(133g) Debottlenecking of Heat Exchanger Networks Using Optimium Pressure Drops

Process Integration technology is now widely applied in grass-roots design, energy saving retrofit and debottlenecking of Heat Exchanger Networks. This technology has been used in a variety of industries and approved to be reliable and applicable in engineering design.

Debottlenecking may apply to a specific part or entire unit, whether it is due to increased throughput or process modifications.

One of the advanced methods for debottlenecking, which is currently used, is based upon fixed allowable pressure drops, through which a retrofit can be achieved without a need for pump and/or compressor replacement. However, this method does not consider optimum pressure drops for the process streams, and hence suffers from the following drawbacks.

1.Capital investment for new exchangers would not be optimum;

2.Operating cost (i.e., Electricity) of the network tends to be sub-optimum;

3.Wrong targeting results causes wrong design initialisation and hence wrong network topology.

This research is trying to resolve the above problems, by development of a new procedure for pressure drop optimisation in debottlenecking. This procedure enables the designer to study pump and/or compressor replacement whilst at the same time optimizing additional area and operating cost of the network.

It is clear that during pressure drop optimisation, the stream pressure drops should be addressed as key variables and optimised according to unit specifications. The more stream pressure drops, the more turbulency and so the higher film heat transfer coefficients. When heat transfer coefficients increase, the required heat exchanger area installed over that stream reduces, and therefore the purchasing capital cost of new exchangers reduces. But, on the other hand, new pumps and/or compressors investment and running cost increase due to higher pressure drops. Hence, there would be a trade-off between the capital cost associated with the new heat exchanger surface areas and capital cost plus additional power cost associated with pumps and compressors.

The new procedure has been effectively applied to a Crude Oil Pre-heat Train, which was subject to some 20% increase in throughput, and the corresponding results proved to be accurate enough. Namely, the results of targeting stage showed a very good agreement with network synthesis and final detailed design.