Моделирование аэродинамических характеристик профиля крыла с использованием искусственных нейронных сетей

The present paper describes a simple guidance strategy for the reentry mission of a generic Reusable Launch Vehicle. The guidance computation starts with a nominal feasible trajectory, which gets updated during actual reentry based on onboard measurements, using both incremental angle of attack and bank angle parameters. An analytical formulation for a modified predictor-corrector algorithm is developed, using angle of attack profile and bank angle reversal strategy. The guidance solution is arrived at under the heating rate and equilibrium glide constraints and the control actuation is implemented through the aerodynamic control surfaces. A scheme to improve the overall performance of the predictor-corrector algorithm is evolved, which is based on the determination of optimum perturbation levels of control variables. In addition, the control vector formulation is provided with a weighting option, for improving the overall terminal interface accuracy. Numerical results are obtained for the time histories of control variables and performance parameters, for a typical nominal trajectory, which show that proposed guidance scheme works well for a specified reentry mission. Results are also obtained for the optimum perturbation levels for the control parameters, which are used for estimating the sensitivity coefficients required in corrector algorithm. Lastly, the results are obtained for the trajectory parameters, by considering dispersions in the aerodynamic drag and lift coefficients, for establishing its robustness. The results clearly bring out the fact that the proposed reentry guidance algorithm is able to achieve the required terminal conditions, even in the presence of large perturbations in the aerodynamic coefficients.

This paper presents the practical application of cm artificial neural network (ANN) to the problem of concrete mix in a factory. After a brief introduction to the complex problem of concrete mixes design and a quick review of the fundamental basis of neurocomputation, an optimal neural network model has been developed to cope with such a problem. For training the net, several control mixes have been fabricated recording in all cases both the characteristic 28 days compressive strength and the workability measured in terms of the slump of the Abrams' cone. After the training process of the net, the power of its predictive ability is checked by comparison of the results obtained with those corresponding to four reference mixes; in this way, it is shown that the considered approach can be used in multicriterial search for optimal concrete mixes.

where C~ and C~ are the stationary aerodynamic lift and drag coefficients; the dash stands for the derivative with respect to the angle of attack a. The conclusions of [1] were verified on the experimental data published in [2]. Large vibrations (with amplitudes greater than a meter) of phase wires, mainly in the vertical direction along the O Y axis (Fig. 1), were observed repeatedly in high-voltage power lines. In the special literature such motions are called a dance or galloping of a single wire of a power line. Condition (1) was obtained by Den-Hartog [3] in studies of one-dimensional motion of a mechanical model of a wire as a necessary condition of self-excited vibrations in the vertical plane. Thus, the consideration of one-dimensional self-excited vibrations of an airfoil led to condition (1), which may rightfully be called the Glauert-Den-Hartog condition. It is interesting to note that , in our opinion, Russian researchers are not familiar with the paper of Glauert [1]. Thus, for example, there are no references to [1] in [4], although stability questions are considered, and other papers of Glauert are cited. Fedyaevskii and Blyumina [5] studied the conditions of loss of aerodynamic damping by an angle airfoil rotating about a fixed axis. They obtained a rather complex inequality containing moment aerodynamic coefficients and came to the conclusion (from experimental data) that condition (1), which they call DenHartog's condition, can also be used as a criterion. Therefore, in our opinion, it would be useful for researchers to know that the book-stacks of the Russian State Public Library of Scientific and Technical Literature contain rather complete sets of the Proceedings of the British Advisory Committee for Aeronautics for 1909-1980. Condition (1) was confirmed experimentally [5, 6] and has been applied in civil engineering. In designing skyscrapers, which are subjected to wind loads, the cross section of the building must be oriented so that the angles of attack with respect to the direction of prevailing winds are beyond the interval in which inequality (1) holds [71. In simulating the behavior of a power-line wire under the action of wind, Van'ko [8] considered motion of an airfoil with three degrees of freedom: motion along the O X and O Y axes and rotation about the center of mass (Fig. 1). The conditions of existence of static solutions of the system of equations of motion were established: if the airfoil is lifting and the lift Y ~ 0, equilibrium positions exist for any wind (air flow) velocity V > 0. In studies of the Lyapunov stability of possible equilibrium positions, the following results were obtained:

The application of artificial neural networks to the estimation of bioprocess variables will be discussed. In fermentation processes, direct on-line measurements of primary process variables usually are unavailable. The state of the cultivation, therefore, has to be inferred from measurements of secondary variables and any previous knowledge of process dynamics. This research investigates the learning, recall and generalization characteristics of neural networks trained to model the nonlinear behavior of a fermentation process. Two different neural network methodologies are discussed, namely, feed-forward and recurrent neural networks, which differ in their treatment of time dependence. The neural networks are trained by backpropagation using a conjugate gradient technique, which provides a dramatic improvement in the convergence speed. The objective is to use environmental and physiological information available from on-line sensors to estimate concentrations of species in a bioreactor. Results of the neural network estimators are presented, based on experimental data available from the ethanol production by Zymomonas mobilis fermentation. The feed-forward and recurrent neural network methodologies are demonstrated to perform suitably as unmeasurable state estimators. Both networks offer comparable abilities of recall, but recurrent networks perform better than feed-forward networks in generalization.

