(248f) A Methodology for Process Debottlenecking By Process Intensification: Application to Ethylene Oxide Production

Improvements in process economics and sustainability are key challenges for the chemical industry. Process Intensification (PI), which has broadly been defined as the development of innovative apparatuses and techniques that offer drastic improvements in chemical manufacturing and processing [1], offers potential solutions to this challenge. PI aims at significant enhancement of physical and chemical phenomena, achieved in process-intensifying equipment, such as e.g. microreactors, advanced heat exchangers or by integration of functions (reaction-separation), introduction of new energy sources or process control methods. As a result, PI technologies may replace large, expensive and energy-intensive apparatuses with smaller, more efficient and economically competitive ones [1]. Moreover, PI technologies may overcome process bottlenecks, which cannot be tackled with the classical equipment. Therefore, the interest in PI is increasing steadily, documented by numerous experimental demonstrations of intensified equipment [2]. Nevertheless, up till now, the number of successful PI implementations on an industrial level is limited [3]. One of the reasons for this is the complexity in identifying the most suitable PI option for a given process, necessitating means for an efficient determination of the possible improvement. Due to time and resource constraints detailed modelling or experimental investigations for all existing PI equipment is generally not feasible. Therefore, this paper presents a systematic methodology enabling process analysis, bottleneck identification and selection of the most promising PI options for a detailed investigation.

Numerous PI options for tackling a bottleneck are identified on the basis of a broad database of PI technologies, such as intensified mass or heat-exchange reactors, reactive separations and hybrid processes. The database is an extended version of the one previously presented by Lutze et al. [4]. The number of feasible options is systematically reduced by consideration of technological and process specific constraints. Application of several short-cut methods for fast quantification of possible process improvement enables the pre-selection of the most promising equipment options. The most suitable option for overcoming a given bottleneck is modelled in detail and investigated by means of rigorous process simulation. The methodology is illustrated for the case study of ethylene oxide production.

Before the methodology for process debottlenecking can be applied, the process that is to be improved needs to be analysed, which either can be done on the basis of process plant measurements or a detailed flowsheet simulation model. Subsequently, the targeted process improvement is defined and formulated as an objective function to be optimized, such as e.g. minimizing operational cost per unit of product. Furtherly, the process bottlenecks need to be identified with respect to the defined objective function. The determination of relevant bottlenecks is achieved based on energy and mass efficiencies calculated for every piece of equipment (what corresponds e.g. to selectivity for a reactor). The values of efficiency factors for the base-case design are subject to sensitivity analysis and manipulated to some value of uncertainty. Based on the results of the sensitivity analysis, the mass or energy efficiency factors having the most significant influence on the improvement of the objective function are identified. Consequently, the equipment and related mass or energy efficiency is selected as the major target for intensification.

Subsequently, the PI database is screened for possible intensified options for the identified bottleneck. The database includes more than 150 intensified technologies previously reported in the literature, which are described by the type of unit operation (e.g. reaction and separation processes, heat exchange) operating parameters (pressure, temperature, available substrate and product phases), and the specific phenomena that present the limitation causing the bottleneck (e.g. mass transfer, heat transfer). Moreover, the database contains characteristic parameters describing the apparatus, as e.g. heat transfer coefficients for advanced heat exchangers, and referenced relevant literature on experimental demonstration and possible applications. First screening of the database is performed in order to identify feasible technologies for tackling the target bottleneck, considering criteria such as feed and product stream phases, as well as constraints on process operating conditions. Additionally, different methods and tools are applied for the determination of the specific physical/ chemical phenomena responsible for the bottleneck (e.g. heat transfer, mass transfer, reaction equilibrium as possible reasons of the bottleneck- reaction selectivity). For instance, based on simplified correlations proposed by Klaewkla et al. [5], the influence of mass transfer limitations on the reaction selectivity is evaluated. If the mass transfer is not limiting selectivity, options for mass transfer improvements are further omitted. One potential option for intensification that may be found by means of such analysis is a reaction with in-situ product removal. In order to identify feasible separation processes that can be coupled with the reaction the thermodynamic insight approach [6] is applied. In case a mass separating agent is required, potential solvents are screened by means of a computer aided molecular design method (CAMD) [7], taking into account restrictions from the reaction and separation.

