一种使用浅层神经网络增强BEM模型的方法

An Approach Using Shallow Neural Network to Enhance Blade Element Momentum Model

水平轴潮流能水轮机是常用的潮流能获取装置,载荷预报是设计过程中的重要问题。BEM方法被广泛应用于水平轴水轮机的载荷预报,由于该方法计算速度快,常在早期选型优化阶段使用。BEM方法的核心步骤是通过迭代方式求解诱导系数,存在收敛性问题。文章使用没有收敛性问题处的数据进行神经网络训练,利用训练后的神经网络对产生收敛性问题处的诱导系数进行预报,实现了扩展运算域的效果。研究发现,BP神经网络模型能够有效解决收敛性问题,拓宽运算域,预测的平均误差在4%以内。通过研究BP神经网络的最优结构,为后续的相关研究提供了设计建议。

Horizontal axis tidal current turbine is a commonly used energy acquisition device,and load prediction is an important problem in the design process.BEM method is widely used in load prediction of horizontal shaft turbine.Because of its fast calculation speed,it is often used in the early stage of selection and optimization.The core step of BEM method is to solve the induction coefficient by iterative method,which creates the problem of convergence.The data without convergence problem is utilized to train the neural network,and the trained neural network is used to forecast the induction coefficient of the convergence problem,which achieves the effect of extending the computing domain.It is found that the BP neural network model can effectively solve the convergence problem and broaden the computing domain and the average error of the prediction is less than 4%.The optimal structure of BP neural network is studied,which provides design suggestions for the subsequent related research.