Aerodynamic layout optimization of reentry vehicle based on multi-objective genetic algorithm

The Aerodynamics Journal is based on multi-objective genetic algorithm for re-entry aircraft aerodynamic layout optimization. Insult, for = 1.t Wei 2. Ma Qiang Zhang Lumin 21 Northwestern Polytechnical University, Xi'an 710072; 2 China Aerodynamics Research and Development Center. The Mianyang 621000 main business wants to use the multi-objective genetic algorithm to determine the aerodynamic layout optimization of the reentry vehicle, and the optimal solution set is compared with the traditional multi-objective optimization method weighting method constraint method. By calculation, the multi-objective genetic algorithm can search for the approximation of the optimization problem in the secondary operation. The optimal solution set provides a sufficient basis for the aircraft designer to make the target compromise decision.

Key 1 Division reentry vehicle; aerodynamic layout; multi-objective optimization; multi-objective genetic algorithm 0 Introduction In general, the lift-to-drag ratio and drag coefficient of hypersonic reentry vehicles have an important influence on their reentry maneuver characteristics, which is a reentry vehicle. An important indicator to be achieved by aerodynamic layout optimization. However, in order to meet the requirements of the lift-to-drag ratio, it is necessary to take into account other performance indicators such as the weight of the utility volume structure, etc., and it is necessary to perform a compromise process at the time of design. This forms the multi-standard optimization design of the reentry vehicle dynamic layout. The traditional method of multi-objective optimization is to convert it into a single-objective optimization problem, which is obtained by solving the series single-objective optimization problem, 30 optimal solution. A subset of the set. Yes, the traditional multi-objective optimization method is a point-to-point search process. In contrast, genetic algorithms deal with a variety of non-inferior solutions that can generate a large number of non-inferior solutions in a secondary run, so that an approximate optimal solution set for multi-objective optimization can be searched. In this paper, the multi-objective genetic algorithm is used to determine the aerodynamic layout optimization of the reentry vehicle, and the 30 optimal solution set is compared with the traditional multi-objective optimization method weighted method constraint method. Through calculation, the multi-objective genetic algorithm can search for the approximation of the optimization question and the 3 optimal solution set in the secondary operation, which provides a sufficient basis for the aircraft designer to make the target compromise decision.

2 Multi-objective optimization method like multi-objective optimization can be described as a fund project National Natural Science Foundation funded project.

Postdoctoral rover work.

Usually, the optimal solution for multi-objective optimization is a solution set and there is no =2,6 points where all targets are superior to other solutions in the solution set. In other words, if there is no such feasible solution It is called the non-inferior solution or the optimal solution. It is called a non-inferior solution in the target space. All non-inferior solutions constitute multi-objective optimization, and the optimal solution is non-inferior solution set. In theory, any solution in the optimal solution set may become the optimal solution, depending on the preferences of the decision maker. As far as possible, the approximation is obtained, and the subset of the 0 optimal solution set is suitable for the basis of the total compromise solution. Therefore, the first step in solving the multi-private optimization is to determine that Parelo4ifjcfSill's traditional approach to multi-scale optimization 1 is to convert it into a single. For the optimization problem, a non-inferior solution set of 1 is obtained by solving the series and the optimization problem. The most common methods are the weighted sum method and the constrained method for the D1 equation. The weighted sum method can be used to find the non-inferior solution set of the optimization question by changing the cargo weight coefficient milk. The weighted sum of the shortcomings of the method seems to be difficult to determine reasonably.

By changing the constraint value of the objective function, a non-inferior solution set of the question can be obtained. Both the explicit weighted sum method and the constrained method are single-objective optimization methods.

Different from the above algorithms, the genetic algorithm can generate a large number of non-inferior solutions in the secondary operation through appropriate selection mechanism or niche technology, so it is possible to search for multi-objective optimization problems, and the optimal solution set. Objective optimization asks a genetic algorithm based on the selection mechanism of branch leagues. The algorithm process is as follows: (1) Encoding uses grading Gray code to encode individuals; branch league selection First, select two individuals for each target without replay from the population. And retain the best individual to get 2 individuals; secondly, according to the target 2 selection strategy according to no playback, get 2 individuals; merge the above individuals to obtain the replicated population; The search collection of the non-inferior solution of the non-inferior solution of the non-inferior solution of the non-playback random pairing is started from the initialization. The searched, non-inferior solutions are saved into a chain; each time a feasible non-inferior solution is identified from the generated progeny individuals and compared with the feasible solutions stored in the chain. Furthermore, a new feasible non-inferior solution chain optimal preservation strategy is formed to retain the two optimal individuals of the parent.

