Optimization of PVC-free materials in cables

5. Processability

5.1 Background
5.2 Strategy
5.3 Initial Rheological Characterization
5.4 Introduction to Computer Simulation
5.5 Mixing Tests
5.5.1 Background
5.5.2 Mixing Experiment
5.6 Computer Simulation in Practice
5.6.1 Rheology - Testing and Modelling
5.6.2 Process Simulation
5.6.3 Test Runs on Extruder
5.6.4 Results and Comments to the Tests
5.6.5 Conclusion on Processing Tests
5.7 Production of Test cables
5.7.1 Optical Test Cables
5.7.1.1 Discussion of Results
5.7.2 Test Production of Power Cable Jacket
5.7.2.1 Extrusion Test Results
5.7.2.2 Discussion of Results
5.7.2.3 Second Extrusion Trial
5.7.2.4 Discussion of Results of the Second Extrusion Trial.

5.1 Background

The HFFR compounds are more difficult to process than conventional polymer materials like plasticized PVC, PE etc. This is due to two main reasons. Firstly the HFFR compounds contain high loadings (up to approx. 65%) of inorganic fillers in order to achieve good flame retarding properties. This means that only a limited part of the compound is polymeric and for this reason the melt viscosity gets very high. Consequently, large shear forces are developed and high frictional energy is created in the extruder during processing giving rise to undesired temperature rises in the compound. The higher the extrusion speed, the higher the heat build up during the extrusion process.

If the temperature gets too high the fillers (hydroxides) liberate water (crystal water). This has three main effects:

1. Development of bubbles in the product and the surface gets rough.
2. The flame retarding effect of the fillers is reduced.
3. The mechanical properties cannot fulfil the requirements.

Thus it is important to keep the temperature as low as possible during processing. For this reason the polymer part of the compounds has been selected with as low crystalline melting points as possible with due respect to other essential features like mechanical and thermal deformation properties, compatibility with filler, chemical stability etc.

Other problems are obtaining a good mixing with colour masterbatch and often also the build-up of large deposits at the die opening.

The main challenge is to find ways to increase the line speed of the extrusion processes to "PVC" like level, maintaining good quality of the extrudate and without overheating the compound.

The challenge can be dealt with in different ways. One is to select the materials with the best processing properties i.e. the lowest melt viscosity. The lowest viscosity is, however, frequently related to the lowest flame retardancy due to low filler concentration.

Another way is to try to optimize the machinery (extruder screw, distributor and die) in such a way that temperature is kept low also at high line speeds.

Normally, the best way forward will be a combination of the two, selecting as "friendly" a compound as possible, still with adequate flame resistance (and other properties, see section 4) and also optimize the processing equipment which normally consists of standard parts. These parts have typically been optimized for general purpose materials like PVC and PE which very seldom present processing problems of the type mentioned.

Due to the many variables involved in the extrusion process it is extremely difficult to foresee the effect of a given adjustment i.e. will the effect of increased temperature settings on the extruder be a higher or a lower melt temperature? The obvious answer is "higher" but in fact it depends on a lot of things like the relative amounts of frictional and conductive heating, the temperature dependence of melt viscosity etc.

Due to the high complexity of the problem we have chosen a new solution method i.e. computer simulation of the extruder process making it possible to perform test runs on a computer for a short time instead of the time and money consuming trial and error testing in the factory.

After having selected the final compounds based on the material test programme the idea is to perform computer simulations with these materials in order to design an optimal extruder screw, distributor and die system for a 63 mm conventional extruder.

Comparative test runs are then to be made with new and conventional equipment.

5.3 Initial Rheological Characterization

Before any attempts to perform computer simulations it is necessary to determine the rheological properties of the individual compounds.

The initial rheological characterization of the compounds was made on a Brabender Extrusiograph measuring extruder. For every compound the recommended processing temperature was chosen for the measurement. The molten compound was extruded through a die of well defined geometry, readings of output and corresponding pressure drop were made and converted to shear viscosity vs. shear rate flow curves.

From the curves the viscosity values have been taken at a shear rate of 100s^-1. The results are shown in table 3.

The measurements only give a rough comparative classification of the processability since the measurements were made at one temperature only and at a rather low and limited shear rate range.

