3.2. Part B: Analysis of the Machine-Specific Behavior of Three Different IMMs at Varying Operating Points
In addition to the start-up behavior, the behavior of the machine in a steady state is of interest. For this purpose, the influence of the machine type on the polymer mass depending on the machine settings was studied. Therefore, the experimental design from
Table 4 was used to set different operating points of the machine. Immediately after each cycle, the mass of the ejected polymer was measured, which is presented in
Figure 4.
It is important to emphasize that the dosing and switchover volumes were set identically for each machine at the individual operating points. Theoretically, the masses ejected at the individual operating points should be the same, but this could not be observed.
Figure 4 prominently illustrates significant differences in mass based on the operating point and the specific machine utilized. The results show that machines A and C show larger variations in ejected mass with changes in setting parameters than machine B. To visualize the individual effects of the setting parameters on the ejected polymer mass, regressions were carried out to create effect diagrams. The regression as well as the coefficient of determination are described in the section “Regression models” of
Appendix A. The effect diagrams depending on the machine used are shown in
Figure 5. Since the dosing and switchover volumes were varied by the same amount, only the dosing volume is used as representative of the screw stroke. The effect diagrams show that the mean ejected mass using machines A and C is strongly dependent on the set cylinder temperature, followed by the flow rate and the injection stroke. In comparison, changes in the operating point have hardly any effect on the ejected mass when machine B is used. This is surprising since machines A and B are hydraulic, and C is all-electric.
To be able to explain the reason for the strong variation in the ejected mass depending on the machine used, the time series of the process parameters were evaluated. For visualizing the multitude of time series and the influence of the machine-specific behavior on the time series of the process parameters, the max. flow rates, the max. injection pressures, and the actually achieved dosing as well as switchover volumes were calculated.
Figure 6 visualizes the calculated characteristic values (green) using three exemplary time series of the flow rate, screw volume, and injection pressure.
To visualize the volumetric deviations, ΔV
d was calculated as the deviation of the screw position at the end of metering to the set dosing volume and ΔV
s as the deviation of the screw position at the end of injection to the set switchover volume.
Figure 7 shows a schematic representation of the screw positions. Positive values of ΔV
d and ΔV
s lead to an increase in the injected volume.
The deviations of the volumes were calculated using the following formulas with V
i,target as the set volume and V
i,actual as the achieved volume. To determine the achieved volumes, the time series of the screw positions were evaluated (see
Figure 6). V
d,actual was calculated using the maximum and V
s,actual using the minimum of the time series.
To convert the volumetric deviations into the deviation of the screw stroke Δs, the following formula was used.
The computed values for the deviations of the dosing and switchover volumes as well as the screw strokes s calculated from these values are visualized in
Figure 8. The values ΔV
d resp. Δs
d show that the hydraulic machines A and B meter more material than set, with machine A showing the highest deviations. In contrast, machine C meters almost exactly the set dosing volume. When considering ΔV
s, machine B shows the greatest precision in approaching the switchover volume, with machines A and C exceeding the switchover point even more. Converted into screw stroke, machine A shows the strongest overrun in the majority of machine settings.
A dependence of the operating point and the achieved dosing volume cannot be identified, which is not surprising since the selected operating points control the subsequent process phases, and the setting variables for parameterizing the metering phase were not varied. For this reason, the effects of the setting parameters are not evaluated further. In contrast, a significant influence of the setting parameters on the screw position during switchover is already evident. For the evaluation of these effects by using effect diagrams, regressions were carried out, which are described in the section “Regression models” of
Appendix A.
Figure 9 visualizes the mean of ΔV
s depending on the machine and the setting parameters. Thereby, a clear dependence on the respective machine setting becomes apparent: at high injection flow rates, the switchover point was overrun by the screw more than at low flow rates. This seems plausible because at a higher screw feed motion, the precision required to slow down the screw is lower. All other parameters of the experimental design show less influence on the overrun of the switchover volume, except for machine C, where increased cylinder temperatures led to a further overrun of the switchover point. Furthermore, it can be seen that in some cases, there are interactions between the setting parameters, which have an effect on the screw position at the end of the injection phase. For the interpretation of the results, the drive types are presented in more detail below.
When considering the different machine designs, a variety of the effects observed can be attributed to the drive system and the respective control system. Machine A has an injection control system that pressurizes only the cylinder chamber that extends the hydraulic piston (number 1 shown in
Figure 10).
When the switchover volume is reached, the hydraulic pressure is reduced and the valves (here: hydraulic 4-3 directional control valve) are switched off, which can result in a typical overrun of the switchover volume. The main drive, i.e., the flow and pressure control, is provided by a speed-controlled radial piston pump with analog control. On the one hand, the hydraulic pipes lead to a certain inertia, but on the other hand, the pump and the control loop also have a certain dynamic. Overshooting due to the I component of the PID-control is typical for this type of injection control.
In contrast, machine B pressurizes both hydraulic chambers (number 1 and 2 shown in
Figure 11) to adjust the position of the screw and is equipped with a position control system with a hydraulic accumulator.
