Towards Economically and Environmentally Optimal Operations in

Towards Economically and Environmentally Optimal Operations in Natural Gas Based Petrochemical Sites. Antonio Gonzalez-Castaño, J. Alberto Bandoni, a...
0 downloads 3 Views 2MB Size
Subscriber access provided by UNIV OF DURHAM

Article

Towards Economically and Environmentally Optimal Operations in Natural Gas Based Petrochemical Sites Antonio Gonzalez-Castaño, J. Alberto Bandoni, and Maria Soledad Diaz Ind. Eng. Chem. Res., Just Accepted Manuscript • DOI: 10.1021/acs.iecr.7b04598 • Publication Date (Web): 25 Jan 2018 Downloaded from http://pubs.acs.org on January 30, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Industrial & Engineering Chemistry Research is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

Towards Economically and Environmentally Optimal Operations in Natural Gas Based Petrochemical Sites Antonio González-Castaño, J. Alberto Bandoni, M. Soledad Diaz* Planta Piloto de Ingeniería Química (PLAPIQUI) CONICET – Universidad Nacional del Sur. Camino La Carrindanga km 7, Bahía Blanca (8000), Argentina. *[email protected]

In this work, we address the economic and environmental optimization of the operations in a natural gas based petrochemical complex through multiobjective optimization. The site is represented by a mathematical model that includes linear and non-linear simplified models for single equipment and plants to calculate site production levels, taking into account main operating variables, utility consumption, intermittent delivery, inventory level profiles and product distribution. The resulting multiperiod, mixed integer non-linear programming (MINLP) models are implemented in GAMS. The bi-criteria MINLP model is solved with the ε-constraint method. The obtained Pareto-optimal curve shows the trade-off between the economic and environmental aspects to pursuit sustainable operations of an existing integrated petrochemical complex. The environmental metrics is Global Warming Potential. ReCiPe is used for life cycle assessment and its eighteen mid-point impact indicators and three end-point indicators are further analysed for the petrochemical complex. The optimal solutions show that the environmental performance of the industrial cluster can be improved by a different distribution of processed raw material and intermediate products between the plants, as well as the means of transportation.

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1. Introduction The petrochemical industry provides numerous commodities for modern society and much effort has been devoted to optimizing production processes and economic performance during the last decades. Sahinidis et al.1 proposed a mixed integer linear programming (MILP) problem for the optimal selection and expansion of chemical plants, addressing varying demands and prices in long term planning. Turkay et al.2 applied logic-based approaches proposed by Turkay and Grossmann3 to solve a mixed integer linear problem (MILP) for the total site optimization of a petrochemical complex. Bok et al.4 proposed a bilevel decomposition strategy for the solution of a multiperiod MILP model for the supply chain of continuous processes. Jackson and Grossmann5 proposed Lagrangean decomposition techniques for the solution of a non-linear programming (NLP) problem for multisite production planning of production, transportation and sales in a chemical company. Schulz et al.6 proposed an integrated multiperiod optimization model for the supply chain of an existing large-scale petrochemical complex with short grid spacing. The mathematical model accounts for supply, production and product multimodal delivery. More recently, a few authors have addressed modeling and optimization of the gas-based industry and supply chain, fostered by the increasing availability of shale gas7. Cafaro and Grossmann8 addressed the long-term planning, design, and development of the shale gas supply chain network as an MINLP problem. He and You9 studied ethylene production from shale gas and bio-ethanol by Aspen Hysys simulation and stochastic optimization. Onel et al.10 and Niziolek et al. 11 studied the optimal plant design for olefins production and for aromatics production from natural gas and by using mathematical programing, respectively.