Closer rheological studies were made with the three selected compounds nos. 6, 11 and 14. The simulations for these are described in chapter 5.6.

Table 3.

5.4 Introduction to Computer Simulation

Three commercially available software packages for the simulation of the extrusion process have been evaluated and tested at the supplier’s work stations. The one giving the best data agreement with the values measured on the Brabender Extrusiograph was chosen. It was the package from Polydynamics Inc., Canada and Compuplast, Czech Republic consisting of the Polycad programme for general 2-dimensional flow analysis, the Extrucad programme for simulating the extrusion process, and two accessory programmes for viscosity data processing and storage, Viscofit and Polybank.

A computer simulation using these programmes proceeds as follows:

1. Measurement of viscosity data i.e. Rabinowitch corrected shear viscosity and elongational viscosity as a function of shear rate and temperature.
2. Fitting the data to viscosity model (Newtonian, Power Law, Carreau etc.) using Viscofit.
3. Polycad simulation of flow in the die and the distributor, primarily to find the pressure necessary for a given extruder output. This simulation is based on the iterative solution of balance equations for heat, mass and momentum and apart from the flow properties it is necessary to know melt density, heat capacity and thermal conductivity. Further the geometry of the die and the distributor, temperature settings and melt temperature must be given.
4. Extrucad simulation of the extrusion process using the above mentioned exit pressure, screw RPM, temperature settings and extruder geometry as input parameters. Also the properties of the solid compound must be given i.e. the crystalline melting point, the heat capacity, the bulk and the mass density, the heat of melting and also friction coefficients of granules against the metal surfaces of the extruder.
The result of the simulation includes all relevant pressures, temperatures, stresses, rates, degree of melting, and heat and mass flows as a function of position in the extruder.
5. Normally the extruder output at the selected screw RPM and the assumed output through the die will not be in agreement. This must be accounted for by adjusting the RPM until the agreement of flow is achieved. Also adjustment of the input melt temperature in the die must agree with the exit temperature of the extruder. After a few adjustments of the data and new simulations the agreement will normally be reached.
6. The data are then evaluated and for the actual compounds the most important item was to make sure that the compound was not overheated and that the extruder motor was not overloaded.

5.5.1 Background

The normal method for colouration and UV stabilization of polymer compounds is mixing with concentrates, masterbatches. In order to get a good distribution of pigment or stabilizer the viscosity of the masterbatches is normally very low. For the very viscous compounds involved in this project this gave problems as a very highly viscous melt does not mix well with a nearly liquid one (viscosity mismatch). Also in the cross linking of HFFR compounds where an optimum mixing is essential this point is relevant. One way of solving this without having to use especially developed concentrates is to use mixing elements in the extruder screw for improving the distributive mixing of colour in the HFFR compound. In the following section a trial series of such elements is described.

5.5.2 Mixing Experiment

On the Brabender Extrusiograph a 19 mm screw was designed in such a way that that the last 3L/D of the 25L/D screw could be exchanged. Three elements were tested, first the normal metering zone without mixing elements and then two types of mixing element.

Two of the compounds were tested, one with a rather low viscosity, 80465 and one with a higher viscosity, compound no. 6.

The results were very clear. The normal screw gave a poor mixing with distinct concentric rings of colour in the extruded round rod whereas the mixing units of both types gave better results with rings that were scarcely noticeable under microscope at 100x magnification of a thin section in transmitted light,

see appendix 7. No significant difference between the two tested compounds was seen.

5.6.1 Rheology - Testing and Modelling

For the computer simulations the rheological characterization was made on a Rosand Precision Capillary Rheometer on the compounds nos. 6, 11 and 14. Measurements were made a two temperatures (160-180°C) and a range of shear rates between 20 and 20000s^-1.

For the shear viscosity the Carreau model was chosen in order to take the lower Newtonian flow region (which is neglected by the Power Law) into account in the simulations. The elongational viscosity was modelled by the Power Law. For all three materials the flow properties are given by the following equations:

Shear viscosity as a function of temperature and shear rate: ^:

(1)

Elongational viscosity as as function of elongational rate: :

.                                       (2)

K, b, C, a, n, E and n_e are constants determined by the Viscofit programme for each of the compounds. T_r is an arbitrary reference temperature.

