# Fusion of Chemical Engineering & IE/OR

 4 Are there any examples or case studies where either industrial engineering or operations research concepts are used in Chemical Industries ? asked 11 Dec '10, 13:35 Ram 519●4●10●29 accept rate: 0% fbahr ♦ 4.6k●7●16

 7 Oh where to start? There are too many case studies to even enumerate. In fact, there is an entire discipline of chemical engineering that is devoted to applying techniques from OR/IE/applied mathematics to chemical process industry problems -- it is called "Process Systems Engineering" (PSE). Many of the chemical engineers in this area are actually applied mathematicians in disguise. For instance, much of the early research in MINLPs was actually driven by chemical engineers like Ignacio Grossmann. Another example: the interior-point NLP solver IPOPT came out of Larry Biegler's (another chemical engineer) group at CMU. The global optimizer BARON is a product of Nick Sahinidis's group (yet another chemical engineer). Here are some of the problems solved by chemical engineers (that intersect somewhat with the OR/IE): All manners of convex/nonconvex optimization problems. (LPs, QPs, SOCPs, LMIs, NLPs, MPECs, CPs) All kinds of stochastic programming problems. All kinds of discrete-continuous problems (MILPs, MIQPs, MIQCPs, MINLPs) Simulation problems. Control problems. All kinds of statistics problems, using multivariate methods (PCA, PLS). As you can see, optimization, control and statistics are the major themes here. Basically, a chemical engineer uses OR tools for: Optimal design of plants (optimization, simulation, numerical analysis) Optimal operation/control of plants (optimization, mostly LPs and QPs, some NLPs) Optimal planning and scheduling (mixed-integer optimization, mostly MILPs, occasionally CPs) Supply chain optimization (optimization, mixed integer programming) Identifying models of plants (statistics, time series analysis) State/quality variable estimation of plants (statistics, filter theory) Fault detection and monitoring (statistics, multivariate analysis, classification) Inferential/soft sensors for unmeasurable variables (statistics) Here's a list of faculty members in the Americas who are in the PSE discipline. http://cepac.cheme.cmu.edu/country.htm If you are an OR practitioner or researcher, I encourage you to make a friend in the PSE area. There are significant opportunities for research in the area of process systems engineering. While it is a fairly mature field, the cross-pollination of ideas from OR/IE may lead to interesting outcomes. OR/IE people study many classes of problems that are of immense interest to chemical engineers (and similarly, chemical engineers have developed numerous techniques that are potentially useful to OR folks, especially in the realm of techniques/algorithms/software for nonconvex nonlinear problems). Unfortunately, for some reason or other, the two sides don't talk to each other very much, so you see a lot of research happening in parallel, and invariably some duplication happens. For instance: the really big thing in PSE right now is large-scale supply chain optimization -- it's something that OR/IE people have been studying for years. Because of the lack of interaction, I'm told that PSE people sometimes end up rediscovering ideas that have been known to OR/IE people for years. answered 11 Dec '10, 23:16 Gilead ♦ 2.3k●5●14 accept rate: 15% Thank you very much. Your answer helped me to judge my track. (13 Dec '10, 13:47) Ram
 4 Check out the work by Ignacio Grossman and his group at Carnegie Mellon. http://www.cheme.cmu.edu/people/faculty/grossmann.htm answered 11 Dec '10, 14:43 Michael Trick ♦♦ 4.1k●5●16●33 accept rate: 20% Thank you for such an inspiring link. (13 Dec '10, 13:53) Ram
 3 I came here to say Ignacio Grossman and Victor Zavala but was beaten to the punch on the first. Victor's page is here: http://www.mcs.anl.gov/~vzavala/ answered 11 Dec '10, 19:50 AnORStudent 323●3●15 accept rate: 0%
 2 A quick search on INFORMS PubOnline found this journal entry "Resource Management in the Petrochemical Industry" Stadtherr, Mark A. 3151169, Rudd, Dale F. 2753711 Here is a JSTOR link to the same article. answered 11 Dec '10, 14:02 larrydag 1 ♦ 3.2k●6●14●26 accept rate: 9%
 2 The entire Ignacio Grossmann's academic family are working on this topic. See the tree at: http://egon.cheme.cmu.edu/pdf/IgnaciosTree.pdf Two recent graduates are Christoc Maravelias at Wisc and Fengqi You at Northwestern: http://www.engr.wisc.edu/che/faculty/maravelias_christos.html http://www.chem-biol-eng.northwestern.edu/people/faculty/you.html Victor Zavala has done most of the work on process control, but his graduate advisor, Larry Biegler is well-known in the nonlinear programming community due to the IPOPT solver: http://numero.cheme.cmu.edu/ answered 12 Dec '10, 22:10 Zoey 21●2 accept rate: 0% Thanks for the links. (13 Dec '10, 13:36) Ram
 2 There are many IE/OR applications at Chemical Industries. I currently work for a chemical company, and the opportunities are endless. I am currently applying genetic algorithms to find optimal campaign schedules. Discrete event simulation is also very popular to solve supply chain problems. I have an article on Chemical Processing Magazine that talks about using Simulation in the Chemical Industry: http://www.chemicalprocessing.com/articles/2010/178.html answered 17 Dec '10, 16:05 dcope 31●1 accept rate: 0%
 1 I'm pretty sure LP models are used to control flows in oil refineries, and I think nonlinear programming is used to determine to control certain chemical synthesis processes (but I'm neither a chemist nor a chemical engineer, so I'm not positive). answered 11 Dec '10, 16:10 Paul Rubin ♦♦ 14.6k●5●13 accept rate: 19% That's actually pretty close. In a refinery, there are usually several optimization layers; near the top there is a Real Time Optimizer, which solves a large NLP to figure out what to produce to achieve maximum profit (based on current prices etc.). It then sends targets to the control system, which solves an LP to figure out how to manage local controllers such that they are able meet targets. The local controllers themselves solve a QP to implement optimal control moves (based on an empirical model of the plant). It's actually more complicated than that, but that's the basic idea. (11 Dec '10, 23:27) Gilead ♦ @Gilead: Thanks for the confirmation ... further support for the adage that "even the blind squirrel occasionally finds an acorn". :-) (13 Dec '10, 23:17) Paul Rubin ♦♦
 1 Half year ago, I joined a software firm for chemical engineering. I am surprised by the extend of OR usage in this area (Non-linear, linear and discrete optimization). answered 22 Dec '10, 22:56 Xiaodong Zhang 61●1 accept rate: 0%
 1 There is a journal called computers and chemical engineering which incorporates a great deal of OR applied in chemical field as shown in the link below http://www.elsevier.com/wps/find/journaldescription.cws_home/349/description#description answered 07 Sep '11, 23:08 Jiadong Wang 70●4 accept rate: 0% thank you so much.........now I understood the importance of this journal (11 Jun '12, 08:55) Ram
 0 Woolsey once wrote that chemist made good ORs because they could understand abstract models but also kept the inputs and outputs balanced. He also has a great story about disappearing orange juice concentrate that is great for on practical experience. answered 13 Dec '10, 18:22 Laurence 41●3 accept rate: 0%
 toggle preview community wiki

By Email:

Markdown Basics

• *italic* or _italic_
• **bold** or __bold__
• image?![alt text](/path/img.jpg "Title")
• numbered list: 1. Foo 2. Bar
• to add a line break simply add two spaces to where you would like the new line to be.
• basic HTML tags are also supported

Tags: