Capital Project Management Support

Tom Wolf, P.E.                    twolf5165@aol.com                                    (281)565-4038


 

Capital Cost Estimates - Understanding, Analyzing, and Evaluating

A book for the person requiring more than just an answer.
 
Refining and petrochemical project management personnel, most without estimating background, are required to understand, analyze, and determine validity of complex cost estimates prepared by others. 

This book is specifically designed and written as a desktop resource. In addition to presenting an organized approach to analyzing cost estimates, results of decades long historical cost data is provided, and  needed assistance to answer the crucial question of “is this estimate reasonable?” is presented.

List Price: $56.00

AVAILABLE at Amazon.com and other fine retailers.
 
ISBN-13:
       978-1461180739
ISBN-10: 
       1461180732
 
8" x 10", 232 pages

Lang Factor Cost Estimates

 

T. E. Wolf, P. E., Sugar Land, Texas

 

(Please note this is copyrighted material, it is provided for your information.  Please feel free to use it, but please don't abuse the copyright.  Thank you.)     

     

 

Overview. 

 

Lang Factors were first developed and presented seventy years ago as a simplified method to obtain a quick estimate of Total Installed Cost for industrial projects.   These factors are still referenced and utilized today.  Are Lang Factor’s still valid? 

 

Lang and Lang Factors.  


After the end of World War II, Hans J. Lang, an engineer for Day & Zimmermann, Inc. in Philadelphia, PA, wrote a series of articles3, 4, 5 suggesting a simplified method for obtaining a quick estimate of approximate cost of process plants.  He went on to elaborate that the simplified method was not intended to replace preliminary estimates but should be used to provide a judgmental cost more reliable than a mere guess.

 

The method presented by Mr. Lang was based on a study of fourteen (14) plant estimates of various sizes and types, ranging from $100,000 to $15,000,000 (late 1940’s USD).  He identified three process plant types, solid processing plants, solid-fluid processing plants, and fluid processing plants.  He defined these as follows: “a distillation unit would be classified as a typical fluid process plant, a coal briquetting plant as a typical solid processing plant, and a solvent extraction plant complete with bean preparation and meal processing facilities as a typical solid-fluid processing plant.” 

 

The method presented required an estimate of the plant equipment delivered to the site and one of three factors (f) applied to this delivered-to-site equipment (Equip) cost.  The three factors for each type of plant being:

 

Solid process plants                3.10

Solid-fluid process plants        3.63

Fluid process plants                4.74

 

The Total Installed Cost (TOT) estimate was then obtained by multiplying the appropriate factor by the equipment cost, or

 

TOT = Equip * f

                       

The overall factor estimate calculated included not only the delivered process equipment cost, but material and labor costs for site improvements, foundations, steel, buildings, piping, electrical, controls, both design and construction costs, and overhead costs.  Overhead cost included insurance, taxes, contingency, field and office expense, temporary construction facilities, and contractor fee.  In other words, a quick approximation of a typical contractor prepared Total Installed Cost (TOT) estimate, however excluding Owner Costs such as land, operating cost, etc.

 

Changes since 1940’s.

 

Over the ensuing decades, several people have elaborated on the Lang factor either directly or indirectly by providing information allowing a Lang factor to be calculated.  Several of these studies and results are contained in the Fixed Capital Cost Estimation chapter of Perry’s handbook7; others include books by Gerrard2 and Page6, and a Dysert1 article among others.

 

However, during that period, many things have changed; do these factors still apply?  There are now governmental rules and regulations in-place, which just did not exist in the 1940s and 1950s.  There are materials and construction methods that are different.  There are digital process controls instead of pneumatic controls.  The computer is used in lieu of the slide rule and there is three dimensional computer design.  Then there is material and labor cost inflation (escalation) over the many decades.  That is just to mention a few.  All of these are significant changes, but the one most quantifiable is the cost escalation. 

 

Effects of Sixty Years Escalation

             

As defined, the Lang factor is a ratio of Total Installed Cost (TOT) to Delivered-to-Site Equipment Cost (Equip).  One could assume that by dividing one project cost component (TOT) by that same project’s cost component (Equip), the resulting ratio would be free of the cost escalation effects and thus could be compared with similar ratios of other projects executed at other periods.  Further, since the Total Installed Cost is made up of both material and labor cost subcomponents, this assumption is based on a theory that even though the instantaneous material cost index and instantaneous labor cost index would not be equal, the relationship of the two instantaneous cost indices would be at least correlated.  The basis of this assumption is that the pricing of each cost subcomponent occurred at the same relative time within the business cycle.  This assumption has been shown to be valid.8

 

Database and Results

 

Over my career, I collected every bit of estimating data I could put my hands on.  Definitions, articles, books, and most important of all data on completed projects.  I then organized the data in a database of over 250 projects (primarily refinery and petrochemical units, with some chemical plants, cogen units, among others), ranging from less than $10 MM to over $250 MM, with over 100 input fields per project the database has over 100 calculated cost to cost and cost to hour ratios.  Unfortunately, not all input fields were available for all projects, but even so, I had a tremendous amount of data to digest.  Using statistical functions available within the software utilized, I calculated means, medians, standard deviations, and plotted frequency curves, data scatter diagrams, and best fit regression curves.  The results were then summarized in tables to help make the information easier to find and use.9

 

One of the database relationships studied included TOT/Equip ratio (the Lang Factor).  Figure 1 shows the scatter diagram of the available ratios, including the straight-line trend line, and Figure 2 the frequency histogram showing the average and ± one standard deviation (one sigma). 

