期刊
ENERGY
卷 283, 期 -, 页码 -出版社
PERGAMON-ELSEVIER SCIENCE LTD
DOI: 10.1016/j.energy.2023.129124
关键词
Mean-line; Steam turbine; Hot steam injection; Wet steam; Neural network optimization; Non-equilibrium condensation
This paper develops an object-oriented in-house mean-line code with the ability of hot steam injection calculation for steam turbines. A multi-objective genetic algorithm optimization and a neural network are used to redesign the steam turbine. The results show that hot steam injection can significantly improve both quality and power compared to the baseline.
Low-pressure turbines usually work in wet conditions which causes both lifetime and efficiency reduction. Hot steam injection (HSI) which has received great interest recently, is a suggested solution to reduce wetness. There is a research gap in implementing HSI in the mean-line procedure, which is a conventional method of designing and analyzing turbines. In this paper for the first time, an object-oriented in-house mean-line code is developed with the ability of HSI calculation for steam turbines. The validation of code is performed with the 3D simulation of a 2-stage axial steam turbine. In addition, considering that HSI may reduce efficiency of the turbine due to mixing entropy generation, a multi-objective genetic algorithm optimization and a neural network is used to redesign the steam turbine. Mass fraction and total temperature of injected flow from the trailing edge of blade rows are the decision variables and, liquid mass fraction and efficiency of turbine are the objective functions. The comparison of Pareto front points reveals that the maximum possible improvement of quality and power relative to baseline is 5% and 10%, respectively. Furthermore, if efficiency is the desired objective function, by enhancing 1% of steam quality, it can be increased by 0.1%.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据