On the control problem in fluid energy milling and vortex classification processes: approaches for experimenting and modeling particulate systems in an industrial scale plant.
Systematic methods and tools for managing the complexity
Process Operation, Monitoring & Analysis (T4-2)
Keywords: Fluid energy milling, classification, modeling, particulate systems.
Particulate processes and powder materials are used in many different industries such as pharmaceutical products, food processing, powder metallurgy, painting, mineral processing and chemicals in general. Particle size distribution and energy efficiency are important performance outputs for size reduction and classification operations. Particle size distribution is important to provide quality characteristics for the adduced products. In fact, it contributes to the taste of the chocolate and also influences the absorption of drugs by the human organism. It can interfere in the print quality of laser printers and modify the mechanical resistance of ceramic cutting tools. In addition to the quality aspects, it is important to emphasize that size reduction processes are known to have very poor energy efficiency. Many important authors have reported that less than one percent of the total energy input is used in the breakage process. For instance, fluid energy mills, where the comminution results mainly from particle to particle collisions, consume most of the energy to put the particles in motion, and only around 0.1% of the total energy is used in the breakage of the particle itself. It is self evident that energy efficiency is probably the most important problem that the industry is facing in the recent past and industrial engineers will continue to face this problem now and for long time in the future.
A considerable effort has been put in by many scientists and powder processing researchers in order to obtain suitable models for this branch of industrial systems. Most of the works describing those systems are based on population balance models, process rate models, and more recently approaches using Markov chains theory. The mentioned phenomenological approaches are often deemed to be very time-consuming and intensive in laboratorial testing when defining and adjusting model parameters related to mechanical properties of the processing particles. It is not very common to find works in open-literature dealing with multiple inputs and output variables models and only very little research has been conducted based on black-boxes and nonlinear identification methods. Fine grinding and particle size classification, two of the most important operations in particulate processes, are described in this paper from the perspective of implementing a control strategy for an industrial scale plant. The multivariable control approach is discussed for the fluidized air jet mill and the rotor classifiers. Control problems are formulated considering pervasive investigation on the more recent literature and a practical and realistic nonlinear black-box approach is proposed to experiment and model such systems in a real-world industrial plant. Particle size distribution and energy efficiency are defined as the most important outputs to be controlled. Difficulties in performing model identification of dynamic processes without considering the availability of a laboratory scale pilot plant and without using an on-line particle size measurement device are discussed and results from preliminary experiment are presented.
See the full pdf manuscript of the abstract.
Presented Tuesday 18, 10:05 to 10:25, in session Process Operation, Monitoring & Analysis (T4-2).
