Optimization methods for decisions in industrial gas network problems

Lead University: Lehigh University
PI: Luis F. Zuluaga
Co-PIs: Alberto Lamadrid

Air Products provides indispensable products like oxygen, nitrogen, and hydrogen to manufacturing, health care, transportation, and other essential industries worldwide. In a 2012 study by the American Chemistry Council, it was shown that industrial gases companies produced approximately $17 billion worth of products in 2010 and employed approximately 60,000 American workers. The objective of this project is threefold: 1) As a continuation of previous PITA-funded work on the optimization of the industrial gas network at Air Products, this project will develop the optimization of the industrial gas network under the posibility of production capacity expansion or reduction. Adding this factor makes the underlying optimization problem much more challenging. 2) As a continuation of previous PITA-funded work on sensor fault detection using sensibility analysis techniques, this project will develop new sensitivity analysis techniques to address the problem of quantifiying the effect of changing parameters used to make production decisions about the industrial gas network. Such parameters measure the company's total cost and revenue from their industrial gas business, and the ability to fine-tune them will substantially affect the bottom line. 3) In a new line of collaboration, this project will help Air Products optimize their participation in future gas markets. This requires handling non-linear and discontinuous functions that reflect the behavior of the contracts, as well as gaining a deep undertanding of these gas markets. To that end, a new approximation technique will be used to make the optimization of the company's gas market participation a tractable one for current optimization solvers. This project will be done in close collaboration with the Air Products R&D group, faculty from Lehigh's Industrial and Systems Engineering (ISE) Department and the College of Business and Economics, and with the participation of at least one graduate student.