ACS Paragon Plus Environment

Page 2 of 45

Page 3 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

The application of multiobjective optimization to the petrochemical industry was addressed by Sophos et at.12 taking into account maximization of thermodynamic availability change, minimization of lost work and minimization of feedstock consumption. They solved biobjective optimization problems applying both the weighted single objective function approach and the ε-constraint approach. Furthermore, life cycle assessment (LCA) was used as a descriptive tool for assessing environmental performance while determining the main sources of impact in a process or product life cycle. In this context, Azapagic and Clift13 proposed for the first time the integration of LCA into multiobjective optimization as an effective tool to expand the capabilities of LCA strategies. The inclusion of environmental objective function was addressed by Al-Sharrah et al.14 for production planning in the petrochemical industry as a single objective MILP problem. In 2005, Hugo and Pistikopoulos15 proposed the application of multiobjective optimization to the longrange planning and design of supply chain networks. Eliceche et al.16 included LCA tools in the optimization of the utility plant from a large-scale petrochemical plant, as an MINLP problem. Gebreslassie et al.17, applied multi-objective optimization, to include life cycle assessment to the design of absorption cooling systems. More recently, Gebreslassie et al.18 addressed the optimization of the design and operations of a hydrocarbon biorefinery, also taking into account environmental and economic objective functions, formulating multiobjective Nonlinear Programming (NLP) problems. Pieragostini et al.19 provided an extensive review on the state of the art of multiobjective optimization in chemical processes. Guillén Gosálbez et al.20 addressed the design of hydrogen supply chains for vehicle use considering economic and environmental objective functions. The design problem was formulated as a bi-objective MILP, which simultaneously accounted for the minimization of cost and environmental impact. The environmental impact was measured

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

by the contributions to climate change (through its characterization factor, Global Warming Potential) by the hydrogen network operation. Furthermore, Guillén-Gosálbez and Grossmann21,22 proposed the use of mathematical programming for the minimization of the LCA impact, by introducing a robust optimization framework. More recently, Sabio et al.23 performed multiobjective environmental and economical optimization on an industrial process network including uncertainty in the LCI data related to the main feedstock. Zhang et al.24 addressed supply chain optimization within a multiobjective framework, using cost, GHG emission and lead time as performance indicators and applied it to a real-world industrial case. González-Castaño et al.25 proposed a multiperiod MINLP model for the multiobjective optimization of a petrochemical complex, taking into account linear inputoutput models for each plant in the site, with the environmental metrics obtained with Eco Indicator 9926,27. In a more recent paper, González-Castaño et al.28 proposed increasingly detailed models for the plants in the industrial cluster and used ReCiPe29 damage models for environmental impact estimation; steam and electricity consumption were obtained from the literature for similar technologies. Yang and You30 compared the production of ethylene and propylene from natural gas liquids-rich shale gas and naphtha, assessing economic and environmental performance, on the basis of process modeling and HYSYS simulation for three process designs. In this work, we propose a multi-objective multiperiod optimization model for the operations of a large-scale petrochemical complex6, considering both economic and environmental objectives. To our knowledge, it is the first time that the operations of an entire integrated petrochemical site, including process unit models, feedstock/product storage and transport, are simultaneously optimized for both objectives. Contributions to the eighteen ReCiPe29 mid-point impact categories are further analyzed for the

ACS Paragon Plus Environment

Page 4 of 45

Page 5 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

petrochemical complex, which is also a novelty of the paper. The site comprises two natural gas processing plants, two ethylene plants, a caustic soda and chlorine plant, a vinyl chloride monomer plant, a polyvinyl chloride plant, three polyethylene plants, an ammonia and a urea plant. As compared to our previous work, we have included nonlinear equations for the description of the operations in main plants, as well as correlations for utilities calculation. Nonlinear mathematical models were derived for most of the plants, based on rigorous existing models tuned with actual plant data. Simplified models consider variations in production with key plant operating variables, such as temperature and pressure in separation units. Available yield data for chemical transformations and utilities consumption are used to model the rest of the petrochemical complex. The objective functions are maximization of gross profit and minimization of global warming potential (GWP, the characterization factor for the climate change impact category), subject to constraints on mass and energy balances, bounds on product demands, equipment capacities and intermediate and final product storage tanks limitations. In addition, there are constraints on the final products distribution by ship, train or truck while storage tanks capacities are satisfied. Environmental metrics are obtained from ReCiPe29 according to the life cycle assessment procedures. The resulting problem is a multiperiod mixed integer nonlinear programming (MINLP) optimization problem, the model was implemented in GAMS 24.4.5.31 The minimum GWP case and the maximum profit cases are further compared in terms of the end-point impact damage categories.

