(373ac) A Model for the Autoclave Low-Density Polyethylene (LDPE) Industrial Process

Low-density polyethylene (LDPE) represented 18% of the polyethylene market worldwide with a 20 Mt production, in 2020. Its primary use is in films, injection, and molding applications [1,2]. Consumers are getting progressively more demanding, forcing the polymers industry to develop more tailor-made polymers to meet their clients’ requirements. In addition, the 2050 net-zero target of Plastic Europe means they also need to focus on minimizing process emissions [2]. This has led to renewed interest in process modeling and optimization.

In this work, we model the LDPE production process at Repsol Polymers Sines plant, using Aspen Plus^® V12.1. The reactor, an autoclave with 5 different zones, operates at high pressure (1200-2000 bar) and high temperatures (170-300 ºC). To control the polymer chain growth and achieve the desired grade, termination agents such as n-butane or propylene are added to the reactors’ feed streams. The plant also produces the copolymer Ethylene Butyl Acrylate (EBA), by adding the comonomer to the inlet mixture. In contrast to previously modeled autoclaves, the inlet stream is divided into 5, that will enter in different points of the reactor. The novelty of the present work lies in exploring a different variant of a LDPE autoclave reactor. Since there is limited information in the literature regarding EBA copolymerization, especially compared to other copolymers like Ethylene Vinyl Acetate (EVA) [3,4], this study provides new insights into EBA copolymerization.

The reactor was modeled as a series of CSTRs, representing each zone of the autoclave. To address non-ideal mixing in each zone, two different CSTRs blocks were used for each zone. Since the kinetic parameters from literature [5–9] led to significant deviations from the plant data, we were forced to determine new kinetic parameters for the several reactions involved. This was done with Aspen’s Data Fit tool, considering process data gathered from 7 different LDPE grades. In a second stage, the EBA grades were incorporated into the model and the cross-reaction and the butyl acrylate reaction parameters were regressed using grades with varying levels of comonomer incorporation. The reactor’s output variables were the number-average molecular weight (Mn), polydispersity index (PDI) and short-chain branching frequency (SCB).

After the reactor, the model was extended to the separation system by estimating the binary interaction parameters of the PC-SAFT EOS thermodynamic model. A non-equilibrium model was built for the higher-pressure separation stage based on previous work [10]. The lower separation phases were modeled with a Flash2 block, assuming equilibrium conditions. Then the compressors were added. The primary compressor was modeled with a MCompr model (3 stages), whereas the secondary compressors used 6 Compr and 4 Heaters to simulate the 2 pistons of the first stage, the intercooling and the 4 pistons of the final stage. The compressors’ efficiencies were then estimated to replicate their energy consumption, which represents a large share of the process energy consumption.

Acknowledgments: Financial support from Fundação para a Ciência e Tecnologia (FCT) through project UIDB/04028/2020.

References:

[1] J. Soares, T. McKenna, An in House Industrial Short Course on Olefin Polymerization Processes for Repsol, 2022.

[2] Association of Plastics Manufacturers in Europe, Annual report, 2022.

[3] I.R. Alleyne, Optimal grade transition policies for a high pressure EVA polymerization plant, 2006. https://doi.org/10.7939/R3-YNRN-XB21.

[4] I.L. Chien, T.W. Kan, B.S. Chen, Dynamic simulation and operation of a high pressure ethylene-vinyl acetate (EVA) copolymerization autoclave reactor, Comput Chem Eng 31 (2007) 233–245. https://doi.org/10.1016/J.COMPCHEMENG.2006.06.007.

[5] N. Agrawal, G.P. Rangaiah, A.K. Ray, S.K. Gupta, Multi-objective optimization of the operation of an industrial low-density polyethylene tubular reactor using genetic algorithm and its jumping gene adaptations, Ind Eng Chem Res 45 (2006) 3182–3199. https://doi.org/10.1021/IE050977I.

[6] P.D. Iedema, M. Wulkow, H.C.J. Hoefsloot, Modeling molecular weight and degree of branching distribution of low-density polyethylene, Macromolecules 33 (2000) 7173–7184. https://doi.org/10.1021/MA991711O.

[7] S. GOTO, Y. K, F. S, S. M, COMPUTER MODEL FOR COMMERCIAL HIGH-PRESSURE POLYETHYLENE REACTOR BASED ON ELEMENTARY REACTION RATES OBTAINED EXPERIMENTALLY, (1981).

[8] J.Y. Ham, H.K. Rhee, Modeling and control of an LDPE autoclave reactor, J Process Control 6 (1996) 241–246. https://doi.org/10.1016/0959-1524(95)00052-6.

[9] P. Lorenzini, M. Pons, J. Villermaux, Free-radical polymerization engineering-III. Modelling homogeneous polymerization of ethylene: mathematical model and new method for obtaining molecular-weight distribution, Chem Eng Sci 47 (1992) 3969–3980. https://doi.org/10.1016/0009-2509(92)85146-3.

[10] A. Buchelli, M.L. Call, A.L. Brown, C.P. Bokis, S. Ramanathan, J. Franjione, Nonequilibrium Behavior in Ethylene/Polyethylene Flash Separators, Ind Eng Chem Res 43 (2004) 1768–1778. https://doi.org/10.1021/ie0302037.