Abstract: The compressive strength after low-velocity impact is a key performance index of the composite laminates. By analyzing data of compressive strength experiment of CCF300/5428 composite laminates after low-velocity impact (LVI) under different energy levels, a prediction model based on generalized regression neural network technology of compressive strength after LVI is proposed. This model uses the impact energy, the dent depth and the damage area as input parameters and the Gauss function as the hidden layer excitation function. The model uses part of the experimental data to train for finding the optimal smoothness coefficient. Then, the model is reconstructed with the optimal smoothness coefficient, and is simulated with part of the experimental data. Results show that the model has good generalization ability of experimental data and is feasible to estimate the compressive strength after low-velocity impact.

Key words: neural network; composite materials; low-velocity impact; compressive strength; estimation