(373r) Optimal Design and Operation of Flexible Energy Polygeneration Systems Using Decomposition Algorithms

Energy polygeneration systems offer several economic and environmental advantages over single product systems as a result of tight integration of multiple processes into one plant [1]. Furthermore, using multiple feedstocks may enable the exploitation of certain synergies, for instance, by heat integration of exothermic and endothermic processing units or blending the different qualities of syngas generated to provide the correct ratio for downstream synthesis processes [2], [3]. In addition, utilizing solid waste as a sustainable feedstock together with a conventional fossil fuel (such as natural gas) has the added benefit of providing an efficient means of waste disposal with only a relatively small negative impact on economics [4]. In this paper, the optimal design and operation of an energy polygeneration system that converts a hybrid solid waste and natural gas feedstock to electricity, liquefied synthetic natural gas (LNG), dimethyl ether (DME) or olefins is investigated. In order to maintain competitiveness in the face of uncertainties (for instance in product prices or environmental policies), a flexible design solution that enables adjustment of operating conditions to maximize the production of the most valuable product(s) is proposed [5]. In this work, 100 % operational flexibility is assumed thus no limits are imposed on the turndown ratios of the product trains. Previous research on the optimal design and operation of flexible polygeneration systems suggests that under most scenarios the plant should produce only one fuel or chemical at a time in addition to electricity [6],[7]. Thus, the proposed process operates in either an LNG, DME or olefins mode, with electricity produced from off-gases in all three modes. In the LNG mode, the syngas generated from solid waste gasification is cleaned and sent to a methanation process to produce synthetic natural gas which is then blended with the natural gas feed prior to liquefaction. In the DME or olefins modes, the different qualities of syngas from solid waste gasification and natural gas reforming are blended in correct proportions and sent to their respective synthesis trains.

The above problem is formulated as a scenario-based two-stage stochastic MINLP to optimize simultaneously design and operational decision variables. This work uses the recently developed GOSSIP software framework that implements decomposition algorithms such as nonconvex generalized Benders decomposition (NGBD) ([8], [9]), Lagrangian relaxation, and modified Lagrangian relaxation for efficient and scalable solution of nonconvex stochastic MINLPs [10]. The results of the stochastic formulation are compared with the deterministic approach to demonstrate the improvement in economic performance as a result of taking uncertainty into consideration.

[1] T. A. Adams II, J. H. Ghouse, Polygeneration of fuels and chemicals, Current Opinion in Chemical Engineering 10 (2015) 87–93.

[2] T. A. Adams II, P. I. Barton, Combining coal gasification and natural gas reforming for efficient polygeneration, Fuel Processing Technology 92 (2011) 639–655.

[3] C. A. Floudas, J. A. Elia, R. C. Baliban, Hybrid and single feedstock energy processes for liquid transportation fuels: a critical review, Computers & Chemical Engineering 41 (2012) 24–51.

[4] C. C. Cormos, Assessment of flexible energy vectors polygeneration based on coal and biomass/solid wastes co-gasification with carbon capture, International Journal of Hydrogen Energy 38 (2013) 7855–7866.

[5] Y. Chen, T. A. Adams II, P. I. Barton, Optimal design and operation of flexible energy polygeneration systems, Industrial & Engineering Chemistry Research 50 (2011) 4553–4566.

[6] Y. Chen, X. Li, T. A. Adams II, P. I. Barton, Decomposition strategy for the global optimization of flexible energy polygeneration systems, AIChE Journal 58 (2012) 3080–3095.

[7] T. A. Adams II, T. Thatho, M. Le Feuvre, C. L. Swartz, The optimal design of a distillation system for the flexible polygeneration of dimethyl ether and methanol under uncertainty, Frontiers in Energy Research 6 (2018) 41.

[8] X. Li, A. Tomasgard, P. I. Barton, Nonconvex generalized benders decomposition for stochastic separable mixed-integer nonlinear programs, Journal of optimization theory and applications 151 (2011) 425.

[9] X. Li, A. Sundaramoorthy, P. I. Barton, Nonconvex generalized benders decomposition, in: Optimization in Science and Engineering, Springer, 2014, pp. 307–331.

[10] R. Kannan, Algorithms, analysis and software for the global optimization of two-stage stochastic programs, Ph.D. thesis, Massachusetts Institute of Technology, 2017.