Premium
Optimal purchasing of raw materials: A data‐driven approach
Author(s) -
Muteki Koji,
MacGregor John F.
Publication year - 2008
Publication title -
aiche journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.958
H-Index - 167
eISSN - 1547-5905
pISSN - 0001-1541
DOI - 10.1002/aic.11494
Subject(s) - raw material , purchasing , sequential quadratic programming , process (computing) , quality (philosophy) , product (mathematics) , refining (metallurgy) , process engineering , purchasing process , coke , computer science , mathematical optimization , nonlinear programming , quadratic programming , engineering , mathematics , nonlinear system , waste management , materials science , operations management , chemistry , philosophy , operating system , geometry , epistemology , quantum mechanics , metallurgy , physics , organic chemistry
Abstract An approach to the optimal purchasing of raw materials that will achieve a desired product quality at a minimum cost is presented. A PLS (Partial Least Squares) approach to formulation modeling is used to combine databases on raw material properties and on past process operations and to relate these to final product quality. These PLS latent variable models are then used in a sequential quadratic programming (SQP) or mixed integer nonlinear programming (MINLP) optimization to select those raw materials, among all those available on the market, the ratios in which to combine them and the process conditions under which they should be processed. The approach is illustrated for the optimal purchasing of metallurgical coals for coke making in the steel industry. However, it is well suited to many similar problems such as the purchasing of crude oils for refining, of ingredients for processed foods and of polymeric materials to blend into functional polymers. © 2008 American Institute of Chemical Engineers AIChE J, 2008