Report to Olive’s Paper Mill plc.

S F Bush

Summary

The investigation has examined (a) the process operation now and as far as can be established on the night of the failure, (b) the current condition of the machinery relevant to the operation of the roll and (c) the condition of the rubber roll cover itself so far as can be conjectured from the evidence of fragments and calculation.

The broad conclusion is that based on (a), (b) and (c), the most likely cause of the roll cover failure was fatigue growth of cracks on the inside of the cover, leading to sudden disintegration.

Report for Richard Grant Accessories Ltd.

S F Bush

Summary

A large proportion, in the order of 90%, of front bumpers made from one lot (about 15 tons) of ABS sheet supplied by Telcon, have been found to fail by cracking in a particular region of the bumper. The vast majority (95%+) of the failed bumpers were found to have failed after they were installed on the vehicles, but before they entered service (i.e. at the pre-delivery checks). The ABS sheets supplied by Telcon were made in turn from Montsanto Lustran QE 1110, and were thermoformed into bumpers by Richard Grant during May and June 1980. This particular batch was the second in a sequence of three using Telcon sheet, the first batch being made in late 1979. Reported failures from batches one and three were very few and generally arose for a variety of reasons not connected with the cracking problem.

The investigation reported below is one of several commissioned by the interested parties to identify reasons for the cracking. Fundamentally the issue is to decide between the ABS granules supplied by Montsanto, the ABS sheet as delivered by Telcon, the processing into bumpers by Richard Grant, the installation by Skoda, and a combination of two or more of those. In arriving at a decision on this it is important to recognise that a fault in one area can show up as a fault in another. Thus processing can significantly alter the microstructure of the ABS material; conversely subtle unperceived changes in the ABS can cause difficulties in the processing.

Taken with the earlier investigations made by others, the new results reported below represent about as comprehensive a range of tests as can now be done, two years after the event. It is believed that the results obtained by all investigators are fundamentally consistent. The broad conclusion of this report is that the ABS material in the batch in question was significantly more crack-prone than that in the first batch.

Report to Petrocarbon Developments Ltd.

S F Bush

Objective

Overall, this is to summarise the likely behaviour of the existing reactor and feed systems and the scope for improvement.

A preliminary report was tendered on 21st April and this report confirms and extends its conclusions.

See also the section on Applications to Existing Products & Processes.

Paper to the Mixing and Polymer Processing Conference of the European Federation of Chemical Engineering, Delft, Holland

S F Bush with E Jongen.

Introduction

A major problem in the manufacture of a number of important polymers such as polyethylene, polyester and nylon, is the occurrence of non uniformities which show up as streaks and blobs in film and cause breaks in filament. The major source of this problem lies in the fact that during manufacture a tiny fraction of the material is exposed to reaction conditions very significantly different from those applying to the bulk of the material, by virtue of the exceptional residence times developed at vessel walls. Such non uniformities are particularly likely to occur at cooling surfaces since the unavoidable tendency to stick there is enhanced by the increase in viscosity through the thermal boundary layer.

The present paper outlines a description of the reaction and flow of polyethylene and ethylene through tubes from which there is a substantial extraction of heat. The driving force is provided by pressure drop.

In the high pressure stirred autoclave process, the reaction in the tube is incidental to its main function of cooling the polymer and reducing pressure (by about 400 atmospheres) at the outlet from the autoclave. In the tubular process, the overall pressure drop through a tube is greater (around 2000 atmospheres), since the tube is itself the main reactor. The chemical kinetics and fluid mechanics equations have been set up to cover both cases, but the applications referred to in this paper arise from the autoclave process.

See also the section on Applications to Existing Products and Processes.

Systems Technology Group Technical Note, ICI Europa.

S F Bush with S Baillie-Strong.

Summary and Conclusions

1. Major variations in texturised yarn dye-uptake were observed which correlate strongly with time at spinning and clearly moved in sympathy with pack pressure and Draw Tension Integrator date at spinning.
2. The most likely cause of these variations was differences in moisture content of the polymer fed to the driers causing hydrolysis and molecular weight changes in the spun yarn.
3. These variations caused about 20% of the total package-to-package variability with shifts of almost 2 dye-shades in mean level in about 36 hours. These magnitudes are for Turquoise CG dye at pressure – a typical continental customer requirement. The recommended quality control dyestuff, Navy D2G, probably underestimates these variations.
4. The cause of gross variations in polymer moisture content should be eliminated. However, since the polymer transport system (Tote Bins) cannot be hermetically sealed and since the chip moisture content is below atmospheric humidity, some variations will always be present. A drying control strategy to deal with these variations with the aid of a drying model should be developed.
5. The same magnitudes of changes in initial water content and yarn i.v. have been obtained by other investigators on 167 f30 at various times from January 1974 to the present.
6. The calculated i.v. to water content sensitivities agree closely with measured values.

Process Technology Dept Technical Note, ICI Europa

S F Bush

Summary

The adaptation of the fibre drawing model to the spin draw model requires an estimate of the heat transfer from the individual filaments in the quench system following the spinnerette. An estimate is made of the growth of the boundary layer surrounding each filament in a typical 44 filament bundle, and the corresponding heat transfer coefficients.

See also the section on Applications to Existing Products and Processes.

Paper A7 to the Institution of Mathematics and its Applications Symposium on Partial Differential Equations and distributed Parameter Control Systems, University of Warwick, England, 13th-16th July 1971.

S F Bush with P Sinclair.

Introduction

By far the most general technical problem in the operation of an industrial-scale chemical process is the optimisation of performance by selecting appropriate steady-state values of the manipulable variables. This is thus more truly a design than a control problem. An exception to this generalisation is provided by the class of vapour-phase hydrocarbon oxidative processes of which combustion and chlorination systems are notable sub-divisions. Here there are often genuine problems of (a) moving between steady-states without entering regions in which explosions or excessive degradation of product can occur, and (b) determining steady-states which are sufficiently stable in some sense.

A feature of these systems is that the reactions are exothermic and carried out in combustion chambers or reactors designed to permit or induce an internal feedback of material by means of which hot reacted gas is mixed with cool feed gas, diluting it and bringing it rapidly to reaction temperature. In the system referred to here, chlorine and hydrocarbon are premixed and enter the reactor in an entraining jet which induces the recirculation and which is also allowed to impinge on the walls of the reactor.

Group II report, ICI Central Instrument Research Lab

S F Bush

Summary

Sudden cessation of reaction has been observed in reactors in which fresh feed gas is brought to reaction temperature by the entrainment of hotter, partially reacted gas. Observations have been made on the plant-scale and pilot-scale versions of these reactors. On the pilot-scale, the instability was induced by small measurable disturbances in the inlet flows. A mathematical analysis of the dynamics of the reacting system is used to predict the range of conditions for which steady reaction is possible, the magnitude of the disturbances necessary to extinguish steady reaction once established, and the influence of geometrical factors on the stability of the reaction system. Within the limits of the analysis, the predictions are in reasonable agreement with experimental results on both the plant-scale and the pilot-scale reactors. The results of the study have been used to modify the internal design of the plant reactors in order to obtain greater stability of reaction. So far as is known the occurrence of unexplained temperature collapses has been largely, if not completely, eliminated.