• Author / Uploaded
• Maria Paula

Citation preview

Accepted Manuscript Title: Economic and environmental analysis of the cumene production process using computational simulation Authors: Pedro G. Junqueira, Patrick V. Mangili, Rafael O. Santos, Lizandro S. Santos, Diego M. Prata PII: DOI: Reference:

S0255-2701(18)30319-2 https://doi.org/10.1016/j.cep.2018.06.010 CEP 7312

To appear in:

Chemical Engineering and Processing

Received date: Revised date: Accepted date:

14-3-2018 9-6-2018 11-6-2018

Please cite this article as: Junqueira PG, Mangili PV, Santos RO, Santos LS, Prata DM, Economic and environmental analysis of the cumene production process using computational simulation, Chemical Engineering and Processing - Process Intensification (2018), https://doi.org/10.1016/j.cep.2018.06.010 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Economic and environmental analysis of the cumene production process using computational simulation

1

IP T

Pedro G. Junqueira1, Patrick V. Mangili1, Rafael O. Santos1, Lizandro S. Santos1*, Diego M. Prata1 Department of Chemical and Petroleum Engineering, Universidade Federal Fluminense, 24210-240, Niterói, RJ – Brazil.

SC R

*Corresponding author: Lizandro S. Santos (E-mail: [email protected])

Highlights

U

• Simulation of five different cumene production processes.

N

• Simulation of utility plants for more realistic results.

A

• Development of six different categories of eco-indicators.

M

• Eco-indicators lumping in a single index for global eco-efficiency comparison. • Economic analysis by estimating the processes’ gross annual profits.

PT

Abstract

ED

• The heat integrated process proved to be the most economical and sustainable.

The need for mitigating environmental impacts has been heading towards the development of

CC E

new technologies that could lead the industries to a greater ecological efficiency. Countless methodologies for evaluating processes’ efficiency in relation to their respective ecological footprint have been employed, among which the concept of eco-efficiency has stood out. In this context, this work aims to compare six different cumene production plants, namely

A

conventional, transalkylation, heat-integrated, dividing wall column, reactive distillation and double-effect distillation technologies, with regard to their economics and environmental efficiencies. The economic analysis was performed by estimating their respective Gross Annual Profit (GAP) and determining the specific production costs indicator, whereas the environmental assessment was carried out by calculating six eco-indicators, namely raw materials consumption, fuel consumption, energy use, CO2 emissions, water consumption and wastewater generation. The processes’ environmental performances were then compared

2 through the Eco-efficiency Comparison Index (ECI) method. The study showed that the intensified processes are not only more economically attractive but also more environmentaly friendly.

Keywords: Computational Simulation, Cumene, Eco-efficiency, Process Economics, Process Intensification.

IP T

1. Introduction

The man-nature relationship has significantly changed as a result of the manufacturing activities

SC R

derived of the first industrial revolution in the 18th century. Greenhouse gases emissions have increased due to the exploitation of natural energetic resources and, as the years went by, the records concerning the global annual mean temperature have increased. In fact, the World Meteorological Organization (WMO)’s report1 indicated that 2016 was not only the warmest

U

year ever recorded but also presented the higher concentration of carbon dioxide in the

N

atmosphere. The European Commission2 has released a global list of the countries with the greatest CO2 emissions, with China being its top emitter with over 10 million kilotons in 2015,

A

almost twice as much as the United States (second major contributor).

M

Governments have established sustainable development politics with restrict environmental laws as an effort to mitigate the ecological impacts resulting mainly from industrial activities.

ED

Hence, industries have been developing more energy-efficient plants, with reduced atmospheric emissions, raw materials consumption and waste generation in order to minimize their ecological footprint. In this regard, Stankiewicz and Moulijn3 stated that new design techniques

PT

are being developed with the intention of increasing production, reducing energy demand and minimizing costs, which is the case of process intensification approaches such as reactive

CC E

distillation columns, ultrasound assisted equipment and multifunctional reactors, among others. In addition, Wang and coworkers4 pointed out that such approach represents an advantageous strategy not only in terms of profit and sustainability but also regarding safety, since more

A

compact process configurations provide smaller volumes of toxic and flammable inventories. For these reasons, process intensification has gained more and more attention of the industrial sector. For instance, the European Comission5 has been fomenting the development of solutions for sustainably improving the production of chemical goods. The U.S. Department of Energy 6, in turn, has recently announced the plan of developing novel and breakthrough technologies in process intensification with the purpose of enhancing energy efficiency through the Rapid Advancement in Process Intensification Deployment (RAPID) Institute.

3 In this context, the purpose of this paper is to demonstrate the relevance of process intensification to the chemical industry by quantitatively and qualitatively comparing the economics and environmental impacts of six different cumene production plants. To the best of our knowledge, cumene processes have only been studied in terms of either their economics or single environmental impacts. Therefore, their ecological performance was evaluated by considering several impacts in order to provide the scientific community with more relevant data. In fact, the environmental analysis was performed by determining the raw materials

IP T

consumption, fuel consumption, energy use, CO2 emissions, water use and wastewater generation eco-indicators. The economic evaluation, in turn, was carried out by estimating the Gross Annual Profits (GAPs) of each technology, as well as the specific production costs

SC R

indicators. The ecological and economic indicators were subsequently normalized and plotted in a radar-type chart in order to determine the most environmentally friendly process by means of the Eco-efficiency Comparison Index (ECI) method.