A constant pump is responsible for charging the hydraulic accumulator. For this purpose, both cylinder chambers are actively controlled via servo-control valves. Since this control is faster than the control of machine A, the control parameters are adapted to the hydraulics accordingly. The position control results in a more precise control of the screw position.
Machine C differs significantly from machines A and B in terms of its type of drive. The position of the screw which is installed on machine C is controlled by an electric servo motor. Ring membrane sensors (eight strain gauge cells on each side of the cylinder) ensure the precise measurement of the position. It is typical with electric drives of this type that the machines can set the target injection flow rate very precisely. The braking of the screw, on the other hand, exhibits a certain amount of lag since the servo motor can only achieve a limited deceleration.
After explaining the respective drive systems, the observed results can be explained as follows. The hydraulic machines meter more material, with machine A overrunning the switchover point the most due to its unidirectional hydraulic system. The servo motor of the electric machine meters the set dosing volume. The hydraulic injection molding machine with positioning control (machine B) enables the highest precision when approaching the switchover volume. This is due to its clamped system of hydraulics. The electric machine (machine C) shows an overrun of the switchover point due to the limited deceleration of the servo motor during braking. Here, both machines A and C show a clear dependence of the overrun characteristics on the flow rate and additionally for machine C on the cylinder temperatures. The screw position of machine B is not significantly influenced by the temperature or the screw stroke.
In the following paragraphs, the characteristics of the time series and their significant differences will be examined in more detail.
Figure 12 shows the time series measured when the machine was adjusted according to test points 3 and 4. The left side shows the time series at a flow rate of 25 cm
3/s (test point 3), while the right side shows the time series at a flow rate of 65 cm
3/s (test point 4). In the upper two diagrams, the red dashed lines represent the setting values of the flow rate and in the lower two diagrams the setting values of the switchover volume.
At low flow rates, all three machines reached the set flow rate, with machine C achieving a smooth asymptotic convergence to the set point. Machines A and B initially exceeded the set value and reached the value somewhat later after settling. In addition, the screw positions show that all three machines achieved an almost linear feed motion. At high injection flow rates, stronger differences between the time series become apparent. Machine C reaches the set injection flow rate asymptotically without exceeding it. In contrast, the hydraulic injection molding machines A and B show transient injection behavior, which can be seen from the fact that machine B exceeded the flow rate and machine A took a very long time to reach the set flow rate and then exceeded the set value. Likewise, the screw positions provide essential information on the feed movements of the screws. Machine B shows the fastest feed motion, with the switchover point being approached most precisely. Machine C, on the other hand, has a slower feed motion, but achieves the set flow rate more accurately. In addition, it can be seen from the time series of the screw position that machine C overruns the switchover point. Machine A shows the slowest feed motion, which remarkably exhibits a non-linear behavior. This is an indicator that machine A has a non-linear feed motion of the screw at high flow rate settings.
Figure 13 visualizes the maximum flow rate depending on the parameters of the experimental design. The set flow rates of the respective operating points are indicated by red dashed lines. The hydraulic machines A and B show a higher oscillation, especially at high flow rates. Furthermore, a dependence on the parameters of the experimental design can be seen for machines A and B, with machine C always reaching the set target value.
In summary, all three machines demonstrate different behavior during injection, which is influenced by the respective operating points.
In addition to the screw positions and flow rates, the resulting injection pressures were evaluated. The resulting injection pressures, two operating point examples of which are shown in
Figure 14, show significant differences. It is apparent that the machine has a significant influence on the injection pressure profile depending on the respective operating point.
The maxima of the injection pressure curves depending on the parameters of the experimental design, shown in
Figure 15, demonstrate a clear dependence on the setting of the injection molding machine and also the machine type. The reasons why the injection pressures sometimes vary significantly (e.g., test points 1 and 3 when using machine A) may be due to material fluctuations or plugs.
One explanation for the difference in injection pressures is the intensification ratio (IR). The IR describes the pressure intensification between the hydraulic pressure applied and the specific pressure in the screw antechamber (
Figure 16) [
31]. Usually, the applied hydraulic pressure affects a larger cross-sectional area inside the piston than the specific pressure in the screw antechamber. As a result, a specific intensification ratio is obtained for each injection molding machine.
Both hydraulic injection molding machines investigated in this work have an identical mounted injection unit. Consequently, the piston diameter for the hydraulics is also identical. Injection molding machine A has an IR of 15.4 with a screw diameter of 25 mm. Thus, a pressurization of, for example, one bar, results in a specific pressure of 15.4 bar in the screw antechamber. With a screw diameter of 30 mm, the injection molding machine B has an IR of 10.7. Consequently, a hydraulic pressure of one bar results in a specific pressure in the screw antechamber of 10.7 bar. For electrically driven injection molding machines, on the other hand, the pressure in the screw antechamber is identical to the pressure applied at the end of the screw. Accordingly, the IR ratio for electric injection molding machines is equal to one. Summing up, the IR factor is a relevant parameter that should be taken into account when creating a machine fingerprint.