·Table of Contents
·Industrial Plants and Structures
A Novel Test-Rig to Measure the Lubricating Effect of Polymer Melts in ExtrudersJulio Roberto Bartoli
Depto.Eng.Metalúrgica e Materiais - Universidade de S.Paulo/USP, Brasil
Kabwe Industrial Fabrics Ltd., Kabwe, Zambia;
Present address: Dept. of Polymer Science and Engineering, Kyoto Institute of Technology, Kyoto, Japan
Alan K. Wood
Materials Science Centre-The University of Manchester Institute of Science and Technology/UMIST, UK.
If a polymer melt film acts as a lubricant in the clearance it may tend to stabilise the screw, by means of a journal bearing effect.Ideally any deflection of the screw would produce a restoring force due to the increased pressure in the clearance at the point of deflection. Worth and Lai-Fook proposed a modification of the flight profile geometry to enhance the lubricating effect of the melt and ensure the presence of the melt film. From theoretical calculations (Theory of Lubrication) they described two suitable designs, these being the Rayleigh step and tapered land bearing . Wood showed that screw wear due to metal-metal contact may be related to the relative magnitude of the melt pressure in the clearance, when compared to the screw channel pressure (ambient pressure). Screw load experiments at 3,400 N and 100 to 150 rpm screw speed, gave melt pressures in the clearance below the ambient pressure (6-12 MPa). Wood and Whittle carried out experiments on the same equipment, using HDPE (high density polyethylene) and three kinds of screw flight profiles (flat, step and tapered). The results showed that melt pressures were higher with the flat-flight design than with the step-flight, the tapered-flight having the lowest. These results do not agree with the model of Worth and Lai-Fook and it might be due to the shorter effective parallel land of the step and tapered flights. Carlile and Fenner investigated the lubricating action of LDPE and PS (polystyrene). A critical shear stress around 105 Pa was found in the clearance, before which there was hydrodynamic/slip lubrication, metal-to-metal contact not occurring. They inferred that the magnitude of this critical stress is comparable to that for the onset of melt fracture in capillary flow, for many commercial polymers, usually attributed to melt slip. Rauwendaal sees a disadvantage in the step or tapered flight geometry when the extruder is running empty or partially empty. As the apparent contact area between screw and barrel is reduced, then the stresses are increased and adhesive wear is more likely to occur.
In this work a novel test rig was constructed to study the feasibility of using this equipment to measure the lubricating effect of the polymer melt, in terms of its pressure and screw-barrel clearance, around the complete circumference of an extruder screw section. The melt processing conditions were investigated[11,12] for three types of polymers: polypropylene (PP), low density (LDPE) and linear low density (LLDPE) polyethylenes, using a standard screw flight design and varying the screw parameters of load, speed, channel pressure and barrel circumferential position.
Test Rig Mechanical Design and Instrumentation
The design of the current test rig comes from modifications suggested by Wood of the equipment used in the previous work. The test rig (Figures 1 and 2) reproduced the operational conditions of a metering zone of a conventional extruder. A plastic extruder 50 mm O.D., 25:1, S+B Plastics Machinery Co. fed polymer melt to the cross-head coupled test rig, this being assembled on a moving frame with the drive system (electric D.C. motor, 7.5 kW) used to rotate the screw. This arrangement allowed the screw shaft to be loaded at both ends. The barrel section was supported by two end plates (bolted in the frame). The end plates had an internal labyrinth that allowed the melt to flow evenly into the barrel through four inlet holes. The barrel could rotate around the screw (Figure 2) and fixed in any circumferential position from 0o to 360o. The barrel had three cartridge heaters of 500 W each, fitted axially along the wall.
|Fig 1: Test rig: 1) barrel 2) inlet flange 3) outlet valve 4) transducer housing 5) screw 6) shaft drive 7) load bearing 8) wire ropes for load 9) end plates 10) seals.||Fig 2: Barrel, a) side view and b) cross section; showing angular marks and transducers positions: pressure (1) and displacement (2,3); dimensions in mm.|
A Kistler 601H piezoelectric transducer (PT1): 5.5 mm sensor diameter,16 pC/atm sensitivity and 100 MPa at 240 ° C (maximum), measured the melt pressures on the screw flight and channel. Two inductive displacement transducers (DT2 and DT3) were specially made for this work and used to measure the clearance between screw flight and barrel, as described in the following section. They were separated in the barrel by 1/4 screw pitch axially and 90o circumferentially, as is shown in Figure 2. A Data Acquisition Unit, ADU, (Mowlem Microsystems) was used to record up to 1000 readings/channel for each transducer, these being relayed to an OPUS PCII microcomputer.
The ambient screw channel pressure was set by a valve in the test extruder outlet (Figure 1). This pressure was measured using another transducer (Dynisco) in the flange connected to the feed extruder. A pneumatic system loaded the screw laterally against the barrel through wire ropes coupled to the screw shaft. The melt temperature was measured at the outlet of the test rig using a contact probe (Comark). A tachometer (Smiths) was used to measure the screw speed of the test rig. The barrel was fixed in one of the angular positions and time was given to stabilise the temperature at 190 °C, before to start up.