In the next section, a brief revision regarding the cumene manufacturing processes is presented.

U

In Section 3, we describe the method for calculating the eco-indicators, whereas the

N

assumptions to estimate the costs and the economic indicator are provided in Section 4. In Section 5, we discuss the methodology required to establish the comparison between the six

A

technologies. Finally, in Section 6, we summarize the main results and provide some suggestion

ED

2. Cumene Production

M

for future works.

Cumene is mostly used as an intermediate for the production of acetone and phenol, used in the manufacture of several products such as bisfenol-A, polycarbonate and epoxy resins, nylon 6,

PT

among others. Due to such aspects, the global demand for cumene has increased significantly, being the Asia-Pacific the leading with appoximately 43% of the overall market revenue7,8.

CC E

The cumene production technology was originally developed during World War II to meet the demand for aviation gasoline, with which it was blended to improve octane rating8,9. Cumene is currently among the world’s five biggest large-scale productions, along with ethylene,

A

propylene, benzene and ethylbenzene10. Cumene is produced by the Friedel-Krafts alkylation of benzene with propylene, an irreversible and exothermic reaction, as represented by Eq. (1). Further alkylation may happen due to the reaction conditions and result in the undesirable formation of p-diisopropyl benzene (DIPB), as shown in Eq. (2). Nevertheless, Pathak et al.11 stated that the formation of DIPB can be partially overturned by using a transalkylation reactor in which DIPB is converted back to cumene, as shown in Eq. (3). Industrially, cumene is obtained in a gas-phase packed bed reactor under high temperature and pressure over an acid catalyst12.

4

C6 H 6  C3 H 6  C9 H12

(1)

C9 H12  C3 H 6  C12 H18

(2)

C12 H18  C6 H 6  2C9 H12

(3)

In the literature, the cumene production process was originally proposed by Turton et al.13 and has been optimized by several authors in order to improve its economics and sustainability. Numerous studies regarding, for instance, the use of different catalysts14, the reduction of raw

IP T

materials flow rates15, the increase in overall conversion by using a transalkylator16 and the

implementation of dividing wall column17, as well as reactive11,18 and double-effect distillation19

SC R

technologies are currently available.

The cumene production plants studied in this paper were simulated in AspenTech’s Aspen Plus® V8.8 software under steady state conditions. In order to obtain consistent results, the same feed conditions (temperatures, pressures, compositions and flow rates), equipment

U

specifications (sizes, reaction kinetics and efficiencies) and thermodynamic model (Peng-

N

Robinson Equation of State) used by the reference authors were considered. The simulation of all processes could be performed since said authors provided sufficient information for

A

reproduction of their respective works and results.The following sections briefly describe each

M

process, whereas a more detailed description regarding equipment parameters and stream conditions is provided by the referenced papers.

ED

2.1 Conventional Technology

Turton et al.13’s design had a significant amount of raw materials that were eliminated in a purge column with the objective of aiding further separation and avoiding build-up at

PT

subsequent steps. Aiming to reduce such raw materials waste, Luyben15 investigated the advantages of increasing both the temperature of the reactor inlet stream (to improve the

CC E

reaction rates) and the reactor’s size (to obtain higher propylene conversion) as well as adjusting benzene recycle to the feed section. The author found that larger reactors not only increased propylene conversion but also required lower temperatures for almost complete (99%)

A

conversion of propene, thus improving process’ economics, despite the expenditures with higher reaction vessel costs. Figure 1 illustrates the optimized process studied in this work, as well as the results obtained from simulation.

U

SC R

IP T

5

N

Figure 1. Cumene conventional production process’ flow diagram.

A

In this process, pure benzene at 98.8 kmol/h is mixed with a propylene stream (containing 0.5 mol % propane, at 101.9 kmol/h). The resulting mixture is vaporized in vessel V1 and preheated

M

firstly in heat exchanger FEHE and then in the fired heater H1 prior to being sent to a tubular reactor R1, in which the reactions described by Eqs. (1) and (2) take place. The reactor has 1500

ED

tubes (0.0763 m diameter and 6 m length) filled with a solid catalyst of 0.5 void fraction and 2000 kg/m3 density. The reaction kinetics are presented in Table 1.

PT

Table 1. Reaction kinetics for the conventional processa. Reaction

Rate expression

Reaction 1 – Eq. (1)

r1 = 2.80 x 107 exp(-104181/RT) Cbenzene Cpropylene

CC E

r2 = 2.32 x 109 exp(-146774/RT) Ccumene . Cpropylene Reaction 2 – Eq. (2) a Cbenzene: Benzene composition. Cpropylene: Propylene composition. Ccumene: Cumene composition. Composition units: kmol . m-3. Reaction rate units: kmol . m-3 . s-1. R: 8.316 kJ . kmol-1.

