Research On Bio-inspired Optimization Of Large Opening Cylinder With Nozzle

Due to the process and structural requirements, opening nozzle has become one of the most common structures of special press equipment. For the geometry discontinue connections of the main shell and manifold, there will be a high stress concentration in the large opening cylinder with nozzle intersecting area under the internal pressure, which becomes the primary part of the severe accident. The new national standard GB150-2011 《Pressure Vessel》, added " Analysis method of cylindrical opening with radial nozzle reinforcement design ", the scope of the standard further is groundbreaking expanded.Based on the related pictures of structural parameters and stress concentration factors given by the GB 150-2011 standard, the neural network was used to simulate them to establish a numerical relationship between structural dimension and the value of concentration factors by the nonlinear mapping ability; the genetic algorithms global optimization function was used to design the structural parameters of cylindrical shell with big nozzle; a design process and optimization software of large opening cylinder with nozzle were developed; the best value of the structural design for a given design conditions was get. The influence of model length on the calculate results of the finite element analysis software ANSYS has been analyzed and a structure dimension suitable for ANSYS analysis has been proposed, by which the optimization results verified. Using the data obtained by genetic algorithms to conduct nonlinear function fitting so as to get the structure’s best ratio equation, by which provides another way to fast computing for the optimization calculation.Firstly, based on the characteristic parameters analyzed of the model of analysis method of cylindrical opening with radial nozzle reinforcement design, a three-input and two-outputs BP neural network is designed to simulate the relationship between the structural dimension p δet/δe、λ and the stress concentration factor Km、K, and then DPS data software is used to design a 165 trials uniform design, creating a good uniform distribution of training samples and test samples, which shows that a three hidden layer with 12 neurons BP network has a better mapping ability compared with other designs. The final determined neural network eventually gets the maximum relative error less than 4%, so the standard data is stored in the three layers BP neural network and the relationship between the structural dimension and the value of concentration factors provides forecasts for the stress concentration factor of opening structure.According to the analysis of optimization model, the volume formula of the large opening cylinder with nozzle is given and the optimization function and constraints are determined. Combined with genetic algorithm to select the genetic algorithm parameters and MATLAB software, a structure optimization program of large opening cylindrical with nozzle is organized to achieve and output automatic optimization design calculations under a given design parameters. The model length of the opening nozzle suitable for the finite element analysis software ANSYS is discussed, which is used in the ANSYS software optimization to verify the reliability of the results of genetic algorithm optimization. The results show that the combination of neural networks and genetic algorithms to optimize the design process can replace the ANSYS design optimization procedures, and has a significant role in reducing the designing time and steps.The data obtained by genetic algorithms are used to conduct nonlinear function fitting so as to get the structure’s best ratio equation, providing another way to fast computing for the optimization calculation, and the "Large opening cylinder with nozzle optimization software system" is developed.Large opening cylinder with nozzle optimization which combines with bionic algorithm has a strong practical value in terms of reducing the cost of equipment, and has certain significance in optimizing the structural parameters of special pressure equipment.