In Publication No 1 of the Nordic Concrete Federation 1990 "The chloride 
diffusion characteristics of concrete - approximative determination by linear 
regression analysis"by Ervin Pousen, he pointed out that the approximation 

erf(z) = (1-z/3^0.50)^2 

is valid when z is between 0 and root 3.

This approach is discussed in my book on corrosion of steel in concrete, in 
the Germann Instruments profile grinding kit manual and is developed into a 
working model along with a carbonation initiation model and a time to 
cracking model in the UK Department of Trade and Industry Corrosion of steel 
in concrete research programme which I worked on with the UK BRE & Trend 
2000.  The reports will be published by BRE later this year.

John P. Broomfield D. Phil (Oxon.), EurIng, FIM, FICorr, MCS
[log in to unmask]  
Consulting Corrosion Engineer
78 Durham Road
London SW20 0TL, UK
Tel + 44 (0)20 8944 6161
Fax + 44 (0)20 8944 6300
Mobile + 44 (0)77 11 00 61 97
http://members.aol.com/JohnPBroomfield/index.htm

Read the Book
Corrosion of Steel in Concrete: Understanding, Investigation and Repair 
Publ. E & FN Spon, ISBN 0-419-19630-7 £65.00 (pounds sterling)
Tel + 44 (0) 1264 343071 (UK)
Fax +44 (0) 1264 343005 
http://www.efnspon.com
E-BOOK SOON TO BE AVAILABLE FROM
www.eBookstore.tandf.co.uk


In a message dated 10/22/01 10:13:39 AM GMT Daylight Time, 
[log in to unmask] writes:

<< I am wondering whether someone knows the a polynomial equivalent (i.e a 
polynomial equation) to the
 error function (erf) used in the diffusion equation. Many thanks. Jamal
  >>