Choosing the right heat exhanger

Many types of heat exchanger are now available. Various types of compact heat exchanger have been developed, ranging from plate-and-frame units to exchangers using 'micro-channels'.

Even when it comes to the traditional shell-and-tube exchanger the engineer is faced with a range of options - baffle arrangements range from various segmental designs to helical baffles and twisted tube bundles, enhancement techniques include low finned tubes and tube inserts. This diversity presents the engineer with a problem. When should the technology that is to be used in a given design be identified? And how?

The driving factor in exchanger selection is likely to be cost. So, let's start by considering the costs associated with a heat exchanger. These are: initial purchase cost; installation cost; running costs; and final disposal cost.

In many people's minds the purchase cost dominates. However, installation cost can be between three and five times the purchase cost. Since installation cost is dependant upon size and weight, quite large differences in purchase cost can be compensated for by savings in installation cost. Basing selection on purchase cost alone is plainly wrong.

In heat recovery systems the purpose of the heat exchanger is the reduction in overall plant operating costs. As will be shown below, economic heat recovery levels are dependant upon the exchanger technology used. Improved energy efficiency can result in cost savings that pay for the exchanger tens of times over.

In <a href='http://www.e4engineering.com/content_images/table1.gif'>Table 1</a> we compare the economics of heat recovery for a standard process integration problem. The importance of considering exchanger selection at the process design stage becomes clear.

Now let's consider the stages through which an engineering project passes. The first stage could be called process design. This is the conceptual stage of a project and it leads to the formulation of the process flowsheet. This stage usually involves the work of a relatively small number of process engineers. It is at this stage that the heat recovery level for the plant is set and the plant layout sketched out.

The next stage of the design can be called the 'detailed design stage'. This sees the introduction of a wide range of engineering disciplines, all working from the base set down during the process design. Given the number of engineers now working on the project, it becomes difficult to make major changes to the process design. Later on the design moves on to equipment procurement and plant construction.

It is clear from this outline that the selection of the exchanger technology must be considered at the very outset of the project. It is part of process design. To leave selection to the 'procurement stage' in the false sense of providing a 'level playing field' is to loose 90 per cent of the benefit of using an alternative technology.

Moving on now to the selection process itself, the first step should be to identify and reject unsuitable types. This involves an assessment of the suitability of the technology for a given process application and can be done quite quickly.

The first step here is to compare required operating conditions (temperature and pressure) with the limitations of the equipment. This is followed with a brief hazard assessment (for example, is leakage likely to result in hazard, is the equipment type prone to leakage during operation or maintenance?). The next stage is a consideration of fouling (does it occur, what is its nature, how is the exchanger to be cleaned?). Many exchanger types cannot handle suspended solids or fibres. Some can only be cleaned chemically. Finally, consider whether or not the exchangers need particularly skilled maintenance staff and the availability of such staff.

Having identified which technologies are suitable, the next stage is to determine the effect of the technology on size, cost and complexity of the exchanger. Fortunately, this can be undertaken quite rapidly for it does not need detailed design of the exchanger. The importance of this stage can be gauged by considering the two designs shown in <a href='http://www.e4engineering.com/content_images/fig1.gif'>Figure 1 </a>and <a href='http://www.e4engineering.com/content_images/fig2.gif'>Figure 2 </a>.

In Figure 1 we see the shell-and-tube exchanger arrangement required for a heat recovery/trim cooling unit on a styrene plant. The heat recovery section requires four individual exchangers arranged in series. The trim cooler requires two separate exchangers. In Figure 2 we see the plate-and-frame option. One single exchanger (used in a multi-stream configuration) is used.

The complexity of the shell-and-tube system arises from the need to use multiple shells in order to avoid temperature crosses. The calculation of the number of shells needed only requires knowledge of inlet and outlet temperatures. The simplicity of the plate-and-frame solution arises from the use of pure counter-flow.

Most significantly, the determination of the number of shells needed for the duty uses algorithms already embedded in standard process integration software.

We can identify complexity easily, but what about size? Well, rapid sizing algorithms are now available. Exchanger size is dominated by available pressure drop. We can relate pressure drop with velocity. We can relate heat transfer coefficient with velocity. Finally, we can relate velocity with some basic features of exchanger geometry (for example, tube diameter plus exchanger area). By combining these relationships we obtain one that relates pressure drop, heat transfer coefficient, mass throughput and exchanger area:

delta P = (k/M) Aa^m The constant, k, is solely dependant upon physical properties and a geometry factor (such as tube diameter in the case of shell-and-tube performance). The exponent, m, is dependant upon how both friction factor and heat transfer coefficient relate to velocity. Equations of this form can be derived for tube-side performance (including tubes fitted with inserts), shell-side performance (for a range of baffle types), plate-and-frame performance, plate-fin performance, spiral exchangers, and micro-channel exchangers.

The size of an exchanger can rapidly be computed by solving the hot-side pressure drop relationship with that for the cold side and the basic exchanger design equation simultaneously.

This algorithm has already been incorporated into the process integration program offered by ESDU International. The current version of the program incorporates shell-and-tube exchanger relationships. Future versions will also handle the alternative technologies.

So, we can easily handle complexity and size. What about cost information? ESDU has been working closely with exchanger manufacturers for a number of years. One area of collaboration has been the development of costing algorithms. Technologies covered include shell-and-tube exchangers, plate-and-frame exchangers, plate-fin exchangers and compact exchangers that use micro-channels. Work aimed at incorporating this information into process integration software is underway.

In summary then, if the full benefits of exchanger technology are to be realised selection must be undertaken at the earliest stage of a project. The means of doing this are clear and the tools needed are being developed.

A past president of the Heat Transfer Society, Graham Polley is editor of Pinchtechnology.com, part of the Environmental Technology Network (www.envirotechnet.com), sponsored by ESDU International