This paper addresses the aerodynamics ofmembrane flappingwings. In a series ofwind-tunnel experiments on the 25 and 74-cm-wingspan models, the stroke-averaged lift and horizontal force were measured at an angle of attack that varied from 0 deg (horizontal) to 90 deg (hovering position). The flapping frequency was held constant, and freestream velocity was varied with the advance ratio range 0–1.2. For high angles of attack, flapping wings do not exhibit a typical abrupt stall seen with fixed wings. This feature in flapping wings allows a smooth transition from a level flight to hovering and back. The angle of attack, corresponding to a balance of forces in the horizontal direction, decreases exponentially with advance ratio. Aerodynamic coefficients are defined using a reference velocity as a vector sum of a freestream velocity and a stroke-averaged wing-tip flapping velocity. The lift and drag coefficients obtained using this procedure collapse well for the studied advance ratios, especially at lower angles of attack. Polynomial approximations for aerodynamic coefficients are proposed that can be used in flight-performance analysis.

To understand the working principle of ornithopters, various surveys were made on the natural flyers with flapping wings and their ability to produce lift and thrust. The crank mechanism is chosen to make the micro air vehicle (MAV) for spy work without being identified. This mechanism is one of the most complex ones, as its flow condition changes along with its wing motion. The flapping wing mechanism is analysed at two different speeds: 3 and 6 m/s. To analyse the aerodynamic characteristics, the lift and drag coefficient is measured at a different angle of attack using ANSYS software. The results are compared at various times with different working conditions to get the most suitable and reliable conditions. The designed flapping wings exhibit more efficiency with speed 6 m/s and the angle of attack at 15°.

This paper introduces an application of artificial neural network use in Chinese coal mines and states its efficiency in mining problem solving. This paper provides a proposal for further development and application of the neural network in the coal mines and points out the technical problems which should receive attention.

Three-wheelers are popular in Sri Lankan roads as a convenient mode of transportation. Presently there are about 500,000 registered three-wheelers. The range of commercially available electric vehicles are limited due to their battery weight and hence used mostly as neighbourhood vehicles. Three-wheelers are used mostly as neighbourhood vehicles and hence ideal for electrification. But no successful commercial electrically powered three-wheeler is available yet. Aerodynamic drag coefficient and rolling resistance coefficient are two basic parameters needed to determine the motor capacity of an electric vehicle. Accurate values of aerodynamic drag coefficient and rolling resistant coefficient are important to calculate accurate power estimate in electric vehicle design. In this paper we present an accurate method to determine the aerodynamic drag coefficient and rolling resistance coefficient of commonly available three-wheelers in Sri Lanka.

The prediction of aerodynamic coefficients on airfoils using machine learning is increasingly popular due to its efficiency in time and cost. Research typically focuses on a single image type without comparing various types and output quantities (single or multi-output). Although convolutional neural networks (CNN) are widely used, their final layer is often suboptimal as a linear operator, and feature extraction results contain many parameters that can still be trained. Support vector regression (SVR) with kernel functions effectively reduces common errors in feature vectors. We propose a hybrid method, AeroCNNSVR, combining CNN as a feature extractor and SVR as a regressor to predict aerodynamic coefficients on airfoils. This study focuses on the shape and position of airfoils according to the angle of attack (AoA) without considering flow conditions. Using 14533 aerodynamic coefficients from 563 airfoil types, we created a dataset of grayscale and RGB airfoil images. Results show the proposed method with grayscale images performs better because combining SVR strengthens the predictive model, while grayscale images accurately represent the airfoil's shape and position. AeroCNNSVR achieves lower RMSE values for Cl (0.101522), Cd (0.016450), and Cm (0.129661) compared to the CNN model’s Cl (0.112493), Cd (0.019060), and Cm (0.130041). Additionally, AeroCNNSVR's R² values for Cl (0.976071), Cd (0.928700), and Cm (0.860574) surpass those of the CNN model (Cl 0.970620, Cd 0.904282, Cm 0.816355). This research contributes by 1) proposing an alternative besides CFD for predicting and identifying trends in aerodynamic coefficients of airfoils in a much shorter time during the design stage; 2) offering wind tunnel practitioners for early detection of configuration errors; 3) providing an overview of the aerodynamic characteristics of the airfoil under test, including the angle at which stall conditions occur.