Based on the results the bottleneck is characterized by specific phenomena responsible for it. On this basis feasible process options are again determined from a database screening. The remaining options are further evaluated by means of short-cut models to evaluate the feasibility and performance improvement of the intensification. Based on these estimates a preselection of the most promising intensified alternatives is performed in an early design phase, at which available data is commonly sparse.

The proposed debottlenecking methodology is applied for the case study of ethylene oxide production. At first, the limitations of the analyzed process are identified based on a detailed simulation of a base case design for the oxygen-based process variant, representing the current state-of-art technology [8]. The major bottlenecks of the process are identified in respect to improvements in operational cost per unit of product. Applying the sensitivity analysis of mass and energy efficiency factors for every piece of equipment, the main process bottleneck is identified to be the selectivity of the reactor towards production of the ethylene oxide. The second process bottleneck is the ethylene loss on the purge stream, being part of the downstream processing. Taking into account the given bottlenecks, the PI database is screened for feasible options. For reactor improvement, the process conditions (temperature> 200°C, pressure> 10 bar, reacting phase- gas) are specified to identify suitable equipment. Furtherly, the reasons for the limited selectivity are studied by evaluation of the impact of heat transfer, mass transfer, reactants concentrations, and reaction equilibrium limitations.

By means of this screening procedure, reaction with in-situ separation is identified as a promising alternative for tackling the selectivity bottleneck. Absorption is identified as potential separation technique based on the application of the thermodynamic insight method [6]. A possible mass separation agent is derived by means of CAMD [7]. The search procedure is repeated for the second bottleneck, which is the ethylene loss on the purge stream. A membrane separation unit is determined as one of promising options for the reactant recovery. The potential of process improvement is quantified by means of simplified models and in the end validated by detailed modelling.

Overall, the presented approach provides a systematic tool for identifying the relevant process bottlenecks and the determination of the most promising intensification options, combining database search, thermodynamic and kinetic studies, as well as a quantitative assessment for the most promising options. Experimental effort and detailed modelling can consequently be focused on such options, offering the highest potential for improvements. The method can therefore be interpreted as support tool for the identification of the PI alternatives which should be investigated in more detail.

References:

[1] Stankiewicz, A.; Moulijn, J.; Process intensification: transforming chemical engineering, Chem. Eng. Prog. 2000, 96 (1), pp. 22-33.

[2] Keil, F; Modeling of process intensification. Weinheim, Chichester: Wiley-VCH; John Wiley 2007, pp. 1-9

[3] Becht, S.; Franke, R.; Geißelmann, A.; Hanh, H.; An industrial view of process intensification, Chem. Eng. Process 2009, 48, pp. 774-780

[4] Lutze, P.; Gani, R.; Woodley, J.; Process intensification: A perspective on process synthesis, Chem. Eng. Processing: Process Intensification 2010, 49, pp.547-558

[5] Klaewkla, .; Arend M.; Hoelderich, W. F.; A Review of Mass Transfer Controlling the Reaction Rate in Heterogeneous Catalytic Systems, in: Mass Transfer - Advanced Aspects, Ed, Hironori Nakajima, InTech,2011, pp. 667-684

[6] Jaksland, C.; Gani, R.; Lien, K.; Separation process design and synthesis based on thermodynamic insights, Chem. Eng. Sci. 1995, 50, pp. 511-530

[7] Constantinou, L.; Bagherpour, K.; Gani, R.; Klein, J.A.; Wu, D.T.; Computer aided product design: problem formulations, methodology and applications, Comput. Chem. Eng. 1996, 20(6-7), pp. 685-702

[8] Rebsdat S, Mayer D. Ethylene Oxide. Ullmannâ??s Encyclopedia of Industrial Chemistry, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, 2012, vol. 13, pp. 547-572