Self-optimising process control strategies for highly segmented mould temperature control in injection moulding



Christian Hopmann



+49 241 80 93838



In melt based processing methods such as injection moulding, the process variables pressure and temperature can be used to influence the specific volume and thus the shrinkage and warpage behaviour during the cooling of the part. If a homogeneous specific volume is realised in the moulded part, a homogeneous shrinkage potential results. Residual stresses and warpage can be minimised in this way. A homogenisation of the specific volume of the part cannot be achieved via pressure, as it is only applied at the position of the sprue and decreases over the flow path. However, the specific volume can be specifically influenced and homogenised via the local manipulation of the mould temperature, considering the local process pressure. The aim of subproject B03 is the targeted control of the specific volume for distortion reduction via the manipulation of the local mould temperature. In the first funding phase of SFB1120, an injection mould with segmented highly dynamic temperature control and in-situ data acquisition was developed in SP B03 (Fig. 1). This makes it possible to control 18 temperature control zones (9 per mould half) individually and dynamically (see video).

  Moving side of the highly segmented injection mould with temperature control zones (left), schematic representation of one temperature control zone (right). Copyright: © SFB 1120 Fig. 1: Moving side of the highly segmented injection mould with temperature control zones (left), schematic representation of one temperature control zone (right).
*** B03_Video1 ***
Video 1: Thermographic recording of dynamic temperature control of the highly segmented injection mould .

As shown in Fig. 1, right, an infrared temperature sensor is installed in the centre of all zones, which measures the local part temperature in real time. In zones 2, 5 and 8, the melt pressure is also measured in real time. With the help of the pressure and the temperature, the pvT behaviour of the plastic can be determined and the moulding process can be controlled.

To develop the methodology, a simple plate-shaped geometry was initially selected and, due to the shrinkage properties and the wide range of applications, the material polypropylene. In Fig. 1, right, it can be seen that there is a distance of at least 9 mm between the heating element and the moulded part, which results in a delay between the activation of a heating element and a temperature change on the cavity surface.

Therefore, based on a discretised one-dimensional Fourier's equation, a model predictive control (MPC) approach was developed, which considers the thermal inertia in the mould of approximately three seconds and thus enables a temporally precise control of the segmented temperature control. Initially, the control accuracy of the MPC approach was comparatively low. In order to improve the control quality, model extensions and optimisations have been conducted for the MPC approach. For example, the accuracy of the predicted temperatures was improved by a combination of injection moulding simulations and thermographic recordings of the real system bahviour, and the influence of neighbouring temperature control zones is considered. Furthermore, optimisations were made in the selection of the temperature control scenarios, so that the required heating power is determined more precisely.

In order to investigate the control quality of the optimised MPC approach, practical test series were carried out in which distortion was to be provoked by means of an inhomogeneous target specification of the specific volume. In addition to validating the control precision, the aim was to investigate how reproducible the warpage of the moulded part can be controlled. The local specific volume and the resulting warpage could be precisely and reproducibly controlled with a relative accuracy of 2.6 % using the local temperature control (Fig. 2). Compared to a PID controller, the control error was reduced by a factor of eight with the help of the MPC approach.

  Evolution of the specific volumes (left) and resulting delay for target, PID and MPC controllers (right) Copyright: © SFB 1120 Fig. 2: Evolution of the specific volumes (left) and resulting delay for target, PID and MPC controllers (right)

The test series shown in Fig. 2 deliberately generated component distortion in order to illustrate the relationship between temperature control, specific volume and warpage. In current work, the aim is to reduce part warpage across the cycles via temperature control using the MPC approach. To achieve this, a self-optimising control module will be introduced, which is intended to homogenise the local specific volumes of the plate-shaped geometry. In addition to a continuous warpage reduction, thermal transient processes in the injection mould are also eliminated with the help of the MPC approach.