A

The highly exothermic characteristic of the benzene/propylene conversion requires a boiler feedwater (bfw) cooling system, thus generating high pressure steam (hps) to be exported. Nonetheless, the process would be expected to have surplus heat after heat integration in heat exchanger FEHE. Therefore, the reactor product is subsequently cooled down in cooler HE1 prior to being sent to a flash vessel V2, in order to separate the fuel gas (containing mainly propane and non-reacted benzene at 9.9 kmol/h) from the process stream, which is sent to the separation/purification section that comprises two distillation columns. The first column (C1)’s top product, containing 94.8 mol % benzene at 106.1 kmol/h, is recycled via pump P1 to the

6 feed section. Cumene is obtained as top product in the second column (C2) at 93.7 kmol/h with a purity of 99.9 mol %, while DIPB is retrieved at the bottom at 2.9 kmol/h as by-product. Both terms “HE” and “HX” were used for heat exchangers in order to distinguish coolers (HE) from heaters (HX) since they are considered for the calculation of different indicators.

2.2 Transalkylation Technology Based on Luyben15’s work regarding an overall optimization of Turton et al.13’s design, Pathak

IP T

and coworkers11 proposed two other process configurations, namely transalkylation technology and reactive distillation technology (discussed later). In the former, a transalkylation reactor was added with the objective of recovering cumene loss over its further alkylation. Pathak et al.11

SC R

investigated a few key process variables that, when modified, presented significant outcomes in terms of capital and operating costs. The first trade-off was associated with the relationship between reaction conversion and reactor size, since although reducing the latter may result in a decrease in capital costs, the fomer would be undesirably lowered. The second parameter

U

studied was the increase in the reactor temperature, as analyzed by Luyben15, aiming at a higher

N

propylene conversion but with the downside of further alkylation. In this design scenario, DIPB formation should be as low as possible in order to minimize the transalkylator costs. Finally,

A

both the benzene feed and recyle were deemed critical for the process’ economics, since higher

M

benzene concentrations favor higher propylene conversions. Maity and coworkers16 further optimized Pathak et al.11’s transalkylation design with the aid of

ED

a built-in Aspen Plus’s V7.2 optimizer. The authors fixed some variables such as, for instance, the mole fractions of benzene/DIPB in the final product and propane mole fraction in the purge column’s bottom product (to guarantee non-accumulation in further steps). Figure 2 illustrates

A

CC E

PT

the transalkylation process studied in this work, as well as the simulation results.

M

A

N

U

SC R

IP T

7

ED

Figure 2. Cumene transalkylation production process’ flow diagram. In the process, pure benzene at 98.8 kmol/h and a 0.5 mol % propane-containing propylene stream at 105.3 kmol/h are mixed, vaporized in vessel V1 and preheated (in heat exchanger

PT

FEHE and then in fired heater H1) prior to being sent to a tubular reactor R1, in which the reactions described by Eqs. (1) and (2), whose kinetic data are present in Table 1, take place. The reactor has 150 tubes (0.05 m diameter and 1.764 m length) filled with a solid catalyst of

CC E

0.5 void fraction and 2000 kg/m3 density. Similarly to the conventional process, bfw is used to cool down the reactor, which generates hps to be exported. Here again, surplus heat would be expected after heat-integrating the reactor’s feed and product in heat exchanger FEHE and,

A

therefore, the reactor product is cooled down in cooler HE1 prior to being fed to column C1. From this column fuel gas containing mainly propane and propylene is removed at the top at 6.9 kmol/h. The bottom product is subsequently mixed with the transalkylator’s product and fed at 161.8 kmol/h to two distillation columns for product purification. The first column (C2)’s top product, containing 95.3 mol % benzene is recycled at 55.7 kmol/h via pump P1 to the feed section. Cumene is obtained as top product in the second column (C3) at 103.5 kmol/h with a purity of 99.9 mol %, while the bottom product is pre-heated in heater HX1 and sent at 2.6

8 kmol/h via pump P3 to a transalkilation reactor R2 where the reaction described by Eq. (3) takes place. The transalkylator has 100 tubes of 0.05 m diameter and 1.764 m length. The reaction kinetics for the transalkylation are presented in Table 2. The use of a transalkylator is aimed at converting DIPB back to cumene and thus increasing the overall conversion. The transalkilator product is recycled back to the purificication section.

Reaction

Rate expression

Reaction 3 – Eq. (3), direct

r1 = 2.53 x 108 exp(-100000/RT) xBz xDIPB

Reaction 3 – Eq. (3), reverse

2 r2 = 3.877 x 109 exp(-127240/RT) xCu

b

IP T

Table 2. Transalkylator reaction kineticsb.

SC R

xBz: Benzene molar fraction. xCu: Cumene molar fraction. xDIPB: DIPB molar fraction. Reaction rate units: kmol . m-3 . s-1. R: 8.316 kJ . kmol-1.

2.3 Two-Flash Energy-Integrated Technology

The energy use in the industrial sector represents a critical burden in regard to greenhouse gases

U

emissions, since it contributes to approximately 21% of the total CO2 emitted20. Balancing energy consumption and environment preservation is a continuous challenge to the industries in

N

order to maximize their economic benefit. In this context, thermal integration of processes

A

stands out as a useful technique for improving energy efficiency.

M

Flegiel et al.12, for instance, based their process design on Sharma et al.21’s work to apply energy integration in the cumene production process. The former authors implemented a Multi Objective Optimization (MOO) to explore the trade-offs among economic (total capital and

ED

utility costs), environmental (material loss) and safety objectives. Flegiel et al.12’s integration was based on the recompression of the product column’s top stream, in order to generate

PT

enough heat duty so it could be used to heat the recycle column’s reboiler, and the use of a gas expander in the reactor product stream to generate electricity to be used as partial energy source

CC E

for the compressor.

