Research On Hydraulic Optimization Design Of Francis Runner Based On Automatic Approximation Of Blade Profiles On Stream-surface

Turbine runner is the core component in energy conversion of hydraulic generator,the hydraulic design technology has been the key points and special difficulties in the process of turbine runner development.Turbine runners hydraulic design is not only strive for high efficiency,but also need the good hydraulic stability and to ensure the high cavitation performance,strong load adaptability,wide efficient region,and requires reliable structure and mechanical property.In order to ensure the requirement of the design of hydrodynamic performance,the author researches from the method of the design,and designs the blades with optimal hydrodynamic performance and can meet the requirements of structural characteristics,manufacturing processes,etc,to maximum reduce the human factor during development process.That's why it is very important to explore a rapid and convenient method for the hydraulic optimization design.In this paper,according to the characteristics and main performances of the hydraulic design of Francis turbine,the author explores a new design of hydraulic turbine runner which is based on the Artificial Neural Networks and Genetic Algorithm.The author use the existing geometric model of Francis turbine runner blades or traditional designed initial blade,along the steam surface and parameterize the original blade to get a series of space blades.Then get the parameterized type of blades by using a series of polynomial curve to approach the 3D data of blade.And take aim at the highest efficiency and minimum cavitation coefficient performance,and set the blade's emplacement angle,wrap angle and head radius as optimization variables,adopt the method based on Aritificial Neural Network by Genetic Algorithm to optimized calculate the blade under designed and non-designed working condition,and eventually come to the optimum blade with highly efficient,small cavitation coefficient and good stability under both designed and non-designed working conditions.This paper put forwards the study of optimized scheme and technique of Francis turbine and main performance requirements: under the designed condition,the work efficiency of runner increased by 0.7%;under off-designed condition,the work efficiency of runner and the pressure value of the lowest pressure point is improved;and the amplitude of pressure fluctuation after designed is obviously better than off-designed one under most working conditions.This paper demonstrates the feasibility of the Francis turbine runner design by using the Artificial Neural Networks and Genetic Algorithm through examples,and to provethe accuracy of optimization and technical method.The hydraulic optimization design based on multi-target and different working conditions could comprehensive predicate and control the performance of designed blade.At the same time,it could shorten development cycles and decrease human factor during development process.And it has certain theoretical significance and practical engineering value.