The selection mechanism adopted by the genetic algorithm makes the parent population of the selected population have at most two chances of being copied. 3 The reentry vehicle aerodynamic layout optimization uses the above multi-standard optimization method to solve the aerodynamic layout such as simplified hypersonic violation. Optimization question, which. Aerodynamics is calculated according to the generalized internal volt-roll flow theory. This method is fast and effective. It can well meet the needs of engineering design and use this dynamic estimate to make up for the lack of genetic algorithm.

01 Aerodynamic layout optimization This example is the astronaut transporter in the European Space Agency's embryonic manned space transfer system. The following simplified optimization model is proposed to balance the bottom diameter and total volume. The lift resistance scale and utility volume 7 are the largest. design variable. = call. The mathematical model can be a product and 5 is the area.

At the bottom of st1, the diameter of the slender body reentry vehicle has a higher lift-to-drag ratio, while the bluff body reentry vehicle has the largest utility volume. Therefore, the two design goals of the above optimization are conflicting. When using the weighted sum method, the weight coefficient is += and 戌=0.250.40.60.8; when using the constraint method, the lift-to-drag ratio is converted into a constraint, and this is taken as 0.7, 0.8, 0.902, and the target is The optimization is solved by the mixed-mixing simulated annealing genetic method. In the solution of the genetic algorithm based on the selection mechanism of the branch league 070, each design variable adopts 10-digit hexadecimal, population size = 005 evolution algebra, and 7 o. 9o hall 1 is inferior. The results obtained by the method are in the case of 2, 0 in 2, and the results of the symbols at the left and right ends are respectively equal to the lift-to-resistance ratio of the human and the effect of 1 volume and 7 volumes. From 2, the method results are quite (1) 15. The multi-objective genetic algorithm gives the optimal solution set for the aircraft designer in the second run, and the optimal design in the target space. 1 rich interest by two! And hunger change; not the same, for a given weight coefficient, the weighted sum method gives a feasible non-inferior solution, but does not fully extend to the entire target space.

From the search for the number of feasible non-inferior designs, the search efficiency of multi-objective genetic algorithm is much higher than that of relying on single-objective optimization. 1994204ChinaAcademicJoumalElectronic 3.2 elbow maneuvering warhead aerodynamic layout optimization 1 for wingless blunt biconical warheads, can make Its head is bent to achieve high maneuverability with an asymmetrical shape. The optimal model of the elbow maneuvering warhead of 3 is that for a given flight condition, under the condition that the bottom diameter is fixed, the optimal gas mathematical model of high lift resistance and small structural quality is sought. The results of the 3 elbow maneuvering warhead model 1 using the constraint method and the genetic algorithm are four. When solving by the constraint method, the structural quality 妒 is transformed into a constraint, and =0.0.50.7,0. Single-objective optimization is solved using an improved simulated annealing algorithm. When using genetic algorithm to solve, each design variable uses hexadecimal, population size = 200, evolutionary algebra 2300, mutation probability =., 5, a total of 121 factories are searched for non-inferior designs.

In 4, the results of the symbols at the left and right ends are the results of single-objective optimization of the lift-to-drag ratio and the structural quality, respectively. As can be seen from 4, the genetic algorithm basically gives an approximate non-inferior solution set for your entire space. And 1 and the single-objective optimization and the results obtained by the constraint method are basically consistent, but the genetic algorithm does not find the point from the large lift-to-drag ratio.

4 Conclusions, the traditional multi-standard optimization method. The multi-standard genetic algorithm can obtain the effectiveness of the optimization of the aerodynamic layout optimization of the aircraft.

1 Operational Research and Optimization Theory Volume. Beijing Tsinghua University Press, 1 inferior 8.

2 weeks Ming. Sun Shudong. Genetic algorithm principles and application examples. Beijing National Defense Industry Press, 1999.

Tang Wei, Zhang Lumin. The reentry vehicle optimizes the aerodynamic layout. Journal of Astronautics. 19943.

PP fiber woven tape T 500