3 replies to this topic

[z_kapetaki]

hello!!!
I am doing a feasibility study and I have collected the cost functions for all the process units. My question is that in many cost functions I noticed that there are exponential terms. Where do these exponents come from? Is it maybe from the geometry of the unit? I would appreciate if someone could resolve this (a bit naive maybe, I have to admit) query. Thank you very much!!!

[kkala]

I am doing a feasibility study and I have collected the cost functions for all the process units. My question is that in many cost functions I noticed that there are exponential terms. Where do these exponents come from? Is it maybe from the geometry of the unit?

I guess that mentioned "cost function" is a capital cost estimate for a unit, containing an exponential term. The exponential term refers to the size / capacity of the unit. Exponent is often (but not always) close to 0.6, hence the "six tenth factor rule" for rough capital estimates.
Perry's "Chemical Engineers Handbook", Process Economics, Fixed capital cost estimation (Table 9-48 in 7th ed, 1997) reports such exponents for several "Processing Plants" with some explanations. For instance, you may want to estimate capital cost of 100000 ton/y of an acetaldehyde (CH3CHO) plant from ethylene (CH2CH2), knowing that this is 13.7 MM$ for a 50000 ton/y similar plant. Wanted value would be 13.7x(100000/50000)^0.70=13.7x1.62=22.2 MM$ (valid for M&S cost index=1000, probably around 1990, as noted in Perry's table). The cost of 22 MM$ should be updated to current prices (cost indexes could be found in Chemical Engineering Magazine).
Values in Perry are assumed for new (Greenfield) Plants (not clearly stated), lower cost is needed for capacity increase of an existing Plant (Brownfield Plants).
Similar rule also exists for equipment capital cost (Table 9-50 of Perry, 7th ed).

[Art Montemayor]

z_kapetaki:

To answer your specific question as to where the exponents from Cost prediction equations come from, let me tell you.

Years ago, chemical engineers found that as they plotted total capital costs against capacity of the equipment in question, almost linear curves were generated on a log plot. When a mathematical regression was made on this data, the "famous" 0.6 exponent was discovered as a factor in most curves. In other words, the result is one of empirical data that has been regressed - as is often the case in most empirical equations.

[breizh]

Hi ,

Let you try this book :


http://books.google....imating&f=false

Hope this helps

Breizh