Abstract
AbstractCompared with traditional reactors, lead‐bismuth fast reactors have broader development prospects. Based on the operating characteristics of these reactors, this article proposes a design scheme for steam turbine generators suitable for small‐scale lead‐bismuth fast reactors. To achieve the design requirements of high efficiency and high‐power density for steam turbine generators simultaneously, a multi‐objective optimization method based on a feedforward neural network surrogate model is proposed. First, the generator losses and power density are analyzed to obtain the structural parameters that affect the generator optimization objectives. The selected structural parameters are then subjected to sensitivity analysis and data sampling. Subsequently, a feedforward neural network model is used to replace the finite element model, and based on this, a multi‐objective genetic algorithm is employed to globally optimize the efficiency and power density of the generator. The final preferred scheme is obtained from the solved Pareto solution set. Meanwhile, the finite element method is used to verify and analyze the optimization results. The optimization results show that while ensuring the generator efficiency, the power density is further improved. Finally, the temperature rise of the generator is analyzed, and the results show that the temperature distribution of the generator is reasonable.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call