(188w) Data-Driven Modeling and Optimization of an Ethane Steam Cracker

Beykal, B., Texas A&M University
Onel, O., Princeton University
Pistikopoulos, E. N., Texas A&M Energy Institute, Texas A&M University
Natural gas liquids (NGL) production has been increasing, more than 100% since 2008, due to the recent advancements in shale gas industry. [1] It is further expected to grow by 27% until 2040 [2, 3] where these trends significantly motivate the industry to build new ethane crackers with combined ethylene production capacity of 12.5 million tons/year in the United States. [4] Thus, it is essential to determine the optimal reactor design and operating conditions for maximum economic benefit from this process. However, the kinetic models describing the steam cracking of ethane are rigorous and commonly characterized by complex partial differential equations and first-principle models. This creates a formidable challenge on the global optimization of this process using deterministic methods since the calculation of the derivatives are not practical. Previously, a mathematical framework was developed for the optimization of steam cracking through the full discretization of the model via orthogonal collocation [5].

In this study, we are optimizing the operating conditions and the sizing of an ethane steam cracker using a data-driven approach to simultaneously maximize the yield of ethylene product while minimizing the reactor coking. We are using the constrained grey-box optimization framework ARGONAUT [6] to sample the feasible space, create tractable approximations using surrogate modeling to accurately represent the input-output relationship for the steam cracker, and to solve these formulations to global optimality using the state-of-the art Mixed-Integer Nonlinear Programming (MINLP) global optimization solver ANTIGONE. [7] We compare the optimal solution of ARGONAUT with the locally optimal solution obtained using full discretization (via orthogonal collocation on finite elements on the mass, momentum and energy balances) as well as with the solutions obtained using other gradient-free and gradient-based algorithms (such as NOMAD and EGO).

[1] Floudas, C. A.; Niziolek, A. M.; Onel, O.; Matthews, L. R., Multi-scale systems engineering for energy and the environment: Challenges and opportunities. AIChE Journal 2016, 62 (3), 602-623.

[2] Energy Information Administration, Annual Energy Outlook, 2015.

[3] Energy Information Administration, Petroleum and Other Liquids, 2015.

[4] Chang, J., New projects may raise US ethylene capacity by 52%, PE by 47%, 2014, ICIS Petrochemicals.

[5] Onel, O.; Niziolek, A. M.; Floudas, C.A., Toward Shale gas to Light Olefins I: A dynamic optimization framework for steam cracking of natural gas liquids (NGLs). 2017. In preparation

[6] Boukouvala, F.; Floudas, C. A., ARGONAUT: AlgoRithms for Global Optimization of coNstrAined grey-box compUTational problems. Optimization Letters 2016, 1-19.

[7] Misener, R.; Floudas, C., ANTIGONE: Algorithms for coNTinuous / Integer Global Optimization of Nonlinear Equations. Journal of Global Optimization 2014, 59, (2-3), 503-526.


This paper has an Extended Abstract file available; you must purchase the conference proceedings to access it.


Do you already own this?



AIChE Members $150.00
AIChE Graduate Student Members Free
AIChE Undergraduate Student Members Free
Non-Members $225.00