464707 Improved Data-Driven Mathematical Modeling and Global Optimization Framework: An Application in Refinery Planning Operations

Tuesday, November 15, 2016: 1:27 PM
Carmel II (Hotel Nikko San Francisco)
C. Doga Demirhan1,2, Fani Boukouvala1,2, Kyungwon Kim3, Hyeju Song2,4 and Christodoulos A. Floudas1,2, (1)Artie McFerrin Department of Chemical Engineering, Texas A&M University, College Station, TX, (2)Texas A&M Energy Institute, Texas A&M University, College Station, TX, (3)Process Optimization Team, Hyundai Oilbank, Seosan, South Korea, (4)Chemical Engineering, Inha University, Incheon, South Korea

Improving the planning and scheduling operations is one of the major challenges in the petrochemical industry considering the tight competition, environmental regulations, and lower profit margins [1]. A recent work of our group developed a data-driven modeling and global optimization-based planning formulation for highly integrated refinery-petrochemical complexes [2]. In this work, an improved and more generalized framework for planning operations is applied to an actual refinery operated by Hyundai Oilbank and located in Korea.

The refinery planning problem has received considerable attention since the introduction of linear programming in 1950’s. Planning is a topic of special interest for many process systems engineers and in this respect, nonlinear models and specialized algorithms have been proposed for refinery operation but global optimality was not guaranteed in any method [3]. Integration and interaction between refineries and chemical plants is another level of complexity that is addressed only by few researchers [2,4]. Although there are commercially available highly detailed mathematical models for simulation of processing units, such detailed models cannot be used in enterprise-wide operations due to computational expense. Data-driven modeling offers a promising way to obtain inexpensive nonlinear models [5,6], which can relate relevant inputs to relevant outputs to describe planning operations.

The first step of this work consists of organizing, analyzing, and processing the raw data provided by the industrial partner. In order to reduce the variability in the daily data, weekly averages are calculated and used. A detailed outlier analysis is carried out and missing points in the data set are imputed. Having decided on the functional relationship between inputs and outputs, data-based nonlinear models are developed for the processing units. The primary approach is to use quadratic, polynomial, signomial, exponential, and logarithmic models to predict product yields and properties for all of the production units. The parameters of the postulated models are globally optimized using the state-of-the-art, commercial deterministic global optimization solver ANTIGONE [7], based on a large data set of daily production and property measurements. All the models are created in a generalized and an automated fashion, which improved the applicability of the framework.

The second step of the work is creating a superstructure containing all possible connections and operating modes in the refinery. The resulting single-period planning model is a large-scale non-convex mixed integer nonlinear optimization model, which is solved to ε-global optimality using ANTIGONE [7]. Several case studies with candidate months of operation are chosen and the profit calculated by the formulation is compared with actual plant profit. All the outlier analysis, data processing, and planning are done with close collaboration with our industrial partner. Results of the case studies are provided to illustrate the efficiency of our proposed model and global optimization approach.


References:

  1. CA Floudas, AM Niziolek, O Onel, LR Matthews, Multi-Scale Systems Engineering for Energy and the Environment: Challenges and Oppurtunities, AIChE Journal, 62 (3) (2016), pp. 602-623
  2. J Lie, X Xiao, F Boukouvala, B Zhao, G Du, X Su, H Liu, CA Floudas, Data-Driven Mathematical Modeling and Global Optimization Framework for Entire Petrochemical Planning Operations, AIChE Journal, 2016, DOI: 10.1002/aic.15220
  3. NK Shah, Z Li, MG Ierapetritou, Petroleum refining operations: key issues, advances, and opportunities. Industrial and Engineering Chemisty Research, 50 (3) (2011), pp. 1161-1170
  4. K Al-Qahtani, A Elkamel, Multisite refinery and petrochemical network design: optimal integration and coordination, Industrial and Engineering Chemisty Research, 48 (2) (2009), pp. 814-826
  5. P Guyonnet, FH Grant, MJ Bagajewicz, Integrated model for refinery planning, oil procuring, and product distribution, Industrial and Engineering Chemisty Research, 48 (1) (2009), pp. 463-482
  6. SJ Qin, Process data analytics in the era of big data, AIChE Journal, 60 (9) (2014), pp.3092-3100
  7. R Misener, CA Floudas, ANTIGONE: Algorithms for coNTinuous Integer Global Optimization of Nonlinear Equations, Journal of Global Optimization, 59 (2–3) (2014), pp. 503–526

Extended Abstract: File Not Uploaded
See more of this Session: Industrial Applications in Design and Operations
See more of this Group/Topical: Computing and Systems Technology Division