A BP neural network prediction model of compressive strength of green bricks in the Ming Dynasty based on CT scanning technique

Due to the influence of natural factors and complex loads, the mechanical performance of ancient brick structures is deteriorating, and assessing the strength of bricks is crucial for structural safety. The sintered clay bricks of the Ming Dynasty Beijing city wall were scanned by computed tomography (CT), and the grey images of the microstructure and morphology inside the brick were obtained. The parameters including porosity, large porosity, average sphericity, graded sphericity porosity greater than 0.8 and fractal dimension have the largest correlation to the compressive strength of bricks. These parameters are used as input data and the compressive strength is used as output data, and then trained with the Back-Propagation (BP) neural network. The results show that the BP neural network model is more accurate than the traditional fitting model in assessing the compressive strength, which can provide the basis and reference for the rapid and non-destructive acquisition of the strength indexes of ancient green bricks.