Font Size: a A A

Optimization For Siphon Pipe Based On ANN-GA

Posted on:2008-11-11Degree:MasterType:Thesis
Country:ChinaCandidate:Z Z LiFull Text:PDF
GTID:2132360215480092Subject:Mechanical design and theory
Abstract/Summary:PDF Full Text Request
Closestool plays an important part in people's model life, and its water consumption rate is very high. To research and produce a type of closestool with high quality, also saving water is vital for the water resources of the world. This paper uses CFD (Computational Fluid Dynamics ) and artificial intelligence way to do research on the optimization of the siphon. And this is also a new way of research in the water machinery field.CFD is widely used in many fields presently. It has strong agility, and can take particular and integrity records of data .It can easily simulate many conditions that can not be reached in experiment. And it can reduce outlays for research and design, shorten cycle of the new product, advance the design rate of the product. In this thesis, there are also many experiments data to prove the dependability of the simulation.Because of the complexity of the siphon, it is impossible to make a normal function between the shape and the effect of siphon. So firstly according to the analysis of experiments and simulation, this paper find out three parameters that influence the siphon effect farthest, and two parameters that can embodiment the siphon effect. Secondly the dynamic simulations are done to substitute for experiments, and 25 samples are gotten for the ANN(Artificial Neural Networks) training. The BP net is built in the Matlab soft environment, and it is the mapping between parameters to optimize and the goal. At last this paper optimize the siphon pipe using GA(Genetic Algorithm).As there are two objects, this paper raises a new kind of the Multi-Objective Genetic Algorithms(MOGA),which is based on ranking of Pareto solutions, sharing and niche technology, and manages to apply it to the Multi-Objective Optimal Design(MOOD). The result is a group of Pareto solutions that decision-maker can choose any result in the group according his desire.Besides the optimization of the single hydraulic control point, this paper also research on a new siphon pipe with double hydraulic control points, and the results show that this kind of pipe behaves very well in siphonage and water saving.
Keywords/Search Tags:Siphon Optimization, Computational Fluid Dynamics, Multi-objective Genetic Algorithms, Artificial Neural Network, Nouble Hydraulic Control Points
PDF Full Text Request
Related items