2. Petrochemical complex description In this work, we study a large-scale petrochemical complex that includes natural gas processing plants, which extract ethane from natural gas; ethylene plants, several

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

polyethylene plants, vinyl chloride monomer and polyvinyl chloride plants, soda plants, ammonia and urea plants. Natural gas is fed through different pipelines and main products from the site are delivered by ship, trucks and railway. Figure 1 shows a simplified scheme of the entire process. One natural gas processing plant (NGI) is located next to the complex, while the demethanizing sector of a second natural gas plant (NGII) is 600 km away from the site, not far from gas wells. In this way, NGII has three sectors: methane separation, NGL (natural gas liquids) transportation through a liquid pipeline and a conventional NGL fractionation plant, next to the complex, as well as a gas pipeline for ethane transportation to ethylene plants. In both natural gas plants, methane is recompressed to pipeline pressure and mostly delivered as sales gas, while a side stream is used as feedstock for ammonia production within the complex. Pure ethane, propane, butane and gasolines are natural gas plants products. Ethylene is produced in two ethylene plants by steam cracking. Ethylene is a fundamental building block for the synthesis of bulk chemicals and intermediates, such as polyethylenes, ethylene oxide, 1,2-dichloro-ethane, vinyl chloride monomer (VCM). There are three polyethylene plants, one VCM plant and a polyvinyl chloride (PVC) plant, as well as a sodium hydroxide plant. The ammonia plant produces ammonia, which is in turn used as feedstock for urea production.

ACS Paragon Plus Environment

Page 6 of 45

Page 7 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

Figure 1. Schematic representation of petrochemical complex Natural gas processing plants The natural gas processing plants in the complex are based on turboexpansion (TE) and they are currently the most efficient ones for obtaining high ethane recovery due to the fact that cryogenic temperatures favor high efficient demethanizing. In a typical ethane extraction plant by TE, inlet gas is filtered and compressed. It is then air cooled and dehydrated to avoid ice and hydrates formation. After conditioning, the feed stream is sent to several parallel cryogenic trains for demethanization. The bottom products from the demethanizers are mixed and sent to a conventional separation train to obtain pure ethane, propane, butanes and natural gasoline. After heat exchanging with inlet gas, the top product

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

from the demethanizers (residual gas) is recompressed to pipeline pressure and delivered as sales gas. The cryogenic sector is the characteristic part of TE plant. In a Basic Turboexpansion Process, inlet gas to the cryogenic sector is cooled by heat exchange with residual gas and demethanizer side and bottom reboilers. The partially condensed gas feed is sent to a high-pressure separator (HPS). The vapor is expanded through a turboexpander to obtain the low temperatures required for high ethane recovery and is then fed to the top of a demethanizer column. The liquid from the high-pressure separator enters the demethanizer at its lowest feed point. Carbon dioxide and ethane distribute between top and bottom streams. Heavier hydrocarbons are obtained as bottom product (NGL). Figure 2 shows the cryogenic sector of a Gas Subcooled Turboexpansion Process. In this modification to the Basic TE scheme, a fraction of the vapor from the high pressure separator is condensed and subcooled by heat exchange with residual gas coming from the demethanizer. The remaining vapor is expanded through the turboexpander and fed to the middle of the column. The subcooled liquid is flashed and can be directly fed to the top of the demethanizer column, as reflux, thus enhancing ethane recovery. The liquid stream from the high pressure separator can be fed to a lower stage of the demethanizer or mixed with the subcooled liquid stream and fed to the top stage. Diaz et al.32,33 embed rigorous TE models within a superstructure to determine the optimal scheme for different natural gas mixtures. Within the complex, NGI has two basic TE trains and a gas subcooled train, while NGII has both trains with a gas subcooled technology. Feed gas composition has been considered as constant for each plant.

ACS Paragon Plus Environment

Page 8 of 45

Page 9 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

Figure 2. Simplified flowsheet of the cryogenic sector of a TE plant. Gas subcooled process Ethylene plants Both ethylene plants are based on steam cracking, being EPII a more recent technology. There are several furnaces in parallel and their feed also includes an ethane recycle stream. In this process, the main components of the outlet stream from furnaces are: hydrogen, methane, acetylene, ethylene, ethane, propylene, propane, butanes, butylene and pentanes. This stream is compressed in a cracked gas compressor and acetylene is hydrogenated to ethylene before entering the separation train. The first column in the separation train is the demethanizer: the top product stream is off-gas, which is used, in this scheme, as fuel gas in furnaces, and the bottom stream is further fractionated to obtain pure ethylene, ethane, propane, propylene, butane and gasoline. The ethane stream from the ethane-ethylene column is partly recycled to the furnaces.

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Polyethylene, VCM and PVC plants Ethylene is raw material for three polyethylene plants (PEI, PEII, PEIII) and a vinyl chloride monomer (VCM) plant, whose product is raw material for a polyvinyl chloride (PVC) plant. Ammonia and Urea plants Ammonia plants use methane and air as raw materials to produce nitrogen and hydrogen, which are reactants for ammonia synthesis. Main processes include sulfur removal by catalytic hydrogenation (and hydrogen sulfide adsorption and removal); catalytic steam reforming to obtain hydrogen and carbon monoxide; catalytic shift conversion to convert carbon monoxide to carbon dioxide and hydrogen; carbon dioxide is removed from the process stream by amines treatment. Finally, pure hydrogen is obtained in the catalytic methanation step and submitted as raw material to the ammonia synthesis reactor (HaberBosch process). Carbon dioxide is later used in the urea production process.

3. Methodology In this work, we consider given demands and specifications for the following final products in the petrochemical complex described in the previous section: propane, propylene, butane, pentane, ethylene, three types of polyethylenes (PE1, PE2, PE3), VCM, PVC, sodium hydroxide, ammonia, urea; as well as economic and environmental data for the entire complex. The objective is to determine optimal operating conditions, taking into account production planning (production rates, operating conditions), responses to demands, product distribution, inventory level management and life cycle impact assessment.

ACS Paragon Plus Environment

Page 10 of 45

Page 11 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

We formulate a mixed-integer nonlinear program (MINLP), in which continuous variables correspond to operating conditions, streams flowrates and compositions, energy consumption and economic and environmental performance metrics. We have included binary variables to represent intermittent product delivery by ship. Main equations include mass balances in process units or sectors, correlations for process units and utility consumption.

3.1 Mathematical Model 3.1.1Natural gas processing plants Feed gas flowrate to each natural gas processing plant is a variable and the individual flows are calculated as:  , =  ,  = ,

; ∀;  = 1 … 

We use linear correlations (Eqn. 2) for ethane (η

t

i, C2H6)

(1)

and carbon dioxide (η

t

i, CO2)

recovery in the demethanizer bottom stream, as function of main operating variables in the cryogenic

sector

(operating

temperature

in

high-pressure

separation

tank

and

demethanizing column top pressure). These correlations are based on simulations with a rigorous tailor-made plant model that was tuned with plant data,33 for similar conditions to the base case operating points6.  , = ,  + ,  + , ; , , < 0;  = ,

;  =  ,  ! ;  = 1 … 

(2)

Product flows (foi,jt) are calculated as:    ", = , ∙ ,  = ,

;  =  ,  ! ;  = 1 … 

where fii,jt is inlet flowrate.

ACS Paragon Plus Environment

(3)

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 45

We assumed 100% recovery for propane, butanes, pentanes and hexane. The demethanizer top stream (composed of methane and nitrogen) is calculated as the difference between the feed gas stream and the bottom stream. Mass balances in mixers and splitters Mass balances in each multicomponent stream splitter s are calculated as:

$

  %&'( )  = %&/ )  ∀; ∀% ∈ 0;  = 1 … 

(4)

$

  %&'( ) = %&/ ) ∀% ∈ 0;  = 1 … 

(5)

 $ %&)  = 1 ∀% ∈ 0;  = 1 … 

(6)

'(∈+,-.

'(∈+,-.



   %&'( )  = %&'( ) %&)  ∀; ∀% ∈ 0; ∀"1 ∈ 0'( ;  = 1 … 

(7)

   %&/ )  = %&/ ) %&)  ∀; ∀% ∈ 0;  = 1 … 

(8)

where Sout is the set of outlet streams and component j represents carbon dioxide, ethane, propane, butanes, pentanes and hexane. Mass balances in mixers are modeled as:   $ 2 / 3  = 2 '( 3  ∀; ∀2;  = 1 … 

(9)

  $ 2 / 3 = 2 '( 3 ∀; ∀2;  = 1 … 

(10)

/∈+45

/∈+45

where Sin is the set of inlet streams to mixer m. In NGII, the bottom stream from the demethanizer (NGL) is pumped along a pipeline to the NGL fractionation plant, requiring several pumping stations.

ACS Paragon Plus Environment

Page 13 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

Mass balances in multicomponent storage tanks There are multicomponent storage tanks at both ends of the NGL pipeline in natural gas plant NGII. Mass balances are as follows:  67,

=

' 67,

+



/ $ 6/,7, /9:



/ − $ 6'(,7, ∀; ∀;;  = 1 … 

(11)

/9:

Equations 12-15 ensure that composition inside the tank is equal to the outlet stream composition. Lower and upper bounds are imposed on tank volume.