Related ContentDistillation columns play a key role at many plants and also often account for a significant portion of a site’s operating costs. So, vendors and research groups are expending considerable effort to enhance tower operation.
For instance, work underway aims to provide early warning about flooding (Figure 1). This afflicts numerous columns — stopping them from separating feed mixture into purified distillate and bottom products, and thus limiting capacity and prompting process upsets. Detecting incipient flooding can be tricky but a new approach promises to significantly help.
It relies on a modified Rosemount 3051S, a scalable device that typically handles integrated pressure, flow and level measurements around the plant. Roger K. Pihlaja, principal engineer – process diagnostics for the Rosemount Division of Emerson Process Management, Chanhassen, Minn., and A. Frank Seibert, technical director of the Separations Research Program (SRP) at the University of Texas Austin, Texas, (and liquid extraction guru for our online “Ask the Experts” feature) first revealed details about the development at the October 2008 Emerson Global Users Exchange in Washington, D.C.
Figure 1. Column flooding: Viewport shows flooding
Pihlaja and Seibert have added a noise detector to the 3051S. “Bubbles are a white noise generator and as there is a detector built into the 3051S it means we have a new technique to detect something that is intrinsically present during flooding. As far as I know it’s the first time that this kind of signal processing has been built into an industrial device,” Pihlaja explains.
It should be possible to develop active incipient flooding control in a way that would give a small capacity increase on any column, the pair contends. This is because the new technology allows the column to operate in the non-linear region that currently is considered too unstable to consider (Figure 2).
Trials took place in 2008 at the SRP’s pilot facilities in Austin, which have a number of distillation and extraction columns suitable for studying all aspects of the processes involved — from control to the hydraulic characteristics of random and structured packings. The runs used 100% reflux; using less reflux currently isn’t feasible at SRP, notes Pihlaja. “Obviously this is a problem in a university setting because with an 18-in diameter column you are going to get a lot of product that will need recycling. They would need to install equipment to deal with this in order to run at finite reflux ratios.”
Talks are underway with chemical companies and refiners about trials of the method at their sites. “Obviously it would be extremely useful to demonstrate this in an industrial setting. The downside is that commercial columns are not usually as instrumented as the one at UT so it’s a little harder to do. Of course a commercial column is making thousands of dollars a minute, too, so companies are extremely reluctant to mess with them. The economic downturn hasn’t helped either,” notes Pihlaja.
Figure 2. Pushing the bounds: Safely operating closer
In the meantime, a U.S. patent application has been filed on the modified 3051S. “Overall, our work demonstrated proof of concept: we can detect flooding based on a change in the physics inside a column,” he says.
The method also might serve to detect fouling of packing. “With a properly instrumented column, it should be possible to determine that flooding is occurring at a lower vapor/liquid loading in a fouled column than would be expected with a clean column. If you have a map of expected column performance [as in Figure 2], then, in principle, by calculating how much your fouled column deviates from clean performance, it should be possible to estimate the ‘degree of fouling’ and report this parameter as some sort of ‘column health’,” asserts Pihlaja.
Pressure drop (ΔP) across the column is the key issue today for members of Fractionation Research, Inc., (FRI), Stillwater, Okla., believes Mike Resetarits, its technical director. The non-profit research consortium’s membership includes top chemical and petrochemical companies, as well as vendors. (Frank Rukovena of FRI handles distillation questions for Chemical Processing's “Ask the Experts” feature.)
Resetarits likens checking ΔP to a doctor taking a person’s blood pressure: “When I give a distillation column a ‘physical,’ I start with the ΔP across the tower. Operators need to monitor ΔP hourly/daily. Increases can mean foaming or fouling. Decreases can mean acid attack on the trays/packing, or — maybe — the reflux rate isn’t what was planned.”
FRI is carrying out a number of research programs. Resetarits points in particular to a collaboration with researchers in the School of Chemical Engineering at Oklahoma State University (OSU), Stillwater, Okla. Under the supervision of instructor Rob Whiteley, graduate students there are investigating a novel approach to real-time identification of system hydraulic limit. It’s based on the premise that high-speed (10+ Hz) pressure or differential-pressure measurements can characterize the change in hydraulic state of a distillation column as it approaches the hydraulic limit. (The proposed mechanism is unrelated to the macro liquid accumulation phenomenon associated with monitoring column ΔP to detect flooding.)
Traditional linear signal-processing techniques can detect variations in the high-speed pressure signatures as the system approaches flood. Preliminary analysis of data collected at the FRI test facility has been encouraging.
If successful, the technology would increase the capacity of existing distillation equipment with negligible capital investment. However, further work is being deferred until other higher priority collaborative projects are finished, notes Resetarits.
