Steve’s material balance worksheet stretched for two dozen feet and was four feet tall. He looked like an Afghan rug maker as he crouched over it, crawling on his knees to mark it up with colored pencils. No wonder he had a heart attack! However, this was how balances were made forty years ago at International Steel Services, Inc. (ISSI) and many other firms. His problems soon became mine in 1998 when I joined ISSI, which designs chemical plants for the mining industry.
It didn’t take me long to see why Steve’s balances never balanced. I spent the next two years performing regression analyses of physical properties, e.g., to find how the density of ferrous chloride solutions varies with temperature. I put in a fair share of lab hours measuring viscosities, heat capacities, looking at the effect of trace components on sample properties, and comparing field data against the regression curves. Were my numbers much better than Steve’s? Probably not — but they were an improvement.
The first rule of material and energy balances is: consistency trumps accuracy. In one of Steve’s columns, the density didn’t vary with temperature. If the ferrous chloride concentration at 90°C is 20% with 2% hydrochloric acid, using the same density at 30°C and ignoring the acid gives an about 0.5% error. This is less than the accuracy of most flow transmitters, which is about 1%. However, the error matters because ferrous chloride is a feed stream and errors in feed streams compound later if there’s a chemical reaction.
That leads, by the way, to the second, third and fourth rules of material balances: get the feed streams right; don’t strive for better accuracy than you can measure because you can’t confirm it; and compare your balance against fresh field data. Field data markedly disagreeing with predictions from the balance may force reassessment of the balance. However, if the data agree, you’ve got a good defense for your balance if it’s ever questioned.
Some problems with balances occur because the equipment sizing software isn’t up to the task — hence rule number five: don’t blindly accept software results. Use caution, as my fellow columnist Andrew Sloley points out in "Software: Show Some Skepticism." For example, if a vapor phase exists at a particular temperature and pressure for an alkylation product stream but the process is all liquid, then either the fluid package or composition is wrong.
The appeal — and power — of such software, of course, is that it can quickly generate physical properties that would take hours to estimate using handbooks. However, if a composition is simple, it’s possible to compare a property from the National Institute of Science and Technology’s Chemistry WebBook, http://webbook.nist.gov/chemistry/. Unfortunately, there’s no way to readily confirm these properties for complex streams.
If you don’t have access to software, you can try your hand at developing it. At ISSI, I wrote my own sizing software for packed towers — it eliminated many discrepancies among different plant designs. Before relying on it, though, I asked three random-packing vendors to run their equipment-sizing packages against mine as a check; it was okay. That’s rule number six: approach the balance from another angle by using other software to confirm your results.
Believe it, or not, using spreadsheets for balances isn’t dead. I’ve written extensive and intricate spreadsheets to calculate the temperature profiles in bricked fluidized beds, hydrolysis beds and polymer reactors, and phase equilibrium and reaction balances. This brings up rule number seven: use field data to get best results from spreadsheets.
One design, a contact absorber with several side-draws, prompted me to ask for field data on a simple contact absorber. These were used to model the more-complex absorber. Ultimately, field data will determine the validity of that model. Use those data not only to verify the model matches current performance, but also as a back-check against the original material balance to see if operation has significantly degraded.
This brings me to rule number eight: never trust rules of thumb. At ISSI, we used a simple ratio of HCl absorption based on a 1950s acid-absorber article. I later back-checked this ratio during commissioning and found it only approximately correct. Generally, rules of thumb based on research are useful for new processes but don’t rely on them without understanding.
DIRK WILLARD is a Chemical Processing Contributing Editor. You can e-mail him at firstname.lastname@example.org.