Increasing energy costs make heat integration more attractive. However, heat integration can complicate control (see: "Control Challenges Can Pinch Energy Savings") and can significantly expand the ramifications of equipment failure. To illustrate the latter point, we'll examine the impact of where product is withdrawn from a stripping column on an upstream reactor train.
High-temperature reactors commonly include feed/effluent exchangers for heat recovery. Typically, after this exchange the feed gets raised to the final desired temperature by a high-temperature utility, often a fired heater, before entering the reactor. Figure 1 shows a simplified system for a hydrotreating reactor where the reactor effluent also provides heat to reboilers for two downstream sidestrippers.
For illustration purposes, the two sidestrippers differ in configuration.
The first sidestripper (upper column) has a conventional arrangement. Product comes from the bottom of the column. The bottoms liquid splits between the reboiler feed and the feed to the bottoms product pump. The unit shown uses a thermosiphon reboiler; a reboiler relying on a pumped feed doesn't alter the analysis.
The second sidestripper (lower column) has an alternative configuration where the reboiler feed only consists of the column bottoms. Liquid product comes from a sidedraw above the reboiler feed. This is an unusual arrangement — in most units, the lowest liquid product typically exits from the bottom of a column, not a sidedraw.
Both configurations meet the fundamental requirements of the process. Each sends an identical liquid to its product pump and its reboiler. The difference between the configurations only becomes apparent when considering failure modes.
Let's examine what happens if the liquid-level control valve fails. In the upper column, the control valve's failing open leads to an immediate loss of liquid in the column. Due to the short residence time in this unit, in a few minutes at most, no liquid is available for the reboiler. At this point, the reboiler no longer acts as a heat sink on the reactor effluent.
By itself, this isn't necessarily a major problem. With six downstream feed/effluent exchangers, the temperature-profile change may be relatively modest. Nevertheless, it can't be ignored.
One common practice in complex heat-integration systems is to vary the materials of construction as process temperatures change. In some cases, this primarily is a cost-saving step. With some allowance for materials' life and corrosion, less expensive materials often handle colder services. In other cases, the situation is more difficult. Conflicting requirements of low-temperature versus high-temperature corrosion mechanisms may make materials' alterations mandatory as temperature profiles change in heat integration systems.
Fortunately, in this case, it's possible to select suitable, albeit expensive, materials that cover the entire operating range.
The lower sidestripper configuration is a step toward alleviating problems from temperature profile changes caused by control valve failure. The product control valve can't remove liquid below the sidedraw nozzle, which is some distance up the tower. So, a liquid inventory remains — giving operators some time to respond if the level control valve fails. Because liquid from the trays falls into the bottom of the tower, the reboiler won't be starved of liquid if the level control valve fails open.
Alternative designs could meet the same objectives. For instance, another option is to use a baffle inside the tower to create a preferential flow path through the reboilers at all times.
While the reboiler and product configuration of the lower stripper is odd, it does have a valid purpose — to meet the requirements of a specific equipment failure with minimum safety and operating problems.
Always investigate and understand the effect of unusual events. Unit configurations that may act identically in normal operation may differ in costs and benefits in unusual situations.
ANDREW SLOLEY is a Chemical Processing contributing editor. You can e-mail him at ASloley@putman.net