The milling process is a complex phenomenon dependent on various technological and material parameters. The development of a fluidized bed jet milling model is of high practical significance, since milling is utilized in many industries, and its complexity is still not sufficiently recognized. Therefore, this research aims to optimize fluidized bed jet milling with the use of fuzzy logic (FL) based approach as one of the primary artificial intelligence (AI) methods. The developed fuzzy logic model (FLMill) of the investigated process allows it to be described as a non-iterative procedure, over a wide range of operating conditions. Working air pressure, rotational speed of the classifier rotor, and time of conducting the test are considered as inputs, while mass and mean Sauter diameter of the product are defined as outputs. Several triangular and constant linguistic terms are used in the developed FLMill model, which was validated against the experimental data. The optimum working air pressure and the test's conducting time are 500 kPa and 3000 s, respectively. The optimum rotational speed of the classifier is equal to 50 s-1, considering the mass of the grinding product, and 250 s-1 for the mean Sauter diameter of the product. Such operating parameters allow obtaining 243.3 g of grinding product with the mean Sauter diameter of 11 µm. The research proved that the use of fuzzy logic modeling as a computer-based technique of solving mechanical engineering problems allows effective optimization of the fluidized bed jet milling process.
Keywords: comminution; fluidized bed jet milling technology; fuzzy logic; grinding; knowledge-based systems; simulation.