You are currently viewing: Articles


Enhancing heat transfer in Texas Towers

Retrofit technology enhances heat recovery in feed-effluent exchangers, increases throughput, reduces furnace load and provides cost benefits

Peter Ellerby and Peter Drögemüller, Cal Gavin Ltd
Edgar Vazquez-Ramirez and graham Polley, University of Guanajuato
sebastian Erlenkämper, BP
Viewed : 8505
Article Summary
Texas Towers are vertical shell and tube heat exchangers, used as heat recovery units and placed around reactors in petro-chemical plants. Since the stream being heated forms the feed to the reactor, from which the hot stream flows, these units are also referred to as feed/effluent exchangers.

In this article, we discuss the design of feed/effluent exchangers, how exchanger design affects operating costs, and how effective retrofit can lead to both significant energy savings and improvement in throughput. A section of flow sheet for a typical unit, together with operating temperatures, is shown in Figure 1.

A feedstock that is liquid in ambient conditions is to be vapourised, heated to a high temperature and then reacted with a gas in the presence of a catalyst. The reaction is exothermic. The reaction does not take place in the absence of a catalyst. It is therefore common practice to mix gas and liquid feedstocks prior to heating. This action has the benefit of reducing the temperature at which the liquid vapourises, thereby giving favourable temperature driving force effects.

Since heat recovery is important to plant economics, the products of the reaction are used to preheat the reactants. The two-phase feed enters the tubes at the base of the exchanger. Within the tubes, the liquid reactants are vapourised, then the gases are superheated before leaving the tower to enter a furnace. Within the furnace, the reactor feed is further heated before it enters a catalytic reactor.

The product stream leaving the reactor consists of a mixture of unreacted feed gases and desired products. This is fed to the shell side of the Texas Tower, where it enters at the top and flows downwards. As heat is transferred to the feed stream, the product stream is first desuperheated and then partially condensed. It is then fed to a series of coolers before the desired liquid products are separated from the unreacted gases.

Most catalytic reactors are subject to decay of the catalyst. A common means of maintaining the production rate as the catalyst decays is to increase the temperature of the reaction. Eventually, the operation will reach either the maximum reactor temperature or the maximum firing capacity of the fired heater and the plant will have to be shut down in order to renew or regenerate the catalyst. The load on the fired 
heater increases across the operating period.

Economic assessment of exchanger operation
The obvious way in which the feed/effluent exchanger contributes to the economics of plant operation is by reducing energy consumption. This benefit is easy to quantify. However, while the saving in energy costs accumulated over the plant’s operating period is substantial, 
this is not the only benefit. For a 
fired heater of fixed capacity, the higher the level of heat recovery 
in the feed/effluent exchanger, the longer the plant’s operating 
period. This has two major cost advantages: the plant has a 
higher capacity; and the plant produces more products for a given batch of catalyst.

The economic benefit derived 
from improving feed/effluent exchanger design is significant. 
The benefits of revamping existing units include energy saving, 
reduction in catalyst regeneration 
cost and improvements in plant throughput in terms of increased production rate and extended operating period.

Identifying energy benefits
The thermodynamics of heat recovery in a feed/effluent exchanger can be determined by superimposing curves showing the heat demand (reactor feed) and the heat release (reactor effluent) for the case described above as functions of temperature (see Figure 2).

The curves have been superimposed at a minimum temperature approach of 50°C. The distance the demand curve (the lower curve) extends beyond the release curve at the hot end of the scale shows the demand placed upon the fired heater. In this example, the reactor feed is heated to a temperature of 300°C, the feed has to enter the reactor at a temperature of 390°C and the load on the fired heater is 6.7 MW.

The total demand of the feed stream is 27.5 MW and consists of 11.7 MW associated with vapourisation of the feed stream and 15.8 MW for superheating the stream to 390°C. Given that 6.7 MW is provided by the fired heater, heat recovery provides 20.8 MW.

The heat released by the reactor product as it is cooled to 130°C is 
24.0 MW. This figure comprises 17.8 MW for desuperheating the gas 
and 6.2 MW associated with condensation of product. The amount of heat being absorbed in the cooler positioned after the feed/effluent exchanger is 3.2 MW. Hence, the 6.2 MW associated with condensation is divided between the feed/effluent exchanger (3 MW) and the cooler 
(3.2 MW).

Problems in Texas Tower design
In the example outlined, we see that in terms of heat transfer mechanisms the feed/effluent exchanger can be divided into three sections. At the 
base of the tower (cumulative heat load 0–3 MW), condensation (in the presence of non-condensables) is occurring on the shell side with evaporation on the tube side. In the mid-section of the tower (cumulative load 3–11.7 MW), desuperheating is occurring on the shell side with evaporation on the tube side. At the top of the tower, desuperheating is occurring on the shell side with superheating on the tube side.

Heat transfer to and from gas streams dominates the problem. However, the mix of heat transfer mechanisms coupled with the desire to achieve full recovery in a single exchanger leads to a number of problems. The first major problem is obtaining a uniform flow distribution on the tube side of the exchanger. The feed stream entering the unit is a two-phase mixture. The tube bundle presents numerous flow paths. In these circumstances, the two phases will distribute so that the overall pressure drop across the unit is minimised.

The easiest way of ensuring a reasonable phase distribution is to set the tube count so that the pressure gradient associated with a well-mixed two-phase stream is less than that associated with the hydrostatic head of the liquid phase alone. Under 
these circumstances, the liquid can only be transported up the tubes as part of a two-phase mixture. However, this can result in another danger: 
the return of liquid down the tubes 
at the bundle periphery. So, as a 
safety measure, phase distributors, such as perforated plates, are 
often installed in the exchanger headers to ensure gas is present below these tubes.
Current Rating :  4

Add your rating:

Your rate: 1 2 3 4 5