Table 4

The determined values of the constants are given in table 4 together with the other necessary properties.

5.6.2 Process Simulation

The extrusion tests were carried out on a 63 mm extruder with an L/D ratio of 22. It was originally equipped with standard PE and PVC screws and a head with distributor and die optimized for standard materials like PE and PVC. From the very start it was evident that optimization of the processing of the selected HFFR materials meant changes in the extruder geometry.

First the die was considered. Simulations showed that with a conventional die a given output of material no. 14 could be obtained if the pressure at the die entrance was approx. 70 MPa. After a few attempts of adjusting the die geometry a die with the same hole geometry was constructed that could transport the same amount of compound at a pressure of 24 MPa. At the same time the maximum temperatures were decreased due to the changes which were primarily opening of flow channel giving lower shear heating. The same procedure was followed for the distributor resulting in a pressure drop of 4 MPa for this part and a total of 28 MPa at the exit of the extruder.

This final pressure was used in the simulations of the extrusion process. Simulations were made with a lot of screw designs and finally (after approximately 100 simulations) the design of a new screw was decided. This screw gave an acceptably low melt temperature, total melting already at 14D from the entrance to the extruder and a torque manageable for the extruder at the desired output rate which was 2-3 times higher than what could be obtained with the normal screws.

5.6.3 Test Runs on Extruder

These simulations formed the basis for the construction of a new screw, called the "Extrucad" screw in the following. Among other features defined through the simulation the new screw contains a mixing zone with a length of 3D at the end of the screw.

Comparative test runs were then carried out with five materials and the three available screws in order to see the results of the simulation procedure in practice. The results are shown on the graphs in appendix 8 and 9.

The five compounds tested were, apart from nos. 6, 11 and 14, two easily processed types, i.e. 80465 and a new material with low viscosity designated material X.

The graphs show the output vs screw RPM, Specific Energy Consumption vs output, and torque vs output for the 5 materials with standard screw, PVC screw and Extrucad screw.

Output vs screw RPM

The results are presented in appendices 8.1 to 8.2. The general trend is that for all the materials the output is highest for the Extrucad screw, about 2-3 times higher than with the traditional screws. Some measuring points are marked with a "B". This indicates that water bubbles have been created due to overheating of the material.

For the low viscosity compound X, the improved output levels off at high rpm due to increased pressure back flow which is not observed with the high viscosity compounds.

For compound nos. 6 and 14 the increase in maximum output is approx. 200% and for no. 11 the improvement is a little lower as this compound in fact gives a better output on the older screws. For no. 14 it was necessary to adjust the temperature settings downwards to avoid formation of water bubbles.

On the graphs for compounds nos. 6 and 11, points calculated by the simulation programme have been included. For no. 6 the agreement is excellent.

For compound no. 11 the agreement is only fair for simulation using the normal viscosity input. For this compound the elongational viscosity determined by the fitting program (E in table 4) is extremely high and may be due to an error. If we neglect the contribution of elongational viscosity the agreement is much better (curve marked sim.-el.visc.).

Specific Energy Consumption (SEC) vs Output

The results are presented in appendices 9.1 to 9.2. In the test runs the SEC was calculated from power consumption and in the simulations the shear stress at the barrel wall calculated by the Extrucad programme was used.

Also for the SEC there is a general trend showing that the specific energy needed to process the compunds is significantly lower with the Extrucad screw than with the standard screw and the PVC screw, typically 2-3 times lower.

Again there is a little different behaviour with compound X for which SEC increases dramatically at high output. The reason is the same as mentioned above, the high pressure flow more or less equals the drag flow and increased rpm does not increase the output very much. The energy consumption increases but not the output. For the other four compounds the observation of low SEC is consistent also at high output rates.

On the graph for 80465, which was one of the materials tested in project 1, we have indicated an SEC measurement from then, obtained on a 30 mm extruder equipped with barrier screw. The SEC was about 3 times higher than on the 63 mm extruder and concerning this point the Extrucad screw was clearly the best of the 3 screws.
In the graphs for nos. 6, 11 and 14 simulated values of SEC are included showing very good agreement with the real values measured in practice.