 

See Figure 1 of 3 below.

See Figure 2 of 3 below.

 

Note some scatter diagram points are far removed from the trend line.  This scatter does not necessarily indicate the data are erroneous, nor does it necessarily imply some particular difficulties were encountered during the project execution, although either or both are possible.  The straight-line trend line (shown) determination coefficient (R2) is 0.809, where the best-fit power function trend line R2 is 0.909.  Determination coefficients vary between 0.000 and 1.000 with the coefficient nearing 1.000 indicating strong correlation.  Both curve fit coefficient’s represent strong correlation between TOT and Equip.

 

The histogram plot shows two peaks and a variance from the average.  Again, do note the data furthest from the average is neither invalid nor necessarily incorrect or represent poorly executed work.  The average (5.122) is the central tendency of the data set and the significance of plus one sigma (6.862) and minus (3.381) one sigma about the average is that a majority of the data set are contained within these two boundaries.  In this particular case 62% of the database ratios cluster about this average.

 

As for the two peaks, further review of the data indicated sulfur units, cokers, and some chemical units tended to fall outside the plus one sigma boundary, while cogen and some chemical units tended to fall outside the minus one sigma boundary.

 

Without a question or doubt, all projects are different, with the variables affecting the differences being almost limitless; all projects face unforeseen problems and difficulties, also almost limitless in description.  However, the majority of the ratios clustered between 3.38 and 6.86, and averaged 5.12.  So, how do these ratio values compare with the sixty-year-old Lang factors?  Figure 3 plots the three Lang factors over the database frequency histogram. 

 

See Figure 3 of 3 below.          

           

 Conclusion

 

Lang’s fluid process plant factor, 4.74, and the database average ratio, 5.12 (majority being refinery units and petrochemical plants) are extremely close, within 8%. 

 

5.12 – 4.74 / 4.47 = 0.0802

 

Mr. Lang’s database of 14 estimates did not include the breadth of plants contained within the cited database, however even then the other two Lang Factors are certainly in close correspondence with the database TOT/Equip ratios.  If you are using the Fluids Lang Factor to obtain a quick rough order of magnitude estimate of a refinery or petrochemical unit the results will be within the estimate accuracy expected, say ±50%.  Another practical use would be the Capital Cost comparison of one revamp scheme against another or one type unit against a similar, competing type unit; just don’t forget to include Owner Cost, Operating Cost, etc, if appropriate.

 

On the other hand, care should be taken when the unit under consideration differs significantly from a fluid processing unit, particularly as mentioned sulfur units, cokers, some chemical units, and cogen units.  Even so, based on the results of a larger database, over a longer period, the results would indicate the use of Lang type factors are still valid within certain limits in obtaining quick, approximate cost estimates.

 

Literature Cited.


  1. Dysert, Larry R. “Sharpen your cost estimating skills,” Chemical Engineering, October 2001, New York: Chemical Week Associates.
  2. Gerrard, A M. Guide to Capital Cost Estimating. Rugby, Warwickshire, U.K: Institution of Chemical Engineers, 2000.
  3. Lang, Hans J.  “Engineering Approach to Preliminary Cost Estimates,” Chemical Engineering, September 1947, New York: Chemical Week Associates.
  4. Lang, Hans J.  “Cost Relationships in Preliminary Cost Estimation,” Chemical Engineering, October 1947, New York: Chemical Week Associates.
  5. Lang, Hans J.  “Simplified Approach to Preliminary Cost Estimates,” Chemical Engineering, June 1948, New York: Chemical Week Associates.
  6. Page, John S. Conceptual Cost Estimating Manual, Houston: Gulf Pub. Co., Book Division, 1984.
  7. Perry, Robert H, Don W. Green, and James O. Maloney. Perry's Chemical Engineers' Handbook, New York: McGraw-Hill, 1984.
  8. Wolf, Thomas E.  Capital Cost Estimates – Understanding, Analyzing, and Evaluating, Appendix 5, Amazon.com, 2011.
  9. Wolf, Thomas E.  Capital Cost Estimates – Understanding, Analyzing, and Evaluating, Chapter 6, Amazon.com, 2011.

 

Tom Wolf has over 40 years petroleum and petrochemical experience, including 25 years in project management.  He holds a BS degree in Mechanical Engineering and is a registered Professional Engineer.

 

© 2013 by Thomas E. Wolf.  All rights reserved.  No part of the information contained in these webpages may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or by any information storage or retrieval system without written permission of the copyright holder, except for the inclusion of brief quotations in a review.