Research On Automatic Optimization Design Of Mixed-flow Pump Impeller Based On Multi-objective

Mixed-flow pump is widely used in agriculture,industry and domestic production.In Nuclear Power Station and Thermal Power Station,mixed-flow pump generally been used as the circulating water pump,which has very strict standard on every item.Impeller is one of the most important parts in mixed-flow pump and its design quality has an important impact on the performance of mixed-flow pump.One-dimensional and two-dimensional flow theories were formerly used for designing mixed-flow pump vanes.These methods can meet requirement in some industrial place.However,these methods can't meet modern pumps' performance indicators and feasibility requirement.Therefore,in order to ensure mixed-flow pump operate safely and reliably and reduce design and production cost,it's urgent to study mixed-flow pump impeller's hydraulic optimization design theory and method.According to performance requirements of hydraulic design,this thesis explores how to achieve hydraulic design method based on multi-objective and multi-condition.This method would be optimized optimize a mixed-flow pump impeller combining some main parameters provided by SI CHUAN CHUAN GONG INDUSTRIAL PUMP CO.LTD.Impeller and guide were designed by traditional method firstly,according to the principle of automatic optimization,the initial impeller parametric model has been fitted,and the blade three-dimensional data was described by a series of cubic polynomial exactly.In automatic optimization,we need outlet of impeller flow field calculation results as boundary conditions,so all passage numerical simulation of initial mixed-flow pump was conducted and the impeller outlet boundary conditions were extracted.In general,we should focus on external characteristics of indicators.Based on genetic algorithms and artificial neural networks,the blade angle was as optimization variable and efficiency was optimized under design condition.In automatic optimization,we also consider some problems on off-design conditions,such as efficiency and head can't get the target;impeller pressure is instability and so on.Based on above research,the second optimization design was conducted.Through the final optimization,the best impeller was gotten.Based on CFD,full passage numerical analysis of optimized impeller was introduced and results were discussed.The results show that optimized impeller's efficiency improves 2.11% on design conditions;the maximum head improves 2.02 m.In addition,optimized impeller's efficiency and head improve at other conditions and there is no hump in operating.According to these results,we can know that the blade angle has important effect on mixed-flow pump's efficiency and head.Comparison of cavitation performance,optimized impeller cavitation resistance is better than initial impeller,the result shows that the blade shape of pump inlet has effect on mixed-flowpump's pressure and cavitation performance.Researched on pressure pulsation,the impeller's pressure pulsation amplitude reduces 84.95% maximally after optimization and high frequency pressure pulsations component is less than initial impeller.According to the comparative results,the method of automatic optimization based on genetic algorithms and artificial neural networks has achieved good results and met the requirements of hydraulic design.The results proved the feasibility of automatic optimization method.The above research results can provide theoretical support for improving the performance of mixed-flow pump and reducing noise,vibration in operation and improving unit safe and stable operation.This research theory settles a foundation for mixed-flow pump's optimal design and other sets of turbo-machinery industrial areas.