Modeling an Efficient Algorithm for Melt Pool Analysis in Metal Additive Manufacturing

Melt pool analysis assesses the size, shape, temperature, and dynamics of molten material in additive manufacturing (AM) to model algorithms based on thermal behavior, material properties, and overall process efficiency. The research defect in a melt pool is using a Harris Hawk Optimized Deep Convolutional Neural Network (HHO-DCNN) based on the essential variables (laser power, scan speed, beam diameter, cooling rate, and shielding gas composition). It used Min-Max normalization to prepare the data, and afterwards, extracted features using the Discrete Wavelet Transform (DWT), Principal Component Analysis (PCA), and Fast Fourier Transform (FFT), to enhance the accuracy of the analysis and prediction of the melt pool data. The proposed technique performs more effectively on the melt pool dimension forecasting with 92-94% accuracy, which is superior to the traditional algorithms because of the R 2 of 0.9123, the RMSE at 0.0003, and the MAE of 0.0131. It makes the analysis of the melt pool more accurate and significantly available. The AM algorithm on metals enhances better accuracy, speed, quality control and optimization of the melt pool to overcome challenges in optimizing production processes with different materials and conditions of production.