The test screw (Figure 3) was the same used by Wood, these having a flighted zone only twice their diameter (45 mm), thus only drag flow needing to be considered. The screw flight-barrel radial clearance was 0.052 mm. Only a standard flat flight screw was used in this work. The screw was made from EN19A steel and flame hardened, hardening being a common procedure to reduce screw flight wear. The screw was hollow, and was fitted as a sleeve on the drive and loading shaft. The shaft had two radial sections through which any loads were transferred to the screw sleeve. Two ball bearings acting as keys were fitted on one of these sections to drive the screw sleeve. The shaft could move into the screw sleeve, due to these bearings, thus the two point loading effect of the radial section was conserved.
|Fig 3: Test rig screw segment (dimensions in mm).|
Displacement Transducer Design
The accurate measurement of the screw-barrel clearance and the position of the minimum clearance around the screw was done using the electrical inductive displacement transducers. The inductive gauge is a copper wire coil (polyimide insulation) wound on a dielectric body (fibre glass polyimide composite) and located in the tip of the probe (Figure 4-a,b). The operational principle is related to the electromagnetic field theory: electric currents (eddy currents) are induced in a metallic target (screw flight) by the electromagnetic field generated from an electrified coil probe (displacement transducer). As the target moves into the field a voltage variation occurs in the output signal. This voltage variation is measured in millivolts and at very short distances (fractions of a millimetre) between the target and sensor device has a linear relationship.
|| Fig 4: Displacement transducer: a) body; b) tip shield distances (X,Y); c) housing.
A very small diameter transducer tip was desired for a better lateral resolution when scanning the screw flight. It was found practically that a 3.5 mm tip diameter gave a suitably high lateral sensitivity, being 0.17 mV/m
m for DT2 and 0.16 mV/m
m for DT3. The transducer depth sensitivity was improved significantly reducing the tip thickness to 1.3 mm (Figure 4-b). The transducer metallic housing (stainless steel) could give an electromagnetic shielding effect, decreasing its sensitivity (Figure 4-c). It was found that this interference is minimised keeping the metal housing apart from the transducer coil at: lateral distance ³
5mm and longitudinal ³
4mm (Figure 4-b). Two high frequency modulator/demodulator signal conditioning units were used: Vibrax IQS 603 (2 MHz iat -24 VDC, by Vibro-meter), being one for each transducer. The output from these units was fed into a simple voltage divider circuit to obtain a suitable voltage input range for the ADU, the data acquisition unit. The transducers were calibrated to ambient and processing temperatures.
Materials and Processing Conditions
Three types of extrusion grade polymers were used to study the lubricating effect of the melts: LDPE: ESCORENE LD 100 BW by Exxon Chemical; LLDPE: ESCORENE LL 6301 XR by Exxon Chemical and PP VC35 12 H by NESTE. The rheological properties of the PP, LDPE and LLDPE melts were determined using a capillary rheometer (Davenport).
A factorial experimental design (Taguchi approach) was used for the processing conditions: four factors at two levels (24) or16 run experiments. The factors and levels were: A) barrel angle position (A1=150°, A2=330°); B) screw speed (B1=100 rpm, B2=200 rpm); C) channel ambient pressure (C1=3 MPa, C2=10 MPa); D) screw load (D1=0, D2=7.2 kN). The barrel angle position at 150° means that the pressure transducer is at the point of maximum applied load to the screw, whilst at 330° it is at the point of minimum load.
|Fig 5: Screw displacements and melt pressures Vs Screw angular position, results of two test rig experiments using LDPE: no loading the screw and loading (7.2 kN); set conditions: Barrel Angle: 330°; Screw Speed: 200 rpm; Amb. Pres.: 3 MPa (DT, displacement transducer and PT pressure transducer).|
The experimental results from the test-rig suggest that the lubricating behaviour of a polymer melt film in the clearance is very dependent on its molecular structure. The lubricating action is also significantly affected by external factors, such as: screw load (due to the screw weight or lateral forces) and processing variables such as screw speed and ambient melt pressure (screw channel). The linear low density polyethylene LLDPE appeared to be an inferior lubricant when compared with the branched LDPE or the PP. The following experimental observations obtained in this work, supported also by the references, confirm the assumptions made by Paller about the poor lubricating performance of LLDPE, which could be the cause of the screw wear by adhesive mechanism:
|Fig 6: Pressure results of two test rig experiments using LLDPE: loading and no loading the screw.|
Metal-metal contact between screw flight and barrel wall may have occurred in at least two experiments with LLDPE, where the clearance was zero. The simultaneous measurements of screw displacement and clearance pressures under these conditions gave a zero clearance and oscillating melt pressures. This instability was always observed when lateral load (7.2 kN) was applied to the screw (at any level of screw speed and ambient pressure used) and the pressure transducer was diametrically opposite to direction of applied load, point of minimum load (barrel angle at 330°). This instability was not observed, even under load, when the pressure transducer was at the point of maximum applied load (barrel angle at 150°).
