“Calculational Prowess” of TK Solver™
A Perfect Fit for Chemical Process Engineering

 
DuPont uses TK Solver to design entire chemical process plants. Brian Hanley finds TK Solver eminently suitable for a particular but massively important part of the chemical process industry: sizing distillation towers.
 
The design of distillation towers drives many other areas of chemical processing and being able to optimize the designs easily is “worth major money to petrochemical companies”, says Hanley, who has a Ph. D. in chemical engineering from the University of Minnesota and 17 years’ experience in the industry. Up to 70% of both the capital and operating costs of the chemical process industry are accounted for by separations processes.
 
For example, being able to accurately determine the right amount of pressure required within a column to separate styrene can be tricky, “but that’s where the real savings occur,” says Hanley. He frequently performs these types of calculations in TK Solver and says that a 3% error could easily add up to $20M a year in lost product and wasted energy.
 
“Creating the models in TK Solver and performing the calculations all in one place, reduces the chance of error and gives me confidence in the numbers,” says Hanley.
 
Hanley credits Professor Michael Doherty—a well-known and highly respected academic in the area of distillation, for sparking his interest in the subject. He describes chemical process engineering as “pretty exact in theory” but says that putting it into practice is quite different. And therein, lies the challenge!
 
Distillation tower design centers on thermodynamic and hydraulic analysis. Ordinarily, Hanley explains, a thermodynamic simulation is done to determine the difficulty of the given separation. Then, a hydraulic simulation is run in order to determine the appropriate column diameter and height (towers can range up to 40 feet in diameter and 120 feet high), and the proper design for the column internals.
 
The equipment customer, on one of several commercially available process simulation packages, typically completes the thermodynamic simulation. The equipment vendor based on the flows and physical properties established by the thermodynamic simulation usually performs the hydraulic calculations. The hydraulic calculations involve proprietary correlations for acceptable column operating rates and real-world column separation efficiencies.
 
Design optimizations involve considerable back-and-forth between the two parties and are frequently incomplete because the parties are often unwilling to share all the details. As a result, chemical companies worldwide wind up spending significantly more on capital, operating, and waste management costs than they might otherwise need to because many of their distillation column designs are not optimal. Hanley says he can get the required calculations done in a “reasonable” amount of time using TK Solver. But because the overall process is a cooperative effort, it’s quite often the source of time delays. Such interchanges can add days or even weeks to the process but Hanley says that if he can get the basic information up front and limit the need for interaction, a two-man week project can easily be reduced to a two-man hour project.
 
“With TK Solver, the productivity enhancements are incredible,” says Hanley.
 
For Hanley and the companies he has worked for, savings in the sizing of distillation towers have been achieved because of a particular way of going about the design process: putting mathematical formulas for thermodynamic simulation and hydraulic simulation together in an environment where they can influence each other, and where any variable or group of variables can be tweaked to see the effect on any other variables. For Brian Hanley, that environment is TK Solver.
 
“I have found the calculational prowess and ease of use of TK Solver ideally suited to my needs,” Hanley says.
 
TK Solver’s features and functionalities permit him to explore design alternatives “rapidly but as thoroughly as is practicable given limited time or manpower,” Hanley says. It also permits solving a design problem outside “the solution schemes envisioned with other process simulators” and, if need be, “to extend the code beyond the options available in other simulation packages.”
 
Hanley says he is making extensive use of the new Solution Optimizer included with the Premium version of TK Solver 5.0. He is “letting it choose things like packing sizes, column diameters and heights, or process flows based on criteria that we establish in conjunction with our customers.” These criteria might include minimizing annual operating expenses, minimizing product loss, maximizing column throughput, maximizing product purity, or meeting existing physical limitations.
 
Packing, by the way, is a combination of materials that allows for the intimate contact of liquid and vapor inside a distillation column.
 
In the analysis of a “typical” binary distillation column, there are four coupled algebraic equations for the overall mass balance, the overall component balance, and the component balances above and below the feed point to the column. Hanley explains: “These equations are statement of the law of conservation of mass: what goes into the column (or column section) has to come out of the column (or column section) somewhere.”
 
Next there are two coupled integral equations for the number of “theoretical stages” – one for above the feed location and one for below. These equations embody the feed and product purities and parameters associated with the vapor/liquid equilibrium and the aforementioned material balances.
 
There is an equation for the top and bottom sections of the column that dictates how much vapor and liquid traffic can be handled by a given column diameter and packing style.
 
“All of these coupled equations must be supplemented with additional equations to predict the physical properties of the liquid and vapor phases,” Hanley says. These include liquid densities, viscosities, and surface tensions, along with vapor densities, viscosities, and diffusion coefficients.
 
The TK Solver models Hanley works with include about 25 variables. How many calculations? “No idea,” he says. “Thousands.”
 
With TK Solver, there is no way to know the number of calculations—and no reason to know. With its rule-based declarative structure and ability to “backsolve”, you put in the formulas and input values, and TK Solver does the rest. With its list-solving and table-lookup properties, TK Solver can work easily and conveniently with massive amounts of data. The software also has a wide-ranging library of preprogrammed models and data—including the complete National Institute of Standards and Technology table of fluid thermodynamic and transport properties, which Hanley uses extensively in his work.
 
Why not use a spreadsheet? “I use Excel because I can buy add-ins for it that allow me to extend the range of calculations I can do with TK Solver,” Hanley says. “I use Excel, but not because it’s a good tool.” A user must, in effect, tell a spreadsheet everything. In contrast, TK Solver, with its declarative structure, in effect—learns as it goes.
 
How much time does TK Solver save Brian Hanley? A single design takes about two minutes and an optimized design, about two hours. How much is the “major money” he spoke of? One of his early models saved his former employer Koch-Glitsch of Wichita, KS, $1M a year in improved column performance.
 
“Overall, the problem involves a relatively large set of coupled algebraic and integral equations, some of which exhibit singularities,” Hanley says. “In spite of all the underlying complexity and ‘bad’ behavior inherent in these equations, we have found that TK Solver handles their simultaneous solution with ease.”
 
Hanley is currently taking a “much needed” break from his accomplished career and says he would welcome the opportunity to one day, teach. Hanley jokes that earning his Ph. D. in chemical engineering was more about work avoidance than academic ambition, adding that he “drifted into” chemical engineering as an undergraduate in the hopes that it paid well!
 
“I guess you do what you’re good at,” Hanley laughs. What does he like to do in his spare time? Hanley enjoys reading, is a regionally ranked competitive fencer, and has recently taken up target shooting.