Optimization Of Stator Winding End Structure Based On Intelligent Algorithm And Approximate Model

Turbine generator stator end winding structure optimization is an important aspect of generator design,and its reasonable design is essential for the safe operation of the generator.If the stator end winding resonates under the action of twice the power frequency electromagnetic force,or the vibration under the action of the electromagnetic force is too large,the end winding structure may be destroyed,which has serious consequences.Therefore,in the design of the end winding structure,the natural frequency corresponding to the modality of the specific shape is required to avoid the resonance region,and the vibration amplitude under the action of the rated load electromagnetic force is smaller than the specified value.Particle swarm optimization has the advantages of simple principle,few parameters and strong implementation,and has been successfully applied in many engineering structure optimization fields.Considering the premature problem that particle swarm optimization algorithm has always existed,the particle swarm optimization algorithm is improved by introducing adaptive inertia weight.In the optimization process,the finite element is used to obtain the natural frequency of the end winding and the amplitude of the end winding under the electromagnetic force as the objective function value and the constraint value.However,since the finite element calculation takes a long time,a support vector machine approximation model needs to be established instead of the finite element calculation,which shortens the optimization time and improves the optimization efficiency.Therefore,the thesis proposes an improved particle swarm optimization algorithm and a support vector machine approximation model to optimize the end winding structure.The main research work and contents of this paper are:1.Firstly,the thesis introduces an adaptive inertia weight and proposes an improved particle swarm optimization algorithm.The design variables in structural optimization are usually discrete variables with constraints,so the variables are discretized and a penalty function strategy is introduced to deal with the constraints to make them more reasonable.Taking the typical truss structure as an example,the improved particle swarm optimization algorithm has good stability and effectiveness,and can solve the structural optimization problem well.2.Secondly,the thesis uses the support vector machine to build an approximate model instead of the finite element calculation.The parameter selection of the support vector machine has a great influence on the performance of the established approximation model.In this thesis,the choice of parameters is regarded as an optimization problem.The improved particle swarm optimization algorithm is used to find the optimal parameters.According to the selected optimal parameters,the structural response approximation model with the optimal parameters is obtained instead of the finite element calculation.3.Finally,according to the approximate model with the optimal parameters,combined with the improved particle swarm optimization algorithm,the stator end winding structure of the turbo generator is optimized.Through the final results,it is verified that the improved optimization method using the improved particle swarm optimization algorithm and the support vector machine approximation model can obtain a good optimization effect,which can provide an effective solution for structural optimization.