[D199]Design Of Heat Exchanger

Heat exchangers are the unsung heroes of many industrial processes and as such they tend to be taken for granted - nobody likes paying for what is often seen to be unnecessary maintenance. Heat exchangers provide duty for so long, that when they start to drop in efficiency, it's usually a gradual process that goes largely unnoticed - until their performance has deteriorated sufficiently to be a problem. Then it really is a problem - and one requiring urgent attention.

What aggravates the situation is the heat exchanger that has never been cleaned properly, coupled with the commercial need to keep it on-line. When the decision is made to carry out cleaning, often nobody knows what the performance of the exchanger is meant to be, either because the drawings have been lost, or no record of any improvement was made after the original cleaning.

When the exchanger finally is opened up to ascertain the extent of the fouling, it's not surprising to find it is so severe that cleaning takes a lot longer than planned. Any benefit that might have been gained by a quick traditional clean is offset by the extended cleaning duration and costs - and, of course, lost production.

If that sounds like a nightmare scenario, bear in mind that this is the sort of situation specialist cleaning companies encounter every week. Cleaning is often carried out without any firm knowledge of how much of an improvement the cleaning will give and how long its effects will last. Having to make 'finger in the wind' predictions clearly is not a satisfactory way to plan maintenance.

One of the most popular and widely-employed heat exchanger configurations in industry, is the straight or hairpin shell-and-tube exchanger. With hundreds or thousands of small-bore tubes bundled together, the extent of quite modest scaling can involve major work to return the exchanger to anything near its commissioned performance. If the outside of the bundle is heavily scaled as well, the cleaning challenge rises by an order of magnitude.

There is potential to bring about a significant improvement in heat exchanger accessibility and 'cleanability', by working more closely with the people who design heat exchangers and fabricate industrial plants.

Better design would lead to improved cleaning - where improved means faster, cleaner and safer, possibly in-situ or even on-line and with better waste containment. It would then be easier and quicker to clean exchangers back to bare metal to return them to duty and their design performance faster.

Plants are generally specified and ordered on the basis of throughput, not accessibility and ease-of-cleaning. Suppliers are happy to comply with this and therefore tend to design heat exchangers with 30-40% excess capacity to ensure that they can continue to provide duty, even when quite extensively fouled. Heat exchangers the world over are currently designed and installed with a view to using one of three systems for cleaning: chemical, pressure jetting and/or mechanical and this approach has remained unchanged for over 50 years.

When it comes to maintenance, refineries - like most of industry - tend to compete on the same basis - a 21-day shutdown is decreed because it's been done that way for maybe the last 20 years. The same cleaning methods are generally used slavishly, with high-pressure water as the cleaning medium.

Most companies look at their heat exchangers in isolation and simply try to extend their run-time, instead of having them designed or re-designed so they can be cleaned more regularly, but faster and better. BP's Coryton refinery, for instance, managed to reduce cleaning time on one shell-and-tube heat exchanger from three days to three hours by applying a different approach to cleaning it.

If a plant is optimised for cleaning, almost full production can be maintained throughout the cleaning process. Relatively minor mechanical changes, such as adding isolating valves to heat exchangers, means that each exchanger, or bank of exchangers, can be taken down and cleaned while the others remain on-line. A redesign of the exchanger so that a header can be removed, means it can then be cleaned with a different system to the standard high-pressure water jetting, in a few hours instead of several days.

At Dow Corning's silicone plant in Barry, south Wales, a tubular boiler and fire tube in the Energy Recovery Unit (ERU) required the removal of a 5mm layer of deposit in as short a time as possible to minimise lost production. Another obstacle was that the unit, which carries waste gases, takes 48 hours to cool and prepare - even with the introduction of a chilled nitrogen purge - before personnel can enter to clean it manually.

The solution involved developing a bespoke remote de-scaler, which was inserted through a small 50cm man-way. Once inside, the de-scaler expanded to fit the hot fire tube, while reaching the full length of the carbon steel tube. With cooling time and man entry eliminated, the shutdown was reduced from five days to three and there was a noticeable improvement in performance of the ERU when it came back on line.

Improved cleaning cycles also mean the rate of future fouling build-up is reduced, which in turn reduces the risk of tubes corroding as a result of the exchanger being open to the atmosphere longer for cleaning.

Heat exchange surfaces therefore remain smoother and provide better heat transfer. If and when the exchanger does foul up, it's easier to clean next time around, using whichever system is preferred. This would represent a change of practice to what has been the norm since the 1980s, for instance, when what was then Mobil in the UK was one of the first refineries to decide that it would extend run-times by abandoning the annual clean and only clean every two years.

Today, typical service intervals have become stretched to three and even four years in some cases, but the apparent operational savings are actually a false economy. Shareholders are indeed happy, because they are getting longer run times, while competing refineries have little choice but to play the same game or lose millions during more frequent shutdowns. Four years down the line, however, the plant will have to come down for major cleaning and maintenance and it will experience a far higher capital replacement cost than ever before.

Article Source : Pg. 5