A comparison with the measured SEC of a PVC compound using the PVC screw shows that the HFFR processing with the new screw is in fact requiring less energy than the PVC on conventional machinery at high output and an equal amount at low output, see fig. 9.

Torque vs. Output

The torques involved in the processes were calculated for the materials nos. 6, 11 and 14 with the Extrucad screw. The results are shown in fig. 10 in comparison with PVC processed with a PVC screw. Obviously the HFFR compounds, and especially no. 14, requires a high torque compared to PVC. In fact for compound no. 14 it was the maximum torque, 2200 Nm that was the rate limiting factor, see fig 10.

Evaluation of the computer simulation method

All the computer simulations performed which could be compared to real measurements have given good agreement with these, typically within 10-20% which is the best one could hope for. Selected typical results for compound nos. 6, 11 and 14 comparing simulated and real values are presented in table 5.

Table 5.

5.6.5 Conclusion on Processing Tests

The computer simulation method has turned out to give very good agreement between simulated and real values obtained on the 63 mm extruder.

A new extruder screw and die system has been designed and produced based on the computer simulation. It has given excellent results in the test runs i.e. the output has increased and the specific energy consumption and maximum melt temperatures have decreased compared to normal screws.

Introduction of the new screw in the extruder will increase the productivity at least with a factor 2.

Single fibre cable

Using production parameters defined during the simulation, the single fibre optical cables (fig. 11) were produced at a speed of 120 m/min. The visual appearance of the cables was OK, but compound no. 14 had a little roughness in the surface as opposed to nos. 6 and 11 which were both very smooth.

The cables were coloured with an orange masterbatch, PE based, and the degree of mixing was tested under microscope and found OK.

The cables were produced with cold cooling water. This is not according to recommendations and in order to be sure that the mechanical properties were OK, tensile tests with samples of cable sheath were carried out.

Fig. 11 Single fibre cable.

All samples had an elongation at break above 150% thereby fulfilling normal specifications for this property.

The three cable samples were tested according to IEC 332-1-C and IEC 332-3-C and they all passed the tests with a good safety margin.

Light Duty Cables

The larger optical light duty cables consist of a flammable core containing 12 fibres with a tight polyester jacket (fig. 1). Earlier IEC 332-3-C tests with 80465 as sheathing material were unsuccessful and to qualify as a standard sheathing material for optical cables the pass of this test is considered necessary.

The cables were produced using roughly the same conditions as before but of course a larger die. The extrusion speed was 50-60 m/min and again the quality of the extrudates was good.

The elongation at break again OK.

The IEC 332-3-C was again passed by all three materials, no. 11 with a large margin but nos. 6 and 14 with a narrow margin only. Addition of, say, 4% masterbatch containing flammable material might mean that the test will not be passed by a coloured sample.

The smoke density test according to IEC 1034 was passed by all 3 samples.

5.7.1.1 Discussion of Results

The test cables were produced without problems at an increased line speed compared to normal. The visual appearance of the cables was OK. Experiences with production of long lengths are not yet available so possible problems with die build-up etc. may occur.

The mechanical properties of the extruded sheaths were OK and so the burning properties.

The improvements compared to the results of project 1 are in our opinion primarily due to the modified screw and die design, but also improvements in the processability of the compounds may play an important role. For the test cables, however, we concentrated on compounds that were not the easiest to process.

5.7.2 Test Production of Power Cable Jacket

Due to the results obtained when testing the new screw design in a 63 mm extruder, it was decided to continue the project in order to scale up the screw design to a 150 mm, 18 L/D extruder.

Extruders with internal diameter in the range 120 to 1810 mm and length between 18 L/D to 24 L/D are typically used in the production of PVC cable sheating for 500 volts to 1000 volts with 3 to 7 cores with a conductor cross section from 1.5 mm² to 16 mm².

According to information given by commercial producers of Halogen Free Flame Retarded compounds, shorter extruders (18 L/D) are preferred for the production of cable jacket.

As input data used for the computer simulation programme during the design of the new 150 mm 18 L/D screw, the required output for two different 750 volts cables were used, i.e. 3 x 1.5 mm² and 5 x 6 mm² with the goal of reach in the required output for maximum line speed in the range of 55 to 65 rpm on the screw, with 80 rpm being the maximum.
Further the design of the used extruder head and used tooling was taken into account during the computer simulation.