The factorial experimental designs showed that PP and LDPE are superior lubricants than LLDPE. This being inferred from the responses: higher clearance pressures and higher minimum clearance, at the 7.2 kN screw load condition. This screw load is equivalent to the weight of a standard 150 mm diameter extruder screw (28:1 Length/Diameter ratio). The double factors interactions were significant among all the variables. Barrel angle-screw speed and barrel angle-screw load were the most significant interactions for both responses (clearance pressure and minimum clearance), and screw speed-screw load and ambient pressure-screw load for minimum clearance.
The location of the point of minimum clearance (attitude angle) in the experiments with LLDPE were found in the main (75% of the results) in the third quadrant, that is top-right section of the barrel. Whilst in the case of the LDPE experiments half results were in the first quadrant and half in the third quadrant. The PP experiments showed that the minimum clearance was in the third quadrant of the barrel. It appears, therefore, that due to the applied load and the lubricating effects the eccentricity of the screw is such that contact is more likely to result at a location other than that of the position of maximum loading. The entire melt pressure profile around the screw circumference needs to be known in order to verify the total restoring force generated by any lubrication action. The forty eight experimental pressure measurements, for PP, LDPE and LLDPE, were taken only at two points around the screw (minimum and maximum load direction). Further work in the lubricating investigations of polymer melts should take benefit of this rotating barrel to produce more information other than melt pressures in only these two load directions.
Fig 7:Melt pressure contours on the F.E. mesh screw section (100 rpm, 3 MPa); legend A= 5.81 x1011 Pa.
Fig 8:Polymer melt pressure vs. Screw angular position, theoretical profiles at: a) 1/3; b) 1/2; c) 2/3 screw pitch; (0° indicates minimum clearance position).
Wood and Bartoli[8,11] reported an approach to analyse the entire screw flight melt pressure profile using a theoretical model for the lubricating effect of polymer melts, based on the Finite Element method and the Reynolds Equation. A 338 element mesh for one-pitch screw section (Figure 7) was generated using the simulation software Patran-Mechanical Computer Aided Engineering. The analysis presented good qualitative agreement with the experimental melt pressure results obtained in the test rig. However, the magnitude of the melt pressures were higher than the experimental ones. The three-dimensional F.E. mesh of the screw section, Figure 7, represents a film of polymer melt around the screw showing its pressure contours. A positive pressure is built up in the screw channel increasing from the contour level E to B (the plot legends are in Pa). It reaches its maximum at the leading edge of the screw flight, contour A, and then it starts to decrease from B to F. At F a negative pressure contour appears in the trailing edge of the flight. Two regions of high melt pressure were found in the flight.
Therefore the F.E. mesh analysis showed three clearance pressure regions separated by two high pressure concentrations. The pressures generated in the model in the region around the point of minimum clearance were the highest and this would be expected to be the main source of restoring force. However, melt instabilities were observed in some of the experiments using LLDPE, these being of the form of pressure oscillations from sub-zero to zero values. In these cases the screw-flight minimum clearances coincided with the position at which the oscillating pressures were observed. Thus the theoretical high melt pressure found at the minimum clearance point in the screw model does not seem to occur practically. The shear rates generally found at the minimum clearance are very high, varying from 5,000 to 120.000 s-1. At this level of shear rate melt instabilities such as melt fracture or wall slip may happen. Hence very low pressures or around zero at the point of minimum clearance may be expected due to the melt slip-stick phenomenon .
The regions of high melt pressure, shown in Figure 7, could act as a cushion for the screw support, the screw moving away from the low pressure zone towards to the top of the barrel. The pressure distribution around the screw (model) is not symmetrical and suggests that the minimum clearance would be located in the third quadrant of the barrel. This could explain why the minimum clearance is often found at the top-right part of the barrel (third quadrant) in the experimental results.
|Fig 9: Polymer melt pressure vs. screw position, experimental results: a) maximum load point (150°) and b) minimum load point (330°).|
The theoretical pressure profiles are very similar to the experimental ones, as shown in Figures 8 and 9. These pressure profiles were reproduced by reading the pressure contours that distinguished the three pressure regions in the F.E. mesh. Three circumferential cross sections were taken along the screw: at one-third, half and two-thirds pitch. The one-third and two-thirds pitch measurements, reproduced in Figure 8, had the same shape of curve as the experimental results generated on the test rig, Figure 9. The pressure measurement at one-half pitch screw is not possible to obtain in the current test rig.
The polymer melt flow in the screw is very complex. This is probably true at the entrance and exit from both the experimental and the model screws. The boundary conditions on the model may have tended to distort the results. A three pitch screw model could give, at the central pitch, pressure contours which are less susceptible to the effects of the boundary conditions imposed. Further, the test rig screw was also short and so it would seem likely that, experimentally, the behaviour observed was also influenced by the effects at the screw ends.
|© AIPnD , created by NDT.net|||Home| |Top||