Manufacture of Epichlorhydrin: Assessment of the prospects for improvement in the Bydgoszcz 200 unit
April 30th, 1981
Report to Petrocarbon Developments Ltd.
S F Bush
Summary
Following massive yield losses (around 50%) on the operation of the 200 unit in July 1980 (run 2) various changes to the operation of the reactor (R201) and reactor-stripper (K202) were recommended by UCC. Most have been implemented and subsequent runs (4-11) have shown clear-cut improvement to between 80 and possibly 89% but still well short of the design (flowsheet) performance of 94%.
The object of this assessment is to see in what ways the various changes made have contributed to the improvements recorded so far and thus what further practicable changes can be made to reach, if possible, design performance. Attention has been focussed on the K202 unit because (a) much the biggest yield loss was recorded there and (b) this unit’s behaviour is much the most complex of any in the plant.
As suspected early on, the behaviour of K202 depends sensitively on the numbers. There is additionally particular uncertainty about the key physical chemistry parameters, which dominate the predicted performance of the unit. General remarks as to trends are thus of little use in this case, in assessing the ultimate performance.
The results of the analysis have therefore been assembled in a way which permits the consequences of both operational changes and any later reappraisal of parameters (which may of course include the real effects of hardware changes). Nonetheless, the results of the analysis and the parameter values chosen on a best information basis, appear to show why the plant has performed as it has, and also ties up with the UCC computer simulation, and the published Italian pilot plant data. It also indicates why a propylene oxide process would be expected to perform better on this design of hardware (cf. UCC practice).