Meanwhile, he offers a warning about plugging. “Approximately half of the world’s distillation columns are packed. It is impossible to climb through such towers. The key to their performance is the distributors, which are easily plugged with dirt, polymer, gums and other materials. During annual turnarounds, all packed towers should be opened and the distributors should be inspected for hole plugging. Better to do these inspections than restart such towers and have them malperform six months later because the half-plugged distributors became fully plugged. Strainers are good, here, too.”
Azeotropes often can pose challenges for distillation — and modeling can provide important insights, says Vikas Dhole, vice president of engineering and product management at AspenTech, Burlington, Mass. He points to a recent project on a Reliance Industries’ plant in Mumbai, India. The challenge was to troubleshoot and debottleneck the acetic-acid heterogeneous azeotropic dehydration column. AspenPlus software simulated para-xylene removal.
“Users often have entrainer losses between columns. In this case, simulation helped to reduce losses by optimizing the packing in the column bed to increase capacity,” adds senior product manager Dave Tremblay. The result is a 0.4 kg/metric ton/yr reduction in entrainer consumption, equivalent to a $550,000/yr savings.
In distillation, capacity and energy consumption are intertwined. “As you squeeze pump-around, vapor flows within the column are reduced, as is energy consumption… There is not only an energy benefit, but a capital benefit, too,” says Dhole.
A project at the China National Offshore Oil Corp.’s refinery in Guangdong, China, illustrates this. Here, the refiner wanted to optimize energy consumption during the design phase to achieve a best-in-class energy intensity index (EII). AspenTech used a combination of pinch and column analysis, together with a total site analysis that considered inter-unit integration. The result was a final design EII of 65.3 compared to the general contractor’s 71. This improvement equals $16 million/yr less in energy costs.
At the Samsung-Total aromatic complex in Chungnam, Korea, process modeling, coupled with pinch, column, distillation-sequence and thermal-integration analyses, improved product purity as well as slashed 20% off the complex’s energy costs, saving $12 million/yr.
“Overall, distillation is a key element for us and we are investing heavily in it,” notes Dhole. In February, the company launched release 7.1 of its apenONE engineering software.
AspenTech certainly isn’t alone in updating its software tools. For instance, Koch-Glitsch, Wichita, Kan., just introduced version 4.0 of its KG-Tower equipment rating program. This assists in specifying mass-transfer equipment, including conventional and high-performance trays and random and structured packings. The latest release provides, e.g., new sieve-tray design/rating capability, an updated tray model and the ability to rate additional packing products.
Hardware, not software, is Koch-Glitsch’s main focus, though — and the company just debuted Intalox Ultra random packing. It boasts low pressure drop/high capacity, high efficiency and high strength-to-weight ratio. For a new installation, this translates into smaller column diameter or less column height. On revamps, benefits depend on the application and include more capacity at current product purity, less energy consumption per unit of product, higher purity at current product rates and lower pressure drop.
Cutting the size of columns assumes more importance — both in terms of cost and fabrication — as the definition of “worldscale” rises to 1,500 metric ton/yr and beyond, notes Greg Wisniewski, product line manager for UOP Process, Des Plaines, Ill.
“Our technology brings a very large benefit in being able to reduce the size of a distillation column significantly. This also allows for situations where it isn’t physically possible to transport or install conventional equipment onto a site,” he notes.
Its latest innovation, SimulFlow, provides a 50%–60% improvement on the company’s normal tray products. “SimulFlow was developed when we felt we were reaching the limits of capacity improvements that could be made with existing tray configurations.”
SimulFlow is designed to handle simultaneous liquid and vapor flow — in other words cocurrent flow. In trials it has achieved what the company describes as extraordinary capacity. It also reportedly provides an excellent height equivalent to a theoretical plate (HETP), a low pressure drop and a greater than 2:1 operating range.
“For the future, we are working on increasing market penetration and improving the technology still further,” adds Wisniewski.
One development of particular note from UOP’s point of view is on the propylene side, which in the past typically has relied on heat pump systems. “However, many greenfield [grassroots] plants now have a very high level of cooling water available, which is driving interest in high pressure systems rather than heat pump systems. This is good for us because with a heat pump the requirement for internals is really low, for example 180–200 distillation trays within the column, because it has a lower operating pressure. Operating at a higher pressure is more difficult and requires more trays, up to 50% more. Here our ability to minimize column height becomes very important.”
Another area UOP sees potential is in improved column control — especially now that it’s fully owned by Honeywell and can take advantage of Honeywell’s process control expertise.
Seán Ottewell is Chemical Processing’s editor at large. You can e-mail him at firstname.lastname@example.org.