpoints of L3 are P3 and P4, and the two endpoints of L4 are P4 and P1, characterized in that, The method includes: With the objectives of minimizing the kinematic fitting error, average transmission angle, and aeroacoustic noise of the roller blade linkage mechanism, and maximizing the dynamic stability, average transmission angle, and aerodynamic efficiency of the roller blade linkage mechanism, an objective function is established and solved under preset constraints to obtain the optimal structural parameters of the roller blade linkage mechanism. The structural parameters of the roller blade linkage mechanism include the coordinates of P1 to P4. and the angle between L4 and the negative half of the y-axis ; .

2. The method as described in claim 1, characterized in that, The kinematic fitting error The calculation formula is: in, For kinematic fitting error, , For the blade swing angle, , , , , These are the lengths of L1 to L4, respectively. Let x be the angle of rotation of the blade relative to the positive x-axis. This is a reference curve for the blade tilt angle; The formula for calculating the dynamic stationarity is: in, For angular velocity stability, For angular acceleration stability, and These are the angular velocity and angular acceleration of the blade's swing angle, respectively. ; The formula for calculating the average transmission angle is: in, For the average transmission angle, , , The formula for calculating the aerodynamic efficiency is: in, For aerodynamic efficiency, For the set of operating conditions, For the first Weighting coefficients for different working conditions , , , air density; The rotor reference area; Angular velocity; The rotor radius; and The first The periodic average thrust and periodic average power under various operating conditions were obtained through simulation. The formula for calculating the aeroacoustic noise is as follows: in, Aeroacoustic noise, , For reference sound pressure level; One motion cycle For the first The aerodynamic noise sound pressure response under various operating conditions was obtained through simulation. The objective function is: .

3. The method as described in claim 2, characterized in that, The preset constraints include: Geometric assembly and mobility constraints: and in, and They represent The minimum and maximum values ​​in and They represent The two values ​​other than the maximum and minimum values; Transmission angle constraint: in, The minimum transmission angle; Dynamic stationarity constraints: in, , These are the maximum values ​​of the angular velocity and angular acceleration of the blade swing angle, respectively. Linkage mechanism capacity constraints: in, and These are the rated torque and rated power of the linkage mechanism, respectively.

4. The method as described in claim 3, characterized in that, When solving the objective function, a pre-trained surrogate model is used to predict and... The initial solutions and the corresponding candidate solutions and To further calculate aerodynamic efficiency Aeroacoustic noise .

5. The method as described in claim 4, characterized in that, The surrogate model is any one or a combination of Gaussian process regression model, radial basis function model, or artificial neural network model.

6. An electronic device, characterized in that, include: Computer-readable storage media and processors; The computer-readable storage medium is used to store executable instructions; The processor is configured to read executable instructions stored in the computer-readable storage medium and execute the method as described in any one of claims 1-5.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing a processor to perform the method as described in any one of claims 1-5.

8. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by a processor, they implement the method as described in any one of claims 1-5.

Images