Sharma and coworkers21 suggested a change in the conventional process’ purge column by replacing it by one flash vessel, which resulted in great losses due to the high amount of gas in the vent (off-gas) stream. Flegiel et al.12 then proposed the use of two flash tanks in series, with

A

a cooler in between. Such modification resulted in a substantial reduction of material losses and therefore, propylene conversion increased from 95.97% to 97.91%. The example studied in this work, as well as the simulation results, are illustrated in Figure 3.

N

U

SC R

IP T

9

A

Figure 3. Cumene two-flash energy-integrated production process’ flow diagram. In the process shown in Figure 3, pure benzene at 350.0 kmol/h is mixed with a propylene

M

stream (containing 0.5 mol % propane, at 325.9 kmol/h) and the resulting mixture is preheated in heat exchanger FEHE prior to being sent to a tubular reactor R1, in which the reactions

ED

described by Eqs. (1) and (2) occur. The reactor has 13,500 tubes (0.0763 m diameter and 7.315 m length) filled with a solid catalyst of 0.5 void fraction and 2000 kg/m3 density. Said reactions kinetic data are present in Table 1. Most of the energy generated in the reactor is consumed by a

PT

bfw cooling system, which generates hps that is also assumed to be exported. Part of the reactor’s product remaining energy is depleted in expander E1 before it is heat-integrated with

CC E

the reactor’s feed in heat exchanger FEHE. The amount of energy converted to electricity is sufficient to avoid surplus heat. The reactor product is then sent to a series of two flash vessels (V1 and V2, with a cooler HE1 in between) in order to separate fuel gas (containing mainly propane and benzene at 20.9 kmol/h) from the process stream, which is sent at 482.6 kmol/h to

A

the separation/purification section that comprises two distillation columns. The first column (C1)’s top product, containing 75.4 mol % non-reacted benzene is recycled at 158.1 kmol/h via pump P1 to the feed section. The second column (C2)’s top product (containing 99.78 mol % cumene at 319.1 kmol/h) is pressurized in compressor K1 and used to heat the reboiler of the first column, while DIPB (99.9 mol %) is retrieved as side product at the bottom at 5.4 kmol/h.

2.4 Dividing Wall Column Technology

10 Broadly present in chemical plants, distillation columns are responsible for the largest fraction (up to 80% total) of energy consumption in process industries22. Efforts have been made to enhance process sustainability, such as finding optimal separation sequences, compacting multiple columns into one shell (dividing wall columns) and merging different unit operations in a single equipment (e.g. reactive distillation) - aiming at reduced ground space use and lower energy requirements11,23. In this context, dividing wall columns (DWCs) have gained more and more attention in the last

IP T

years due to their evident advantages regarding energy and capital costs, sustainability and

operating flexibility24. In the literature, DWCs have been studied by taking into consideration different approaches such as design and control25,26, process expenditures17, and ecological

SC R

footprint22, among several others.

With regard to cumene production, Zhai and coworkers17 proposed a dividing wall column design in order to reduce the process energy requirements. The authors replaced both the purge

U

and product columns by a single DWC. With the purpose of reducing the total annualized costs (TACs), the process parameters were optimized by manipulating certain variables such as the

N

conventional column’s reflux ratio, the DWC’s feed tray location, the number of stages in both

A

columns and the liquid/vapor interconnection flow rates. A sensitivity analysis was carried out to determine the optimum conditions and, as a final result, the process with the DWC operating

M

at 3.2 bar was deemed ideal for achieving the minimum TAC. Figure 4 illustrates the example

A

CC E

PT

ED

studied in this paper, as well as the results obtained from simulation.

Figure 4. Cumene dividing-wall column production process’ flow diagram.

11 In this process, pure benzene at 107.3 kmol/h is sent to a conventional distillation column C1 that is also fed by a stream resulting from the mixture of the products from a tubular reactor R1 and a transalkilator R2. The distillation column’s top product is divided into fuel gas, mainly comprised of propane at 2.9 kmol/h, and condensate, consisting mainly in cumene and benzene at 1.0 kmol/h. The column’s bottom product is fed to a dividing wall distillation column (DWC) C2 via pump P1 in order to separate cumene as top product, at 100.7 kmol/h with a purity of 99.9 mol %, from DIPB (bottom product, at 0.6 kmol/h with a purity of 99.6 mol %). The

IP T

distillation column’s intermediate product, containing mainly benzene at 496.3 kmol/h, is divided into two streams that are fed via pump P2 to each reactor, while the DWC’s

intermediate product, containing mainly DIPB at 21.1 kmol/h, is pre-heated in heater HX2 and

SC R

recycled via pump P3 to the transalkilator in order to increase the overall conversion. The

reactions kinetic data for the both tubular reactor and transalkylator are present in Tables 1 and 2, respectivelly. The reactor has 150 tubes (0.05 m diameter and 1.764 m length) filled with a solid catalyst of 0.5 void fraction and 2000 kg/m3 density, whereas the transalkylator has 100

U

tubes of 0.05 m diameter and 1.764 m length. Since the benzene/propylene conversion is highly exothermic, the reactor is cooled through a bfw system, thus generating high pressure steam,

column C1.

