ANN assisted optimization of a multimode linearly tapered bimorph PYT-5 cantilever beam for low frequency energy harvesting

The data on the cantilever's dimensions are considered as input and the output resonant frequency of the first mode and the generated power are considered as output. They are used for training the artificial neural network (ANN) that would provide us with the fitting function that we modify (scalarize) and use in algorithms for the optimization aiming to achieve minimal resonant frequency and maximal generated power. Two methods were used for the training of the ANN, Levenberg-Marquardt (LM) and Scaled Conjugate Gradient (SCG). For optimization we used the goal attainment method (GAM) and genetic algorithm (GA).本数据集以悬臂梁的尺寸数据作为输入，以其一阶模态的谐振频率与产生的功率作为输出。上述数据被用于训练人工神经网络（Artificial Neural Network，ANN），所得拟合函数可经标量化修改后，应用于优化算法以达成最小谐振频率与最大输出功率的优化目标。本次训练共采用两种方法：莱文贝格-马夸尔特法（Levenberg-Marquardt，LM）与缩放共轭梯度法（Scaled Conjugate Gradient，SCG）。优化阶段则采用了目标达成法（Goal Attainment Method，GAM）与遗传算法（Genetic Algorithm，GA）。