TECHNOLOGY | PROCESS MODELLING
best heating power for different heater cartridges used in the die head, according to Schmudde. FEM simulations can also be used to troubleshoot extruder vibration problems or to predict and prevent them. As twin-screw extruders have been pushed to higher output rates with high drive power, the rotating parts (such as motor, gearbox, screws)
can result in structural vibration, according to Coperion. “With FEM
models of our machines including all drive parts – the process section and the base frame as well as the connection to the foundation – we are able to simulate vibra-
Above: Simulation of flow velocity for polymer melt passing through a screen mesh
tions at different operating windows of the extrud- er,” says Schmudde. Eliminating potential vibration sometimes
requires only a small change to the machine design, he explains. “In other cases, the machine design is correct but the foundation was not done according to the vendor’s recommendation. But even then, some small adaptions can be made to the machine to improve the vibration level, and the result is very important for the lifetime of the machine.”
Right: FEM analysis can help optimise the thermal profile in the die head,
enabling more uniform pellets
Modelling dispersion Scientists from Chemours Company in the US used a modelling program in a study of the effects of twin-screw extrusion process parameters on titanium dioxide pigment dispersion in a highly loaded masterbatch. Other modelling experiments had not experimentally connected dispersion quality with extrusion parameters for a given screw design, according to the paper given at the Society of Plastics Engineers Color and Additives Division regional technical conference (CAD RETEC) in 2016 by Davis, Niedenzu, Reid, and Sedar. Using computer simulation, the researchers
looked at various screw designs and how they impacted fill factor, residence time distribution and viscosity, and then how these factors affected temperature and the ratio (Q/N) of feed rate (Q) to screw speed in RPM (N), which are two key param- eters for dispersion quality for any screw design. The simulation allowed them to collect many experimental points and get a clear picture of what was happening in the screw. Actual experiments in an extruder, with dispersion quality measured by
38 COMPOUNDING WORLD | March 2018
www.compoundingworld.com
screen pack dispersion values, seemed to corrobo- rate the modelling results, the scientists reported. Quantisweb, from Quantisweb Technologies, is an adaptive, expert-driven stochastic approxima- tion optimisation (SAO) software system that automates the process of creating a model and generating a production recipe. “A veteran employee can go into a lab and get good answers based on their vast experience. A new hire, however, has to either develop that experience over time or be taught by the veterans. The software bridges that skills gap by taking the art and turning it into knowledge that can be reused and refined,” says Gilles Gagnon, President and CEO of Quantisweb. This skills gap is a is a universal problem in
maturing industries such as plastics, says Bill Blasius, who is the “Voice of the Customer” advisor to Quantisweb. He says the software’s predictive analytics can use legacy data to create new knowledge and build models so veterans can “train forward.”
Seeking patterns SAO uses a mathematical approach to find pat- terns. “You start with the specifications for a product, a range for those specifications, and their relative importance, along with a list of ingredients and process variables. Quantisweb then generates a minimal dynamic design of experiments (md- DOE),” says Gagnon. If a process has 18 variables, for example, a mdDOE would require 19 experi- ments while a traditional DOE would require hundreds (Figure 5). “You go to the lab and run these experiments, then put that data into the software. The SAO part of the software finds the patterns and creates a behavioral model. Another part of the software finds the optimal formulation and process variables using this model and the original specifications. You then go back to the lab with these optimal
PHOTO: COPERION
PHOTO: COPERION
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76 |
Page 77 |
Page 78 |
Page 79 |
Page 80 |
Page 81 |
Page 82