M

2.5 Reactive Distillation Technology

A

N

which is interpreted as being exported. Any surplus heat present is depleted in cooler E1 prior to

As previously stated, Pathak and coworkers11 proposed two cumene production plant

ED

configurations, namely transalkylation technology and reactive distillation (RD) technology. The former focused on increasing the overall conversion by recovering cumene loss due to its further alkylation in the tubular reactor. Nevertheless, such design strategy resulted in high

PT

energy duties and equipment costs. In view of such drawback, the authors suggested the implementation of a catalyst-packed RD column, acting as both a purge column and a reactor

CC E

(that involves both alkylation and transalkylation), followed by a conventional distillation column. With such design, top fed benzene encounters propylene along the column and the unreacted benzene is reboiled, forced back to the reactive zone.

A

The most significant variable investigated was the RD column’s operating pressure. Higher values resulted in a higher reactive zone temperature due to the higher bubble point on the reactive tray, resulting in higher reaction rates. For this reason, both catalyst costs and benzene losses in the vent could be reduced. On the other hand, since higher operating pressures may reduce the benzene-cumene relative volatility, higher reboiler duties are required to avoid benzene contamination at the bottoms. For optimization, a built in Aspen optimizer was used, fixing the product purity at 99.9% and 0.09 benzene mole fraction in the RD distillate vent. For a better understanding of the influence of the operating pressure, Pathak et al.11 compared

12 different designs at 2 atm, 3 atm and 4 atm, confirming that higher pressures result in more attractive economics. The example studied in this work, as well as the simulation results, are

SC R

IP T

shown in Figure 5.

U

Figure 5. Cumene reactive distillation production process’ flow diagram.

N

In this process, a pure benzene stream at 100 kmol/h and a propylene stream (containing 0.5

A

mol % propane, at 105.3 kmol/h) are fed to a reactive distillation column C1, in which the

M

reactions described by Eqs. (1), (2) and (3) occur. The reaction kinetics are shown in Table 3. Table 3. Reaction kinetics for the RD processc. Rate expression

ED

Reaction Reaction 1 – Eq. (1)

0.96 0.87 r1 = 6.98 x 105 exp(-63742/RT) Cbenzene C propylene

Reaction 2 – Eq. (2)

0.61 0.92 r2 = 4.00 x 109 exp(-79162/RT) Ccumene C propylene

CC E

PT

Keq = 6.52 x 10-3 exp(27240/RT) Reaction 3 – Eq. (3) c Cbenzene: Benzene composition. Cpropylene: Propylene composition. Ccumene: Cumene composition. Composition units: kmol . m-3. Reaction rate units: kmol . m-3 . s-1. R: 8.316 kJ . kmol-1. The use of reactive distillation is particulary interesting in this scenario due to the highly exothermic reactions that take place in the column. The main advantage relates to the direct heat integration between reaction and separation, that is, the reaction heat is used to evaporate ligh

A

ends and hence reduce the total energy requirements for separation. From the RD column fuel gas (containing mainly propane) is retrieved at the top at 6.3 kmol/h, while the reaction products (comprised mainly of cumene) leave at the bottom. Such products are fed at 133.5 kmol/h to a distillation column C2 in order to separate cumene (at the top) at 99.7 kmol/h, with a purity of 99.9 mol %, from the unreacted components, which are retrieved at the bottom at 33.8 kmol/h (comprising 20.8 mol % cumene and 79.2 mol % DIPB) and recycled via pump P1 to the RD column.

13

2.6 Double-effect Distillation Tehcnology In Nourouzi et al.19’s design, the liquid-phase alkylation of benzene on a zeolite catalyst was the object of study, using a four-bed reactor configuration and a separation train to recover the product, similarly to Hwang and Chen27’s work. The design optimization was based on an economic model involving grass root capital investment, operating and raw materials costs, revenue, plant life span, taxes and inflation. The optimization was performed by firstly minimizing the process’ utility requirements through topological changes and by secondly using

IP T

a statistical method to increase the net present value (NPV).

As the first change, Norouzi and coworkers19 proposed the injection of propylene in four

different spots along the reactor, in order to decrease the utilities consumption and maintain

SC R

benzene/propylene proportion. This is an important factor for controlling the reactor’s

temperature and also for preventing the formation of linear hydrocarbons. The authors found that, when propylene flow rate is reduced by one fourth, the benzene is also reduced to keep the

U

same proportion, thus expressively increasing the NPV.

As the second modification, Norouzi et al.19 aimed their efforts toward changing the benzene

N

column’s configuration by proposing a double-effect distillation system. The previous single

A

large column was divided into two columns, C-2 and C-2+. The operating pressures were fixed at 1 bar and 8 bar respectively, such that the C-2’s reboiler temperature is at least 10° C lower

M

than C-2+’s condenser temperature, for energy integration purposes. Double-effect distillation refers to a system in which the feed stream is evenly split into two streams that are fed to two

ED

columns operating at different pressures. The overhead product from the high-pressure column supplies heat for the low-pressure column’s reboiler. This heat integration has the main

PT

advantage of considerably reducing utilities requirements and, therefore, the operating costs. Depending on the feed compositions and the components’ volatilities, such strategy may halve the overall heat duty28. For this reason, double-effect distillation has been studied for several

CC E

applications such as n,n-dimethylacetamide-water separation29, ethanol production30, bioethanol distillation31, acetone-methanol separation32, among others. After establishing said topological changes, Norouzi et al.19 used statistical models for

A

analyzing the interactions of the economic function NPV with several variables, such as the alkylation reactor’s length, inlet temperature and recycle flow rate, as well as the flow rate of the recycle to the transalkylation reactor. A full factorial design and an analysis of variance (ANOVA) were used to find a NPV model that better described the process’ behaviour. The optimization was then performed by the Simplex method33 and a pinch analysis was carried out to further improve the process in regard to reducing utilities requirements. Figure 6 illustrates the example studied in this work, as well as the simulation results.

SC R

IP T

14

U

Figure 6. Cumene double-effect distillation production process’ flow diagram.

N

In the process, pure benzene at 100.0 kmol/h is mixed with part of a benzene recycle stream, which is subsequently pressurized. A 0.5 mol % propane-containing propylene stream at 105.0

A

kmol/h is pressurized and then evenly divided into four streams. The first stream is mixed with

M

the benzene feed and heated in heat exchanger FEHE prior to being fed to the first bed of a 1.3m diameter, 7 m length four-bed reactor R1, while the second stream is fed to the second bed. The first product is mixed with the third propylene stream and cooled in an intercooler HE1 in

ED

order to keep the reactor’s temperature below the bubble point. The resulting cooled mixture is fed to the third bed, while the fourth propylene strem is fed to the fourth bed. The kinetics for

PT

the reactions (1) and (2) that take place in the four-bed reactor are shown in Table 1. The reactor’s final product is cooled in FEHE. Once again, surplus heat would be expected after integrating the reactor’s feed and product in FEHE. Thus, the product is subsequently cooled in

CC E

cooler HE2 prior to being sent to a distillation column C1, from which fuel gas containing mainly propane is removed at the top at 5.0 kmol/h. The bottom product is mixed with the transalkylator’s product and split into two streams. The first, at 136.7 kmol/h, and the second, at

A

91.2 kmol/h, are fed to the low (C2a) and high (C2b) pressure sections of the double-effect distillation system, respectively. From C2a most of the benzene is retrieved at the top, at 60.1 kmol/h, whereas the bottom product is sent at 106.2 kmol/h to the cumene column C3, from which cumene is retrieved at the top, at 99.9 kmol/h, and DIPB is obtained at the bottom, at 6.3 kmol/h. From C2b, in turn, benzene is retrieved as distillate at 61.6 kmol/h and mixed with the C2a’s benzene product, while the bottom product is fed to C2a at 29.6 kmol/h. The benzene stream from the double-effect distillation system is split into a 88.8 kmol/h stream, which is recycled to the feed section, and a 32.9 kmol/h strean, which is pressurized and pre-heated prior

15 to being fed to a 0.7-m diameter, 1.3 length transalkylator R2 where the reaction described by Eq. (3) takes place. The kinetics are disclosed in Table 2. The transalkilator product is mixed with C1’s bottom product.

3. Environmental Analysis Global warming, excessive use of raw-materials and inefficient water distribution are some of the most recurring subjects in recent days. Farming-related activities and energy generation are

IP T

the main responsible sectors to be pointed out in terms of global warming, which can be observed from the increase of 20% in nitrous oxide (N2O), 40% in carbon dioxide (CO2) and

150% in methane (CH4) emissions between 1970 and 201434. The scenario becomes even more

SC R

obnoxious when it comes to water distribution, since the numbers indicate that almost 1 billion people lack access to safe water and more than 2 billion people do not have access to appropriate sanitation35.

In light of this context, numerous discussions regarding the world’s current climate situation

U

have been raised. The need for mitigating environmental impacts has been heading towards the

N

development of new technologies that could lead the industries to a greater ecological efficiency. In this regard, several methodologies to assess the environmental performance of

A

chemical processes have been developed. Jia et al.36 applied a Process Environmental

M

Performance Assessment (PEPA) to characterize and quantify the environmental impacts of two ethanol production routes. Koroneos et al.37 and Brito and Martins38 used a Life Cycle

ED

Assessment (LCA) approach to evaluated the ecological effects of biomass-derived hydrogen production processes and butanol production technologies, respectivelly. In fact, LCA is broadly applied to the sustainability assessment of industrial processes since it provides a wide

PT

range of ecological indicators to estimate environmental impacts. Vaskan et al.39 applied the Eco-indicator 99 framework to address the optimization of utility plants. Such indicator was also used by Guillén-Gozálbez et al.40 for selecting the optimal process flowsheet in order to

CC E

balance economics and environmental impacts. Jacquemin et al.41 reviewed how several LCA assessment methodologies (e.g. Eco-indicator 99, CML 2001, IMPACT2002, etc.) are used to evaluate sustainability of the processing industry. Ruiz-Mercado and coworkers42, in turn,

A

provided 66 ecological indicators useful to the analysis of the ecological performance of chemical processes, being most of them expressed per either sales revenue or product mass. Countless methodologies for evaluating processes’ efficiency in relation to their respective environmental burdens have been employed and, therefore, a noticeable absence of a standard approach would be necessary. In this regard, the international standard ISO 14045:201243 was formulated with the purpose of providing the guidelines required to carry out eco-efficiency assessments. The document defines eco-efficiency as an “aspect of sustainability relating the

16 environmental performance of a product system to its product system value” and proposes its quantification through the determination of eco-indicators. Such strategy was also suggested by the Economic and Social Commission for Asia and Pacific (ESCAP)44, which pointed out the significance of eco-indicators for assisting decision-making tasks (e.g. selecting a particular plant configuration or technological route with basis on its environmental impacts).

