Processing a PBT by twin screw extrusion : how to fix appropriate operating conditions by using numerical simulation

Processing a saturated polyester material

This study has been performed by SCC with the support of Proplast, at the occasion of a specific workshop on twin screw extrusion issues, organized by Proplast, in Rivalta Scrivia (Italy).

This practical shows the interest of using the numerical simulation for fixing appropriated operating conditions for processing a saturated polyester material by twin screw extrusion.

Indeed, sometimes, experimental data are not enough for undertaking what is happening all along the process. That is the reason why simulation, within the Ludovic® software is used for recording the whole thermo-mechanical history during the extrusion process.

In this study we demonstrate that the measurements performed at the screw head (such as temperature, pressure and RTD) consist in a first information about the material. But they are not necessarily relevant of the complete thermo-mechanical history all along the process. Hence it is necessary to deeply analyze the impact of the process conditons on the material.


The studied polyester is a polybutylen therephtalate (PBT B 2550, Ultradur® from BASF), processed on a lab extruder. The envisaged profile is classic for compounding applications and is defined this way (on a ZSE 27mm Leistritz extruder):


Real screw design (top picture) and modeled screw design in Ludovic (bottom)

The profile is then modeled in the Ludovic® software. Like on the real design, the screw holds three main functional zones : a first zone composed by successive kneading blocks is used for melting the product. Then the mixing zone is followed by two areas aimed at the dispersion (for this case they won’t be really used).

Results comparison

As pressure and temperature are measured at the screw head, they are compared to the Ludovic® simulations, for fitting the trials campaign. Results of the simulations are really meaningfull (see the opposite picture) and fits to the measurements.

Comparison of measurements and simulations results

Comparison of measurements and simulations results

Simulation as a predictive tool is really useful as it describes the material reaction during the complete process (and not only in the sensors areas). In this case, it is really important to consider the temperature during the process and NOT only at the screw head.

Indeed, the PBT temperature is continuously decreasing after the first melting zone. Its maximum temperature is really higher in the upstream screw elements than in the downstream elements (difference of 20°C). That is the reason why its measured temperature (at the exit) is not really relevant of the process (on the contrary to the PP whose temperature profile is really straightforward – difference of 7°C).

Ludovic Results : relevant temperature - comparison of PP and PBT behaviour

Ludovic Results : relevant temperature – comparison of PP and PBT behaviour

In this case, simulations results underlines what measures can not show.


As some limitation criteria appear (such as maximum temperature for the product or maximum torque of the machine), some combinations of operating conditions become forbidden. In order to optimize the chosen parameters, the automatic DOE (Design Of Experiments) of Ludovic® is used. It allows to fix targets as consign values; such as the maximum torque supported by the machine (here roughly 120n.M) and the maximum authorized temperature of the PBT (about 280°C as precognized by BASF). These only two limitations open a functioning domain for the envisaged product and extruder

Design of Experiments : how to explore a wide functioning domain

Design of Experiments : how to explore a wide functioning domain

It appears as a green area on the DOE. This zone matches to some couples of operating conditions included between 200rpm/20kg/h and 700rpm/150kg/h.

Ludovic® benefits – saving time thanks to virtuals experiments

In a few hours, using the numerical simulation within Ludovic® affords to explore a large functioning domain with about 300 configurations. 30 min are necessary for the simulation set up. Then in roughly 90 minutes, global trends analysis are performed.

In comparison, the installation of a single screw profile and the first tests of operating conditions mean some hours in a lab.


Using the Ludovic® software for preparing a campaign trials is useful as it allows to

  • get relevant data about the material behaviour (analaysis of the main trends)
  • define an adapted functioning domain, depending on the target and the product/machine limitations

Special thanks

To all the Proplast team who performed the experiments and measures

To Maria Rosa CONTARDI for her welcome on the Proplast site

To Matteo LAVASELLI and Alberto PRIORE for their support

Leave a Comment