Science
ACTUAL PROBLEMS OF AVIATION AND AEROSPACE SYSTEMS
Kazan Daytona Beach


Aviation and Astronautics-2011

10-th International Conference

(November 8-10, 2011, Moscow, Russia)

S.S.Krylov, A.A.Puntus

Moscow Aviation Institute (National Research University)

4, Volokolamskoye shosse, Moskow, 125993, Russia

On November, 8-10-th, 2011 at the Moscow Aviation Institute (National Research University) 10-th International conference , organized by МАI and Central scientifically-research Institute of mechanical engineering, has taken place. This conference was devoted to 100 Anniversary from the date of a birth of the legendary Soviet designer of space-rocket complexes, graduate МАI of academician Michael Kuzmich Jangel and is directed on revealing of problems of current problems in aerospace branch, and also search of their decisions. Participants of conference have got acquainted with the basic stages of a course of life and achievements of the academician of M.K.Jangel - the Chief Designer of space-rocket systems whom, having arrived in МАI in 1931, has connected the life with aircraft, and later with rocket production. Being the student, it in 1935 arrives for work in Nikolay Nikolaevich Polikarpov design office which became the head of its thesis. Looking through M.K.Jangel degree project under the name , N.N.Polikarpov has seen a lot of interesting and new in design decisions and a technique of calculations. He has become interested in M.K.Jangel destiny. Having heard the story of the student about a long and hard way from deaf Siberian village in aircraft, Polikarpov has told: M.V.Lomonosov make the way in a science. It is laudable! It is laudable! Only don't stop halfway, the diploma of the engineer - there is more to come. It is necessary to go further>. M.K.Jangel having ended МАI with distinction, more than 10 years worked in aviation design offices and on aircraft factories. Having taken part in creation of space-rocket techniques, he in 1954 has headed one of leaders of design offices of the country. In 1960 for the big contribution to development of a science to it the scientific degree of a Dr.Sci.Tech. is awarded, in 1963 it is selected the academician of AN UkrainSSR, and in 1966 - the full member of Academy of Sciences of the USSR. In 1959 M.K.Jangel becomes the Hero of Socialist Work, and in 1961 for development of new space-rocket techniques it is awarded by the second Gold Star of the Hero of Socialist Work. M.K.Jangel - the winner Lenin (1960) and State awards (1967), is awarded by a gold medal of S.P.Korolev (1970), awarded by Lenin's 4 awards, an award of October Revolution, medals. The president of Academy of Sciences of the USSR M.V.Keldysh, speaking about value of works of M.K.Jangel, has told: .

On October, 25-th, 1972 on a building of the Moscow Aviation Institute the memorial board has been established: - M.K.Jangel.

The international conference is the large-scale Event, which participants of a steel more than 500 persons, among which scientists and experts - representatives more than 60 enterprises of aerospace branch, teachers, research assistants, post-graduate students and students of 15 scientific and higher educational institutions of Russia, near and far abroad. Within the limits of conference 8 sections worked, has been heard more than 250 reports.

This survey, prepared by Members of conference Organizing Committee, is presenting the materials on some papers of students and post-graduate students of faculty of МАI. The topics of these researches are connected with mathematical problems, related to engineering applications, including aerospace industry. Short statements, approaches, solutions and results are discussed.

 

A.A.Krasnukhin, A.A.Puntus.

(Moscow Aviation Institute, National Research University, Moscow, Russia).

Forecasting problem of satellite flight on high elliptical orbit.

This survey covers the problem of forecasting satellite flight on high elliptical orbit. This orbits' specialty is in apogee altitude exceeding much perigee one. Satellites, using high elliptical orbits have rather high speed at perigee, but it reduces dramatically at apogee. When spacecraft is near apogee for ground observer it may seem that the satellite do not move for several hours. That means his orbit becomes quasi-geostationary. Quasi-geostationary can be located at any point on the globe. This quality is used in southern and northern latitudes, distant from the equator.