The previously measured flow data and compound characteristics for compound no. 11 were used during the computer simulation.

Compound no. 11

The temperature profile recommended by the compound supplier was used in the first extrusion trial.

Without the extruder head the required output for maximum line speed was reached at 15 rpm at the screw.

At 60 rpm an output of approximately 2250 kg/hour was reached, which is far too high - four times higher than required.

After mounting the extruder head it was impossible to increase the screw rpm higher than 4 - 5. This was due to overload of the extruder. The max. amper consumption for this extruder motor is 350 Amp.

With screw rpm between 4 and 5 the output is approx. 150 kg/hour.

The mass temperature of the compound was already in the second zone of the extruder above 190° C causing thermal breakdown of the compound.

Attempts to correct/reduce the mass temperature by altering the temperature settings of the extruder and screw were not successful.

5.7.2.2 Discussion of Results

It is our opinion at least three factors contributed to the failure of the attempt to scale up the screw design from the successful trials on a 63 mm extruder to a 150 mm internal diameter extruder:

1) A limitation in the computer software to deal with screw dimensions above approx. 115 mm may have been the prime reason for the lack of success.
2) Further lack of cooling efficiency of the extruder barrel.
3) Too narrow flow channels in the head caused a too high back-pressure build-up.

It is a requirement for future screw design experiments to reduce the compound output per screw rpm, in order to reduce the build-up of frictional heat and back pressure in the extruders when extruding HFFR compounds with fillers of the same type as used in compound no. 11.

5.7.2.3 Second Extrusion Trial

Due to the discouraging results of the testing of the computer designed screw described in 5.7.2.1 an older 120 mm 20 L/D single flight screw was re-designed to fit to a 120 mm 24 L/D extruder, i.e. prolonged with a new tip with a small increase of the screw core diameter (4 mm).

To use a single flight screw is very often recommended for the extrusion of Halogen Free Flame Retarded compounds by commercial producers, especially for compounds based upon the same type of filler as compound no. 11.

The temperature setting on the extruder line was in accordance with the recommendation given by the compound manufacturer. During the trial we increased the temperature 10° C in order to reduce the friction heat developed in the feeding zone of the extruder. This was done in an attempt to reduce the ampere consumption of the extruder motor as it was very close to the maximum limit for the extruder motor, and this was due to a very high compound mass temperature.

When the mass temperature of the extruded compound exceeds approx.

175 - 180° C crystal water will begin to evaporate. The generated steam will cause formation of bubbles in the compound causing a rough surface and a dramatic reduction of the mechanical properties.

When the screw revolution exceeded 10 rpm equal to an output of 80 kg/h the measured compound mass temperature was found to be 172° C and amper econsumption 290.

A tendency to formation of air bubbles was seen.

The maximum screw revolution for this extruder is 60 rpm and the maximum ampere consumption of the motor is 317 ampere.

When increasing the screw revolution to 20 rpm the compound mass temperature increased to approx. 200 - 205° C and severe formation of air bubbles was seen.

Attempts to reduce the temperature increase in the compound caused by friction heat by adjusting the temperature setting on the different parts of the extruder were not succesful.

When producing cables with PVC-compounds the output required at normal extrusion line speed for this extruder line will be approx. 500 kg/h (corrected for differences in density between PVC- and HFFR-compounds).

5.7.2.4 Discussion of Results of the Second Extrusion Trial

With the single flight screw we achieved an output of approx. 80 kg/h of compound no. 11 without too severe formation of air-bubbles caused by evaporation of "crystal water" due to a too high and uncontrollable increase of compound mass temperature (heat generated by friction).

Further it was only possible to remove the screw from the extruder at the end of the trial using force.

Unfortunately the achieved output is only 15 % of the output when producing cables with standard PVC-compounds.

In order to be able to optimize the screw design for a larger diameter extruder (the design of flow cannels in the extruder head has also to be re-considered) for cable production with Halogen Free Flame Retarded compound based upon the same type of filler as compound no. 11, a number of further design - and production tests have to be carried out.