To address the engineering challenge of installing sensors directly on aircraft lens frames for vibration measurement, this paper proposes a vibration response prediction model (CPO-BP) based on a Crested Porcupine Optimizer Algorithm (CPO)-enhanced Backpropagation (BP) neural network. The model employs a two-step optimization strategy: first, CPO optimizes the number of hidden layer neurons and the learning rate; subsequently, it refines the initial values of the network's weights and biases, thereby improving both the model's convergence speed and generalization capability. Experimental results demonstrate that the CPO-BP model achieves a high coefficient of determination (R2) of 0.97976, with a mean absolute error (MAE) as low as 0.02020 and a mean squared error (MSE) of 0.10112. Its predictive performance significantly outperforms COA-BP, DA-BP, PSO-BP, and traditional BP models. The predicted response curves closely match the actual responses across acceleration, velocity, and displacement dimensions. This study provides a highly accurate and reliable solution for vibration condition monitoring at critical locations within optical equipment.
Zheng et al. (Fri,) studied this question.