3.1 Eco-indicators Eco-efficiency indicators are defined as a measure of the ecological performance of a process in

IP T

relation to its product or service value, which is given by a relationship between an

environmental variable and an economic variable45,46. They are conveniently useful due to their potential to establish a realistic basis for defining objectives and thus assisting in the

SC R

identification of process areas where improvements are conceivable. Particularly for the

industrial sector, the main eco-indicators are related to water and raw material consumption, greenhouse gases emissions, energy consumption and waste generation47.

U

Eco-indicators have been used by numerous authors for the analysis of the ecological footprint

N

of different chemical processes. Silalertruksa and coworkers48 assessed the sustainability of sugar cane biorefinery by estimating total output value and greenhouse gas emission indicators,

A

whereas Chen et al.49 determined the global warming potential (GWP) of ethylene production

M

processes. Wang et al.50 applied the concept of eco-indicators to estimate both resource productivity and emission efficiency of the cement industry. For biorefirenies, SacramentoRivero51 described several sustainability indicators such as land use (SLU), raw materials cost

ED

ratio (RCR), health and safety ccomliance (HSC) and biotechnological-valorization potential (BVP), among others. For the petrochemical industry, Al-Sharrah and coworkers52 presented a

PT

methodology for evaluating the toxicity potential of processes through the determination of the median lethan concentration (LC50), median lethal dose (LD50) and threshold limit value (TLV) indicators. Mangili et al.45, in turn, used such metrics to compare two acetone-methanol

CC E

separation technologies with regard to their environmental performance. The eco-efficiencies of the previously described cumene production technologies are compared here through the determination of six indicators, namely raw materials consumption, fuel

A

consumption, energy use, CO2 emissions, water use and wastewater generation eco-indicators. The indicators applied in this work are conveniently useful when compared to the traditional LCA metrics (e.g. CML, Eco-indicator 99, ESEERCO, IMPACT2002, LIME, etc.) due to the latter’s drawbacks regarding the dependence on available, quality data and time-consuming methodologies. LCA methods are limited to the data that may not be readily available53 and, hence, gathering sufficient information can be quite arduous. Furthermore, some endpoint

17 methodologies such as Eco-indicator 99, for instance, rely on relatively narrow frameworks that are based on specific territories, which may lead to imprecise results54. In this paper, however, the indicadors can be determined from simulation results without undue work and take into account different territories, since the parameters used in the calculations (e.g. CO2 emission factors, utilities conditions, etc.) take the plants location into consideration. In addition, the metrics applied in this work consider economic aspects (i.e. production rate and production costs), which are usually neglected by LCA indicators55. The calculation procedures

3.1.1

IP T

are described below. Raw Materials Consumption

SC R

Raw materials consumption is of great significance to the industry since it contributes to a major portion of the overall process’ expenditures. According to Park and Behera56, it refers to the total amount of materials (either used for conversion or pre-fabricated) acquired by the company. Hence, the calculation of the raw materials consumption eco-indicator (RMCi) for the

U

processes studied was carried out by summing up the feed streams’ flow rates and dividing the

3.1.2

N

result by the total production rate (i.e. flow rate of cumene produced). Fuel Consumption

A

Intertwined to energy use – and therefore to CO2 emissions, due to the greenshouse gases

M

(GHGs) emissions resulting from combustion – the fuel consumption of process equipment must be taken into account when designing new processes. For the production plants analyzed

mng 

ED

in this paper, the total amount of fuel (natural gas) used in the process is determined by Eq. (4).

E 1  i NCVng ni

(4)

PT

Where ṁng corresponds to the natural gas mass flow rate, NCVng refers to the natural gas net calorific value (approximately 0.048 GJ/kg according to the International Energy Agency57) and

CC E

Ei and ni correspond to the required energy and the thermal efficiency of equipment i (fuel consumer), respectively. In this work we considered that only the utility plant’s boiler and the fired heaters consume natural gas. Both equipment were assumed to operate with 80% thermal

A

efficiency58,59.