Special mathematic motion model, based on corresponding geo-potential and standard atmosphere models, heeding moon and sun gravity and light pressure, is applied to forecast satellite flight. Motion at high elliptical orbit is described in mathematical model by a system of differential equations rectangular geocentric inertial coordinate system. Speed vector and vector of coordinates are initial conditions for integrating the system of differential equations. Adams method is applied as a basic method of numerical integration of system of differential equations with the known initial conditions of spacecraft motion. Runge-Kutta method is used as auxiliary one to calculate first eight points for Adams method. At each step values of right sides of differential equations are calculated. Moreover, coordinates and speeds are calculated at the beginning of every circuit. This is performed through interpolation by Adams method of forth order. This methodology makes it possible to obtain the forecast of satellite flight with high accuracy for a long period.

 

A.V.Naumov, S.V.Ivanov.

(Moscow Aviation Institute, National Research University, Moscow, Russia).

The problem of distribution of the investment capital allocated for reconstruction of the aerospace complex.

Nowadays the reconstruction of the companies and structures of the aerospace branch is the actual task. The aerospace complex is a foundation which provides the launch of rockets, space groups control, maintenance of the infrastructure which is necessary for development and effective use of products and systems of the aerospace branch.

We present a mathematical model which is offered for investments of capital which are assigned for reorganization of the structure of the aerospace complex. This model is based on two-stage bi-level stochastic programming problem with quantile criteria.

A lot of indefinite and random factors fluent on the quality of the aerospace complex. In the offered model demand for products of the aerospace complex is random factor. We use quantile criteria for adequate risk discount. The quantile criteria is the level of loss in realization of the project  which cannot be exceeded with given level of probability.

The process of making a decision in this model has two stages. On the first stage we determine the volume of investment capital allocated for development of different branches of the aerospace complex. On the second stage we define the optimal price for realized production. This price depends on the realization of random demand.

Participation of the competitor on the market models by bi-level optimization programming problem. The model supposes participation on the market of two players: the leader (the state, the owner of aerospace complex) and the follower (foreign competitors who produce the similar products). The follower chooses his own price for products depending on the price which the leader can set.

We have made an analysis of such model. We have got analytical formulas of optimal price for the production of the leader and the follower depending on the first stage strategy and demand realization. We have got evident specie of the criteria function of the second stage problem. This circumstance allowed us in scalar case (in case of one branch) to build the deterministic equivalent to the original problem. We show the decision of the problem in scalar case. Possible methods for solving the problem in common case are discussed.

A.I.Emelkin, A.A.Puntus, M.N.Mjakshina, V.I.Polezhaev.

(Moscow Aviation Institute, National Research University, Moscow, Russia).

Parametrical research of heat exchange and current structure in an inclined layer at face heating.

The convection and heat exchange in an inclined layer in the presence of a longitudinal gradient of temperature (a case of heating of a layer from an end face) are studied much less, than in case of an inclined layer at a cross-section supply of heat. However, this case also is interesting to the theory of a thermal convection and has applied value as can serve as convection model at cultivation of crystals in special horizontal capacity, at heating from below vertical slices in mines, at cooling of various chinks, at research of corresponding currents in atmosphere, ocean, an asthenosphere and in other cases. In this work the heating from an end face horizontal is considered and having warmed up from below vertical layers, and also the intermediate positions reached at turn of a layer. Calculations are executed for the numbers of the Rayleigh limited to value 10*5 at different and relations of the parties (L/H=1, 5, 10, 30) for a finding of typical modes and current and heat exchange characteristics. As starting at increase in length of a layer, and also for tests the case of square area was considered also at different angles of slope. Even today here the interesting features caused by interactions of convective mechanisms of the various natures are found out. The basic attention is given revealing of parameters at which extreme values of heat exchange and physical interpretation of features of structure of a current and heat exchange which lead to extreme effects are realized.

Calculations are executed on the basis of Network Computer Laboratory of Institute of problems of mechanics of the Russian Academy of Sciences. Comparison to the results received in computer laboratory COMGA (Cоnvection in Micro Gravity and Applications) is given.

 

A.I.Kibzun, A.I.Chernobrovov.

(Moscow Aviation Institute, National Research University, Moscow, Russia).

Minimization of the CVaR by a stochastic quasi-gradient algorithm to obtain upper bound of quantile.

Generally, there are plenty of random factors in the tasks from the fields of technology and economics related to the optimization of resource allocation. In the optimization problems with random factors quantile criterion (Value-at-Risk) is frequently used. For example, in aircraft control terminal quantile is used as a criterion (i.e. deviation from the target (the miss), guaranteed with some probability).

Practical solution of the problem of finding the quantile's minimum is complicated by the fact that the explicit form of the quantile is often impossible to find. A stochastic quasi-gradient algorithm for solving this problem is also known. That algorithm does not use the explicit form of quantile; but it requires the convexity of quantile and fulfillment of some hardly verifiable conditions.

Conditional Value-at-Risk (CVaR) describes the average loss in the worst cases (when the objective function is more than quantile). Finding the explicit form of CVaR is often not easy. But that form has some useful properties, particularly it is easy to check its convexity. For example, if the objective function is convex by strategy then CVaR is also convex by strategy.

We consider the method of constructing an upper bound of quantile based on the solution of the problem of minimizing the CVaR. Stochastic quasi-gradient algorithm is applied to solve this problem. The convergence of that algorithm is proved with unitary probability.

This method for approximation of quantile has the following useful properties:

-            applicability to a wide class of problems;

-            does not require the convexity of quantile;

-            if the level of reliability is close to 1, the upper bound is close to an exact value.

The paper provides examples of applying the algorithm as well as some practical aspects of building the numerical solution methods.

 

V.G.Komov, O.N.Tretiyakova.

(Moscow Aviation Institute, National Research University, Moscow, Russia).

Development of the controlling program for kinematic system of the machine of a high-speed laser-beam cutting.

Software development for management system by multi-coordinate machine of a high-speed laser-beam cutting was the purpose of the given operation.

Any machine of a high-speed laser-beam cutting represents the industrial technological complex providing handling of various materials on physical properties fulfills high-precision cutting preparations which are used in various industries, including by the space. Machine basic elements are the coordinate desktop on which preparation, the laser, the controller, the computer is allocated. Simple kinematic systems where the coordinate desktop has one or two levels of freedom, don't allow to reach high speeds are sharp. At magnification of speed are sharp with usage of simple systems strongly quality of performance of preparations worsens. For magnification of speed are sharp it is necessary to increase capacity of engines, and it conducts to growth of mass of setting and power inputs.

For the decision of this problem the projected plant considered in given operation which includes system with parallel kinematics is offered. In it concerning the main large coordinate desktop, the easy additional coordinate little table which can make relocation with the big accelerations is moved to planes of its parallel surface. It provides necessary accuracy of leading of preparation to a demanded point since allows to accelerate and to brake an additional little table quickly. Besides, such kinematic system of the machine of a high-speed laser-beam cutting provides lowering of expenses of the electric power, i.e. does setting economic while in service at its implementation in mass production for cutting preparations.

In operation the algorithm of the synchronous control has been offered by similar machines in which basis modelling of the relative driving of the main and additional coordinate desktops of setting is necessary. The controlling program implementing given algorithm is created. The program is written in C#, the controller works with usage G-codes, transmission of controlling commands from the computer to the controller is provided. The program is used at designing of the machine with parallel kinematics for a laser-beam cutting in Open Company NPTS ТМ>.

 

V.N.Panovskiy.

(Moscow Aviation Institute, National Research University, Moscow, Russia).

Realization of methods of global optimization using interval analysis.

While designing aircraft structures it is often necessary to optimize certain parameters. They are usually the weight, flight range, aerodynamic characteristics, etc. Optimized parameters are specified by the designer from some restrictions that appear depending on the physical formulation of the problem or resource limitations. The solution of our problem is the object that most closely matches the required specification - this shows the importance of developing an effective optimization method. The aim of this study was to develop an efficient algorithm for global optimization using interval analysis.

During the study of the possibility of using interval methods the main problems hindering the development of algorithms were identified and subsequently solved. The first problem was that the interval analysis works not with real or complex numbers, it works with completely different structures - real intervals. From that comes the necessity of redefining all operations of elementary algebra and figuring out how to change the basic operations in order to get the right result. The second problem is to establish a inclusion function - a notion that is associated with real function. Making wrong inclusion function causes algorithm handling a large number of values ​​that did not actually reached by the initial function.

The solution of two mentioned above problems has allowed to create two global optimization algorithms:

-            dichotomy of the direct image;

-            imaginary values cutoff.

The result of the work is the development of algorithmic support and software for new methods of global optimization. The developed methods have been tested on standard tasks of conditional global optimization of functions with simple and complex structure of isolines. In all cases, the answer was found with a given accuracy and in an acceptable time, which confirms the possibility of using the apparatus of interval analysis. Thus, the obtained results allow us to hope for the successful application of the methods for the design of aircraft structures. In the future we plan to improve the developed methods by creating new algorithms for constructing the functions of inclusion, and develop new ones, using both sets of interval analysis and other mathematical sciences.

 

G.U.Shevchenko, O.N.Tretiyakova.

(Moscow Aviation Institute, National Research University, Moscow, Russia).

Development of elements of software for process control of laser cutting.

This work is a continuation of work on the creation of programs for automation of technological processes and systems to create innovative laser technologies. In [1] was described software for technology systems for the enterprise .

In this paper we presented the software for controlling the laser processing equipment with computer numerical control (CNC), which was developed for SPC Ltd. . CNC is a computerized control system that reads the instructions of a specialized programming language, used in our case, G-codes, and controls the drive machine tools and machine tool accessories.

The software is intended for general machine control laser cutting, to load into memory controller technology programs, to run these programs, for monitoring the treatment and diagnosis for the general state of the control system.

Program interface is a standard application window of the operating system MS Windows. For control used standard input-output (keyboard and mouse), it is also possible to connect additional devices, such as machine control, encoders, touch panels, etc.

This software is designed and used on the machines for laser cutting, Ltd. established in SPC .

References

1. O.N.Tretiyakova, G.U.Shevchenko. Development of control program to install laser controlled thermo-splitting semiconductor materials. In Sat: 9 International Conference Moscow, 16-18 November 2010, M., Publishing House MAI-PRINT, 2010, 318-319.

 

D.V.Metlitskaya.

(Moscow Aviation Institute, National Research University, Moscow, Russia).

Application of the modified method of artificial immune systems to an optimal control problem of discrete dynamic systems.

The method of artificial immune systems (AIS) concerns met heuristic methods of search and is the representative of evolutionary methods. AIS uses ideas borrowed from immunology, simulating work of immune system of a live organism. AIS finds the decision which can differ a little from optimum, but thus is comprehensible from the practical point of view. Unlike classical methods of optimization, AIS can be applied, when almost completely there is no information on character and properties of investigated function.

In the given work application of modified AIS of search of a conditional global extremum to an optimal control problem of the discrete determined systems is considered. The detailed algorithm of the decision of a task in view is offered. On its basis the complex of the software is generated, which allows to solve a task in view for various models of control systems and various functionals. The complex has the convenient interface with which help it is possible to change easily problem and method parameters, and also to analyze overall performance of a method.

Some test examples are solved by means of the developed complex. On examples it is shown that the offered method allows to select parameters in such a way that the received approximate solution coincide with the optimum with enough accuracy.

AIS is applied to a wide range of problems of designing of difficult aerospace systems, including optimal control problems of trajectories of aircrafts of various types.

 

E.A.Aleshina.

(Moscow Aviation Institute, National Research University, Moscow, Russia).

Bacterial Foraging Optimization for optimal control of discrete deterministic systems.

Bacterial Foraging Optimization (BFO) proposed by Y.Liu and K.M.Passino is a swarm intelligence met heuristic; search of the minimum is done by set of agents, then the collected information is aggregated and it gives pivots for search on the next iteration.

The main principles of BFO are the following:

-        the search is done by a colony of bacteria (particles);

-        every particle tends to reach areas with better values of criterion and avoid noxious areas where the criterion value is worse;

-        search direction is randomly chosen;

-        particles interact, attracting and repelling to each other;

-        some bacteria with the worst criterion values restart search from a random point in the decision space or from the location of the currently best solution.

To apply BFO to optimal control problem of discrete deterministic systems a control vector was optimized; a control criterion was chosen as a function to minimize. To find criterion value firstly the state vector was recalculated and then based on its value and control vector the desired was evaluated.

Based on BFO algorithm, a program for solving the mentioned problem was created. Graphical user interface of the utility is convenient for putting in problem characteristics and setting up the parameters of algorithm. The program allows user to make series of experiments with the same algorithm parameters in one click. This is especially important since the nature of BFO is stochastic.

Solving of sample problems shows that it's necessary to examine different modifications of BFO to have the results acceptable for practice use. This examination should take into account the experience of other met heuristics, and especially of Particle Swarm Optimization as it's among the most effective ones.

 

I.E.Evdokimov, V.S.Nikolaenko, G.S.Filippov, B.Yu.Jashenko.

(Moscow Aviation Institute, National Research University, Moscow, Russia).

Indicatrix mathematical modelling of the difficult radiator.

Objective of this research is working out the method and on its basis the program of calculation of difficult radiators in which the radiant stream from diffuse radiation sources is formed by system mirror, diffuse and diffuse-mirror reflectors with the set optical characteristics.

The developed method is received on the basis of an iterative-zonal method and the programs used in computer graphics for definition of light exposure of objects. It allows to count radiation of warm surfaces taking into account various reflections and to define distribution of radiant streams in space taking into account easing in the environment.

The design procedure is realized in the form of the program in language Borland DELPHI. It's algorithm was checked by means of computing MathCAD environment. The developed technique yields good results at calculation of distribution of a stream of radiant energy from geometrically difficult figures which surfaces are set in the form of a set of points with coordinates (X, Y, Z).

Reliability of the developed technique is confirmed by numerical and modelling experiment with use of light model of a radiator.

In numerical experiment good coincidence of results of calculations by the developed technique with the data of analytical indicator calculations, mentioned in reference books is received.

In modelling experiment measurements of indicator of the ax symmetric light model consisting from diffuse of the radiating ring and system of formation of a radiant stream, executed in the form of two coaxial mirror reflectors () were spent. Good coincidence of results of calculation to the data of modelling experiment is received also.

Result of the executed research is the package of the applied programs, allowing to spend mathematical modelling of difficult radiators.

 

N.F.Kudimov, O.N.Tretiyakova.

(Moscow Aviation Institute, National Research University, Moscow, Russia).

Turbulence Models study for analysis of thermodynamic and gas-dynamic processes in

solar-battery cells.

As well as in the previous works [1, 2] we consider a solar panel is installed on an object moving at subsonic speeds on the ground surface. The real engineering application is the car equipped with the solar panels used as primary power supply replacing the internal combustion engine.

In [1] has been shown that an inefficient heat abstraction has occurred in air flow separation zones. To improve the cooling there are offered to turbulize somehow an air flow in order to increase the effective thermal conductivity, and therefore the task of turbulent flow's analysis using various turbulence models becomes an important.

For the purposes of analysis of heat transfer in solar cells there has been resolved additional dual problem consisting of the calculation of radiative-conductive heat transfer in solar cells and calculation of gas-dynamic flow near the surface of solar cell by using various turbulence models. (Spalart-Allmaras, k-ε, SST, RSM, Cebeci-Smith, Baldwin-Lomax).

The analysis of numerical experiments allows authors to drawing up the conclusions about the feasibility of a particular model.

References

1.      N.Vinnichenko, I.Uvarov, I.Znamenskaya. Analysis of convective flow on solar cells cooling. ISFV14 - 14th  International Symposium on Flow Visualization, June 21-24, 2010, EXCO Daegu, Korea.

2.      N.F.Kudimov, O.N.Tretiyakova. Mathematical model of the radiative-conductive heat transfer in solar cells in convective heat transfer conditions. In Sat: 9 International Conference , Moscow, 16-18 November, 2010, M., Publishing House MAI-PRINT, 2010, 319.

 

O.V.Ryazantseva, A.V.Panteleyev.

(Moscow Aviation Institute, National Research University, Moscow, Russia).

Application of adaptive greedy method of random search global extremum to optimal control of discrete systems.

A characteristic feature of the solution of topical problems of the design of complex aerospace systems is the widespread use of global optimization met heuristic methods, allow one to obtain an approximation to the extreme in a reasonable (from a practical point of view) time. In the given work discuss the application of the adaptive greedy method of random search global extremum to the problems of finding the optimal control of discrete deterministic systems. The method uses the idea of multi-start that is repeated to find a solution where each iteration consists of two phases: construction phase and local search. The first phase (phase of construction) combines the and the accident. As a result, it generated solutions of , from which begins the second phase - the local search. Then the obtained points are taken as initial for the first phase and the process continues. The best of these solutions is taken as an approximate solution of the problem. To solve the discrete optimal control problem to find the optimal control and, therefore, the optimal trajectory. For this we consider a block column of the coordinate values of the control vector at each step. To calculate the criterion corresponding to this column will be searched, the trajectory of the system of equations of state with known values of controls contained in the column, and given initial condition. It then computes the value of the criterion. Thus, the column was similar to the block in the set of feasible solutions, and the criterion corresponding to that - the objective function whose minimum is sought by the greedy method, adaptive random search global optimization.

To implement this method, a program was developed. It implements the following features: input parameters of the problem, input parameters of the method, the solution of the problem; tabular and graphical presentation of the solutions, and view the protocol solutions and help information. It was decided a few test cases with this program help. The results indicate the effectiveness of the adaptive greedy method of random search global extremum in optimal control of discrete systems.

 

O.N.Tretiyakova.

(Moscow Aviation Institute, National Research University, Moscow, Russia).

Experience in the development and application of Internet technologies for learning physics students MAI.

Developed by us and used for several years at the Department of Physics, MAI system of distance learning physics students of technical university can make to the learning process the following elements:

-        carrying out remote labs in physics,

-        carrying out course work in physics and disciplines of specialization,

-        an online test at the rate of Physics and disciplines of specialization,

-        organization of independent work of students,

-        an electronic library of textbooks on physics for students of technical universities.

When you create all these developments used the concept of level differentiation in teaching students of the university. All system components are presented in an educational website Department of Physics 801 Moscow Aviation Institute at:   http://www.kaf801.ru/

 

P.A.Panteleev, A.G.Badalova.

(Moscow Aviation Institute, National Research University, Moscow, Russia).

Software package .

Developed a package of programs in MathCAD allows you to define the best strategy for risk management in the enterprise aviation industry, specializing in the production of new equipment, as well as to select the most preferred method of risk management, depending on the level of risk. We consider the following strategies: self-containment, a harmonized risk management strategy, risk transfer, and implement their methods: avoidance, transfer, distribution, funds reservation, association, localization, diversification, limiting the compensation, warning.

To solve the problem generated by the authors applies the algorithm of the analytic hierarchy process. Compiled by the hierarchical model, which includes four levels of hierarchy: I - the risk of the enterprise; II - risk levels (low, below average, average, above average, high); III - risk management techniques; IV - Risk Management Strategy. Based on the method of paired comparison, the corresponding matrices are composed of judgments. The method consists in finding the largest eigenvalue of the matrix of judgments and the corresponding eigenvector (usually normalized). Eigenvector reflects the ordering of priorities, and the eigenvalue is a measure of the consistency of judgments. The final vector of influence strategies at risk (priority strategies) is obtained by finding the product of the matrix effect on the fourth level of the third, the third level in the second, the second level to the first.

Application of the described software program showed that, in general, for the enterprise strategy is the most preferred strategy harmonized risk management strategy further risk transfer and risk-taking strategy.

If the risk level is low, better to adopt a strategy of risk-taking, more harmonized strategy for management and risk transfer strategies. If the risk level below the average, it is better to apply the harmonized management strategy, a strategy more risk taking and risk transfer. If the risk level of average or above average, it is preferable to a harmonized strategy management, strategy further risks transfer and adoption strategy. If the risk level is high, it is better to adopt a strategy of risk transfer, then the strategy of a harmonized management and risk-taking strategy.

The results obtained can be used in the feedback loop strategic risk management systems.

 

S.D.Makhijanova, A.A.Puntus.

(Moscow Aviation Institute, National Research University, Moscow, Russia).

About limiting cycles in double oscillator, raised by a nonlinear friction.

In the presented research work mathematical modelling double oscillator with a slice-nonlinear dry friction is spent and research of corresponding model is conducted, and also problems of modelling of phase portraits double oscillator n the personal computer are considered. The method of research of considered mathematical model consists in the analytical decision of the independent system of the differential equations received as a result of modelling, modelling of the given system on the computer further is spent, then the corresponding numerical method gets out, the program of calculation with use of computer facilities is made and realization of program calculation on the computer facilities is spent. As a result modelling process on the computer of corresponding phase portraits is spent. In the course of performance of the given scientific work the mathematical model double oscillator with a slice-nonlinear dry friction has been constructed, the analytical decision of the equations in two areas is resulted. Analytical and numerical modelling of the given system on the computer is considered, corresponding phase portraits are received as a result of the given research. Limiting cycles are received by the following technique. On pressing of button of the program the algorithm of the decision on the entered time interval is carried out. After its end final values of variables register in fields of entry conditions, and at following start of algorithm are used as entry conditions. So that to receive a limiting cycle, it is necessary to carry out algorithm while the picture on phase planes won't cease to change.

So, in the given research work the analytical decision received as a result of mathematical modelling double oscillator independent system of the differential equations is passed. Besides, in the given research work the method of research of considered mathematical model double oscillator and as a result applications of this method construction of limiting cycles in such double oscillator, raised by a nonlinear friction is executed is presented. For task in view realization, construction and research of considered limiting cycles is executed, modelling of the given system on the personal computer is spent, the choice of a corresponding numerical method is realized, the necessary program of calculation is made, calculation on the computer facilities is realized and phase portraits double oscillator corresponding to various entry conditions and in various considered areas of a phase plane are received.

 

S.S.Leonov, A.A.Puntus

(Moscow Aviation Institute, National Research University, Moscow, Russia);

A.N.Vasilyev.

(St. Petersburg State Polytechnical University, National Research University, St. Petersburg, Russia).

The methods of neuron network modelling at the solution of typical problems of hydrodynamics.

The purpose of the present work is efficiency checking of the methods of neuron network modelling at the solution of two typical problems of hydrodynamics.

In the first problem (the Couette flow) there considered flow of an incompressible viscous fluid in horizontal layer of the given thickness. The lower bound of the layer is motionless, and upper moves with the constant horizontal velocity directed along the axis of abscissa. The axis of ordinates is directed vertically upwards. The stationary solution is searched. We suppose that the flow is flat. Moreover, the flow is horizontal. Also there is no horizontal pressure gradient.