The paper deals with the applications of artificial neural networks in the control of the DC drive. In the paper three control structures are discussed. The first control structure uses a conventional PI controller. The second structure uses a neural network predictive control. The last structure is a sensorless control of the DC drive using feedforward neural network. The DC drives were simulated in program Matlab with Simulink toolbox. The main goal was to find the simplest neural network structures with minimum number of neurons, but simultaneously good control characteristics are required. Despite used neural networks, which are very simple, it was achieved satisfactory results.KeywordsArtificial Neural NetworkArtificial Neural Network ModelModel Predictive ControlDirect Torque ControlInduction Motor DriveThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

A new semiempirical dynamic stall model is presented to predict the aerodynamic coefficients of an airfoil in unsteady conditions. The model is mainly based on two modifications on the airfoil angle of attack which introduce an equivalent angle of attack to implement in the proposed aerodynamic coefficients. The first modification includes unsteady wake effects of the airfoil in the proposed model which are not considered in most of the semiempirical models. Hence, an effective angle of attack is introduced based on an unsteady aerodynamic theory using the Wagner function for pitching and plunging oscillations of the airfoil. The second modification takes into account time delay effects and dynamic effects due to the flow separation on the airfoil. These effects are included in the model by developing a semiempirical relationship between static and dynamic angle of attack published in the literature. Finally, the modifications form the equivalent angle of attack to implement in a set of equations based on Kirchhoff flow theory. The proposed approach dynamically approximates trailing edge separation point on the airfoil by a completely new way and contributes its effects in the aerodynamic coefficients. Moreover, apparent mass effects and the airfoil leading edge vortex contributions are sufficiently included in the model. The presented method effectively predicts unsteady lift and pitching moment coefficients of the airfoil undergoing stall conditions. Consequently, the proposed model is verified against experimental data under various test cases where obtained numerical results are in good agreement with the test data. Furthermore, the performance of the proposed model is compared with various dynamic stall models including Leishman–Beddoes model, ONERA model and Boeing–Vertol model. The comparison shows that the proposed model minimizes the number of the semiempirical parameters determined from an experiment in dynamic stall modeling compared with Leishman–Beddoes model and ONERA model. It also demonstrates that the presented model performs as well as other well-known models using the various approaches.

Application of artificial neural networks to assessment of ship manoeuvrability qualities This paper presents an attempt to applying neural networks for assessment of parameters of standard manoeuvrability tests, i.e. circulation test and zig-zag test. Methodological approach to application of neural networks as well as applied network structures and neuron activation functions are generally presented. Also, results of simulations performed by means of the elaborated networks are given in comparison with test cases selected at random. In order to analyze and reveal general trends, correlation relationships between results from network simulations and test cases were calculated and are presented as well.

Aircraft post-flight reduction is one of the most important research topics, since post-reduction is effective in all parameters of bird performance, such as maximum speed, fast acceleration, shortening the runway required, high maneuverability, and reduced fuel consumption. Therefore, it is important to examine how the force applied to the airfoil changes in different streams from an industrial point of view. In this study, NACA( National Advisory Committee for Aeronautics) Airfoil 0012 has been studied. Numerical solution is used for the study. In the simulations, the air fluid and Mach number 0.5 are considered and the lift, drag and pressure coefficients at different attack angles are investigated. The results show that this airfoil has a lift factor of 0 due to its symmetric geometry at angle of attack 0 but its drag coefficient is about 0.01384 and its pressure coefficient is about -0.483. By varying the angle of attack from -12 to +12, it was observed that for the positive attack angle the lift coefficient increases and for the negative attack angle increases the lift coefficient decreases. Also with increasing angle of attack the drag coefficient always increases and the coefficient of pressure coefficient decreases. Cite this Article Ali Kamranpay, Alireza Mehrabadi. Numerical Analysis of NACA Airfoil 0012 at Different Attack Angles and Obtaining its Aerodynamic Coefficients. Journal of Mechatronics and Automation. 2019; 6(3): 8–16p.

This paper proposes the application of Artificial Neural Networks (ANN) in forecasting miners' occupational disease incidence, examines the effect of BP network topology and parameters on training time, describes a neural network that uses Legendre polynomials as nonlinear weights and tries to find the most suitable network model to solve the problem via analyzing the forecast result. Through computer simulation, it clarifies that this model could forecast miners' pneumoconiosis incidence commendably.

An experimental application of Artificial Neural Networks to Eating Disorders is presented. The sample, composed of 172 cases (all women) collected at the Centre for the Diagnosis and Treatment of Eating Disorders of the 1st Medical Division of the St. Eugenio Hospital of Rome, was subdivided, on the basis of the diagnosis made by the specialist of the St. Eugenio, into four classes: Anorexia Nervosa (AN), Nervous Bulimia (NB), Binge Eating Disorders (BED) and Psychogenic Eating Disorders that are Not Otherwise Specified (PED-NOS). The data base was composed of 124 different variables: generic information, alimentary behavior, eventual treatment and hospitalization, substance use, menstrual cycles, weight and height, hematochemical and instrumental examinations, psychodiagnostic tests, etc. The goal of this experiment was to verify the accuracy of the Neural Networks in recognising anorexic and bulimic patients. This article describes 6 experiments, using a Feed Forward Neural Network, each one using different variables. Starting from only the generic variables (life styles, family environment, etc.) and hematoclinical and instrumental examinations, a Neural Networks provided 86.94% of the prediction precision. This work is meant to be a first contribution to creating diagnostic procedures for Eating Disorders, that would be simple and easy-to-use by professionals who are neither psychologists nor psychiatrists nor psychotherapists but who are, however, among the first to meet these patients and who are therefore called upon to give such patients the very first pieces of advice on seeking proper treatment.