(300b) Integration of Crude-Oil Scheduling and Refinery Planning By Lagrangean Decomposition Approach | AIChE

# (300b) Integration of Crude-Oil Scheduling and Refinery Planning By Lagrangean Decomposition Approach

Conference

Year

Proceeding

Group

Session

Time

## Authors

Carnegie Mellon University
Carnegie Mellon University
Carnegie Mellon University
Integration of crude-oil scheduling and refinery planning by Lagrangean decomposition approach

Haokun Yang, David E. Bernal, Ignacio E. Grossmann

Chemical Engineering Department, Carnegie Mellon University, Pittsburgh PA.

Problems of crude-oil scheduling and refinery planning are usually considered hard to solve to optimality at an enterprise level. The former deals with the crude oil inventory considering several types of crude oil which are delivered by different vessels and used as feedstock for the Crude Distillation Unit (CDU). The latter deals with the oil production starting at the CDU until final products involving refining and blending operations. Mixed-integer Nonlinear Programming (MINLP) models have been developed with both discrete-time and continuous-time formulation for the scheduling problem and solved with linearization based algorithms, e.g. [1] and [2]. On the other hand, the refinery planning often involves the pooling problem and usually consists of optimizing feedstocks, unit settings, as well as final product blending and shipping. The CDU modeling has used traditionally used approaches like fixed yield or swing cuts Linear Programming (LP) models. Recently, Alattas et al. [3] developed a nonlinear model of the CDU based on fractionation indices and integrated it in an MINLP model involving production planning, also considering the changeover of crudes [4]. Since both problems aim at an economic objective and are linked by the feedstock of the CDU, it becomes interesting to solve an integrated problem which may result in a better solution than the one obtained by solving each one separately. The challenge of this approach is the time scale difference between the two problems which may result in a large-scale MINLP model.

Lagrangean decomposition is a solution method which decomposes an optimization problem with complicating constraints into more easily solvable subproblems which solved separately and iteratively converge to the solution of the original problem by dualizing those complicating constraints. This method has shown a significant advantage in solving scheduling and planning integrated problems [5]. Mouret et al. [6] integrated a crude oil scheduling model with a continuous-time formulation and a simplified refinery planning model using an Artificial Neural Network (ANN) for the CDU modeling [7]. In Mouretâ€™s work, the advantages of considering an integrated approach of the crude oil scheduling and refinery planning problem were showed by achieving a time efficient framework for these problems using Lagranean decomposition.

The integrated refinery problem formulation and solution method using a Lagrangean decomposition algorithm are described in this work. We have studied the effect of having a continuous-time and a discrete-time formulation for the crude oil scheduling model and used several modeling approaches for the CDU. Two different kinds of comparisons have been made using the results of this work. Our model, which includes both the scheduling and planning simultaneously using Lagrangean decomposition with linking constraints for the CDU, is compared against a non-integrated approach to the crude oil scheduling and followed by the refinery planning. The results show an improvement of the economic objective function with respect to the sequential approach. Furthermore, we present a comparison of the solution time against solving the monolithically integrated model using state-of-the-art MINLP solvers, showing the advantage of using a Lagrangian Decomposition algorithm for this problem.

[1] H. Lee, J. M. Pinto, I. E. Grossmann, and S. Park, â€œMixed-integer linear programming model for refinery short-term scheduling of crude oil unloading with inventory management,â€ Ind. Eng. Chem. Res., vol. 35, no. 5, pp. 1630â€“1641, 1996.

[2] S. Mouret, I. E. Grossmann, and P. Pestiaux, â€œA novel priority-slot based continuous-time formulation for crude-oil scheduling problems,â€ Ind. Eng. Chem. Res., vol. 48, no. 18, pp. 8515â€“8528, 2009.

[3] A. M. Alattas, I. E. Grossmann, and I. Palou-Rivera, â€œIntegration of nonlinear crude distillation unit models in refinery planning optimization,â€ Ind. Eng. Chem. Res., vol. 50, no. 11, pp. 6860â€“6870, 2011.

[4] A. M. Alattas, I. E. Grossmann, and I. Palou-Rivera, â€œRefinery production planning: Multiperiod MINLP with nonlinear CDU model,â€ Ind. Eng. Chem. Res., vol. 51, no. 39, pp. 12852â€“12861, 2012.

[5] B. Brunaud and I. E. Grossmann, â€œPerspectives in Multilevel Decision-making in the Process Industry,â€ pp. 1â€“34, 2017.

[6] S. Mouret, I. E. Grossmann, and P. Pestiaux, â€œA new Lagrangian decomposition approach applied to the integration of refinery planning and crude-oil scheduling,â€ Comput. Chem. Eng., vol. 35, no. 12, pp. 2750â€“2766, 2011.

[7] T. Gueddar and V. Dua, â€œNovel model reduction techniques for refinery-wide energy optimisation,â€ Appl. Energy, vol. 89, no. 1, pp. 117â€“126, 2012.

## Topics

### Checkout

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

### Pricing

#### Individuals

##### 2018 AIChE Annual Meeting
 AIChE Pro Members \$150.00 AIChE Graduate Student Members Free AIChE Undergraduate Student Members Free AIChE Explorer Members \$225.00 Non-Members \$225.00
##### Fuels and Petrochemicals Division only
 AIChE Pro Members \$100.00 Fuels and Petrochemicals Division Members Free AIChE Graduate Student Members Free AIChE Undergraduate Student Members Free AIChE Explorer Members \$150.00 Non-Members \$150.00