Dr. Pierre Lennox Bhoorasingh became the second PhD graduate from the CoMoChEng group this week! He defended his thesis, “Automated Calculation of Reaction Kinetics Via Transition State Theory”, on June 8th, 2016. Pierre is now working for Pill Pack in Boston, MA.
Dr. Fariba Seyedzadeh Khanshan officially became the first PhD graduate from the CoMoChEng group this week! She defended her thesis, “Automatic Generation of Detailed Kinetic Models for Complex Chemical Systems”, on January 29, 2016. Fariba is currently working as a post-doc at Shire in Lexington, MA.
At the 9th US National Combustion Meeting in Cincinnati OH, progress on our mechanism importer project was presented. The presentation was titled “Identification, Correction, and Comparison of Detailed Kinetic Models” (Presentation date: May 20, 2015). Unfortunately we couldn’t attend the meeting in person, but Dr. C. Franklin Goldsmith was kind enough to give the presentation for us, and by all accounts did a wonderful job.
This material is based upon work supported by the National Science Foundation under Grant Number 1403171. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation
On September 15, 2014, the CoMoChEng lab participated in the Northeastern University College of Engineering Undergraduate Research Fair.
We have many opportunities for undergraduate researchers at all levels. The flyer above has “Programmers” writ large, but most of our many successful researchers started with no programming experience, and many of our projects will not require you to touch the source code, so do not be discouraged if you can’t code!
Of course, if you can code, then what are you waiting for? Come and put your skills to good use, and improve them in the process!
Richard, Pierre, and Fariba are in San Francisco this August for the 35th International Symposium on Combustion. The Computational Modeling in Chemical Engineering group is presenting four posters on Wednesday. If you’re at the conference, you should come and check them out! Our graphical abstracts are below. Contact us for the full posters.
At the 2nd International Workshop on Flame Chemistry, Dr. West presented the group’s work on two projects related to automatic mechanism generation. The presentation was titled “Reaction Mechanism Generator: Toward High-Throughput Transition State Calculations, and Interpreting Existing Kinetic Models” (Presentation date: August 4, 2014)
ChE Assistant Professor Richard West was awarded a $140K NSF Grant to “Identify and resolve discrepancies in kinetic models of hydrocarbon combustion“.
Computational kinetic modeling of combustion chemistry has made significant progress in recent decades. Dozens of recent models, which describe tens of thousands of simultaneous reactions between thousands of intermediate species, are capable of explaining complicated combustion phenomena, allowing increasingly accurate engine simulations, and screening novel biofuels. However, these ever-proliferating detailed kinetic models are incompatible and inconsistent, are seldom compared directly, and often contain undetected mistakes.
The commonly used format to publish these models, devised in the 1970’s when input was limited by the width of 80-column punch-cards, forces model-builders to abbreviate species’ names, thereby losing their chemical identity, and to discard other metadata. The main challenge in comparing these models is in recognizing, for example, that the name “C3KET21” in one model represents 1-hydroperoxypropan-2-one, which another research group may have named “CH3COCH2O2H” in a different model.
This project develops tools to help identify the chemical species in a kinetic model, to facilitate comparison of models. The new tools will be built upon the open-source Reaction Mechanism Generator (RMG-Py) software, that we have been developing in a collaboration between Northeastern and the Massachussets Institute of Technology. A web-based user interface will make it easy for users to import models to the database, and provide instant reward for doing so (the ability to check the model, fill in gaps, and merge with other models). The proposed work will massively reduce the barriers to converting detailed kinetic models into a machine-readable format with clear and consistent species definitions. This will enable: (1) comparison between models, leading to more replicable science, (2) error identification, leading to higher quality data, (3) better rate estimates and a broader impact for mechanism generation software, and (4) wider use of process informatics tools. The unified database of all previous kinetic models will greatly assist progress in combustion modeling.
CoMoChEng graduate student Belinda Slakman successfully defended her PhD thesis proposal this spring. Belinda’s presentation, “Using Reaction Mechanism Generator (RMG) to study complex liquid-phase systems”, was held on April 11, 2014. Well done Belinda! The CoMoChEng group now officially has three PhD candidates.
The Computational Modeling in Chemical Engineering group attended the 32nd Regional Meeting on Kinetics and Dynamics on January 25, 2014. The conference was held at Trinity College in Hartford, CT. The group had a great presence at the meeting with Fariba, Pierre, Belinda and Victor presenting the following four talks:
- F. Seyedzadeh Khanshan and R.H. West. “Generating Mechanism for Biofuels: Updating Reaction Families.”
- P.L. Bhoorasingh and R.H. West. “Group-Additive Estimation of Distances for Transition State Geometries.”
- B. Slakman and R.H. West. “Automatic Mechanism Generation for Liquid-Phase Systems.”
- V.R. Lambert and R.H. West. “Comparing Detailed Kinetic Models.”
In 2013 we were awarded a $100k Doctoral New Investigator (DNI) Grant by the American Chemical Society Petroleum Research Fund.
In January 2014 we began work on the project, titled “Transition-State Prediction for High-Throughput Calculation of Accurate Chemical Reaction Rates”.
Even conventional engine designs require hundreds of hours of costly tuning to optimize, for each fuel; novel engine designs depend even more critically on the combustion behavior of the fuels. More accurate kinetic models will allow proposed fuels, and engine designs, to be screened more reliably in silico, enabling significant savings in time and money.
Modern quantum chemistry methods allow reaction rate coefficients to be calculated ab inito with high accuracy, but the current pace of these high-accuracy rate calculations is at least an order of magnitude too slow. With the advance of High Performance Computing the bottle-neck is no longer the computation, but the human interaction required to set up the calculation. Only once these calculations are automated will predictive kinetics realize the full potential of ever-increasing computational power.
This project aims to determine how to locate transition state geometries automatically, so that quantum mechanical electronic structure calculations can calculate accurate reaction rate expressions without human input. The hypothesis is that the transition state geometry is mostly determined by the reaction family, the chemical structure near the reacting site, and the geometries of the reactants; a hierarchical group-contribution scheme, similar to that used to estimate reaction rates, should therefore, with modification, be able to estimate transition state geometries. Incorporated into a software platform that automatically generates detailed reaction mechanisms, and coupled with increasingly high-performance computers, this will enable the high-throughput calculation of thousands of accurate reaction rates for petroleum combustion and other fields.