The fuel consumption (FCi) eco-indicator was then determined by dividing ṁng by the total production rate of each process. 3.1.3

Energy Use and CO2 Emissions

Energy consumption is a frequent topic raised by experts when it comes to global warming, since it is directly associated with GHG emissions. Notwithstanding the current concerns, the world has headed towards an alarming situation due to the projections for world population

18 growth, which will result in a significant increase in energy consumption. For this reason, industrial energy use has gained more and more attention of specialists who study possible ways of reducing processes’ energy requirements. The energy use (EUi) and CO2 emissions (CO2Ei) eco-indicators represent a means to estimate the main energy consumption sources in a chemical process, thus allowing the implementation of energy-related optimization approaches in order to minimize such environmental burdens. These indicators were estimated in this paper with basis on the methodology described by

IP T

Mangili et al.45. The former was calculated by adding up the energy demands of all heating

sources for each process and subsequently dividing the result by their respective production

rate. The latter was determined by converting the respective equipment’s energy requirements

SC R

into CO2 emissions. Pumps and compressors were interpreted as electricity consumers operating at 75% efficiency, whereas heaters and reboilers operate with steam generated in the utility plant’s boiler, which has an efficiency of 80%45. The energy requirements for the cooling tower fan stacks were calculated from the procedures available in Perry60’s book. The convertion

U

factors for electricity and natural gas thermal energy were 0.0227 tCO2/GJ61 and 0.0561 tCO2/GJ62,

N

respectively. The lower electricity-CO2 emissions factor for Brazil is due to the fact that the country’s energy mix is predominantly comprised of hydroelectric plants. This value may vary

A

significantly depending on the country, thus influencing the calculation of the above-stated

M

indicators. In Section 6.1.3 the CO2 emissions are later evaluated by considering conversion factors from different countries with different conversion factors for electricity

ED

One should note that the CO2 indicator applied in this work differs from the GWP since the latter corresponds to a “from cradle to grave” LCA indicator and, therefore, also includes the CO2 equivalent emissions due to other environmental burdens such as hydrofluorocarbons,

PT

perfluorocarbons, etc. The detailed description regarding the guidelines and procedures required for the calculations is provided by Mangili et al.45. Water Consumption

CC E

3.1.4

Energy and water are directly related since the use of water/steam in cooling/heating applications are associated with energy requirements by the respective consumers. This inter-

A

dependency tends to increase in the coming years, which is mainly due to the increase in the global energy demand and expansion of the energy sector. For this reason, several methodologies, such as eco-efficiency evaluation, have been applied to the industry in an effort to estimate possible solutions to minimize energy consumption of industrial processes. The water consumption eco-indicators (WCi) for the cumene processes were estimated according to the procedure described by Mangili et al.45. All cooling/condensing equipment were considered to use cooling water, with the exception of both cooler HE1 and C1’s

19 condenser from Flegiel et al.12’s process, which were assumed to use chilled water (5° C – 15° C) due to their lower temperature requirements as described by said authors. The total cooling water demand from the coolers and condensers, as well as the total amount of water required to generate low, medium and high pressure steams in the boiler, and the total heating steam requirements in heaters and reboilers were estimated through simulation. The utility plant, illustrated in Figure 7, was designed with basis on both such calculations and heuristics provided by the authors. Cooling water was assumed to be supplied at 30° C and return at 45°

IP T

C13. Deaerator was assumed to receive water from the feed tank at 30° C and dispatch it to the boiler feed pump at 90° C, whereas the high pressure heater was considered to operate with

boiler’s bleed steam to heat up water to 160° C prior to sending it to the boiler63. The parameters

SC R

used for low (LPS), medium (MPS) and high pressure (HPS) steams were retrieved from Seider et al.64 and are disclosed in Table 4. Cooling water tower evaporative losses and deaerator vent were assumed to be 1%39 and 0.2%65 of the water inlet flow rate, respectively. Other losses and blowdown heuristics are detailed in the next section. To simplify, solids from the ion beds and

U

chemical injection were neglected in the simulation.

N

Table 4. Steam conditions. Temperature, ° C

LPS

254.0

MPS

185.5

HPS

135.0

A

CC E

PT

ED

M

A

Class

Figure 7. Utility plant flow sheet.

Pressure, bar 43.4 11.4 3.1

20 In Figure 7, LPS, MPS, HPS, BFP and CW stand for low pressure steam, medium pressure steam, high pressure steam, boiler feed pump and cooling water, respectively. 3.1.5

Wastewater Generation

The costs associated with waste treatment correspond to an important drawback for industrial activities due to strict environmental laws of waste management and nature protection. Waste streams require complex treatment processes, since toxic compounds cannot be irresponsibly discharged to the environment due to the harmful effects in several biosystems. Catalyst-packed

IP T

ionic columns and electricity-powered aeration are some examples of equipment that contribute to higher capital (CAPEX) and operating (OPEX) expenditures in industrial processes66. In this

SC R

regard, the minimization of waste generation is a common objective among the industries.

The total wastewater flow rate produced by each process was determined by summing up the process losses and equipment blowdown, according to Figure 8. The respective eco-indicator

A

CC E

PT

ED

M

A

N

U

(WGi) was then calculated by dividing the result by the total production rate of each process.

Figure 8. Wastewater discharge schematic diagram (adapted from the Smith’s book67). Condensate losses were assumed to be 20% of the total steam flow rate (80% condensate return), while losses in the boiler feed water treatment were considered to be 2% of the boiler feed water (BFW) flow rate. Boiler and cooling water blowdowns were assumed to be 3% of the circulating flow rate. All values stated above were retrieved from Mangili et al.45’s work, with the exception of cooling water losses (considered to be 1%68 of the total water flow rate required

21 by the consumers). To simplify, wastewater from freshwater consumers and rain water were assumed to be the same for all processes and, thus, were neglected in the calculations.

4. Economic Analysis The development of processes must undergo continuous economic evaluations in order to determine whether they are financially feasible or impracticable. