CFD–ANN dataset and surrogate-optimisation code for a Savonius VAWT: power coefficient and torque ripple across the TSR–overlap-ratio design space

This dataset contains 170 CFD-simulated operating points for a two-bladed Savonius vertical-axis wind turbine, sampled across the tip-speed-ratio and overlap-ratio design space using Latin hypercube sampling, together with the Python code used to train ensemble ANN surrogate models and perform constrained multi-objective optimisation. It supports the associated manuscript on CFD–ANN surrogate optimisation of a Savonius VAWT.本数据集包含170个经计算流体动力学（Computational Fluid Dynamics，CFD）模拟得到的双叶片萨沃纽斯垂直轴风力机运行工况点，采用拉丁超立方抽样在叶尖速比与重叠率的设计空间内完成采样；同时附带用于训练集成人工神经网络（Artificial Neural Network，ANN）替代模型并执行约束多目标优化的Python代码。本数据集可支撑关于萨沃纽斯垂直轴风力机的CFD-ANN替代优化相关学术手稿的研究工作。