In the second problem (the Poiseuille flow) there considered flow of an incompressible viscous fluid in the layer of the given thickness and the given length with solid boundaries under the influence of the pressure difference enclosed to the edges. As well as in the previous problem we are searching the stationary solution. The flow is flat and horizontal.

Process of solution of these problems implies the realization of the following algorithm:

-        assignment of domain in which the solution is being searched (definition of sub-domain with the ratios changing, definition of boundary sections with the changing of the admissible condition type);

-        assignment of operators appearing in the equation, initial-boundary conditions, other relations and constraints;

-        selection of the neuron decomposition to find the solution (the monolayer perceptrons are used at the solution);

-        assignment of the kind of generation of test points with possibility of the assignment of different laws of allocation of these points depending on geometry of the area;

-        assignment of the error functional as a characteristic of the model quality;

-        minimization of the error functional in enrolment of the test points.

The present algorithm was realized in MathCAD 14 system. The result of work is:

-        receiving the approximate solution for Couette flow and Poiseuille flow;

-        receiving inaccuracy of the approximate solutions considered problems via meaning of the error functional;

-        visual comparison of both exact and approximate solutions.

 

Yu.O.Khan, A.I.Fedyushkin, A.A.Puntus.

(Moscow Aviation Institute, National Research University, Moscow, Russia).

Mathematical modelling of a natural convection at the normal and lowered gravitation.

The purpose of the given work was construction of mathematical model for calculation of a thermal, concentration and thermo-concentration convection in dimensionless variables at various numbers of the Rayleigh. The problem of a natural convection of a viscous incompressible liquid in the square closed area with heat-isolated horizontal walls and with the set temperatures on boundary vertical walls is considered. The mathematical model of the given problem represents the two-dimensional equations of Navier-Stokes in dimensionless variable Boussinesq in approach. For speed on all borders the sticking condition is laid down, and for temperature such conditions are laid down: the bottom and top horizontal borders - heat-isolated, and the right and left vertical borders - isothermal. As entry conditions the motionless liquid with linear temperature distribution is set.

For modelling of a considered convective problem dimensionless record of the equations which includes dimensionless parameters - numbers of Prandtl, the Rayleigh, by Grashof, Schmidt was used. The system of the equations of Navier-Stokes in approach of Boussinesq was used. Calculations have been spent with use of modern hydrodynamic program Fluent. Comparison of the received results with decision for different numbers of the Rayleigh has shown good accuracy of calculations. Results of calculation also have shown that at set boundary and entry conditions at increase in number of the Rayleigh the maximum value of the module of speed is displaced to area borders, and isotherms aspire to accept horizontal position.

The concentration model of a convection is defined by thermal number of the Rayleigh and Schmidt's number. The field of concentration under the influence of a convection starts to change with smaller numbers of the Rayleigh, than a temperature field as Schmidt's number is more than number of Prandtl. The number of Prandtl isn't enough, that is conductivity of heat much more, than viscosity, therefore a field of temperature under the influence of a liquid starts to change only at enough great number of the Rayleigh (more than 100). And when the temperature field invariably, a concentration field already considerably changes. The conclusion from here follows that concentration heterogeneity under normal conditions is less, than in the conditions of the lowered gravitation.

As a result of the given work dependence of a thermal and concentration convection, influence of number of the Rayleigh is received and influence of changes of numbers of Prandtl and Schmidt on liquid movement in the closed volume is investigated.

 

Sergey Sergeevich Krylov, the dean of faculty of the Moscow Aviation Institute (National Research University). Area of scientific researches:

information systems and technologies, information systems and technologies, program engineering, problems of preparation of qualified professionals.

Artur Agafonovich Puntus, Professor of department of differential equations of the Moscow Aviation Institute (National Research University). Area of scientific researches:

the qualitative theory and applied problems of differential equations, problems of teaching of courses of higher mathematics, problems of preparation of qualified professionals.



[Contents]

homeKazanUniversitywhat's newsearchlevel upfeedback

© 1995-2008 Kazan State University