Effects of temperature and time on chocolate three dimensional printing using optimized backpropagation and lattice Boltzmann method

Abstract

Three dimensional (3D) printing introduces a new method for customized food processing. Material extrusion (MEX) can serve as a reference for achieving free-form chocolate fabrication. However, limited accuracy hampers the technology’s ability to produce parts with complex or subtle surface details. This work examines the impact of two key parameters, temperature and time, on printing accuracy. Fluid flow analysis is performed to explore the relationship between flow and accuracy. Both temperature and/or time directly influence viscosity; changes in viscosity result in different outlet velocities (also known as extruded velocities), which in turn affect printing accuracy. The most significant contribution is the proposed integrated frame combining rheological characterization, LBM numerical simulation, artificial intelligence (AI) and experimental validation, which explains the potential relationship between the rheological properties and printing quality and reveals the mechanism of how time and temperature affect printing properties. Based on viscosity prediction, the average root mean square error (RMSE) is reduced by 66.56% compared to fitted equations. The viscosity model is incorporated into the external force lattice Boltzmann method (LBM) to determine actual extruded velocity and develop a velocity prediction model (R² approaching 1, RMSE of 1.7311 × 10⁻⁴), aiding in selecting process parameters for future experiments. Printing tests on straight lines and circular rings demonstrate that width increases with rising temperature and/or time. The optimal melting temperature range is identified as 37.2–37.6 °C. The width prediction model, which uses extruded velocity as input, shows an average R² of 0.9705 and an average RMSE of 0.0072. The research can provide theoretical basis for extruder structure design and printing parameter selection, especially for printing process optimization of hot-melt paste represented by chocolate.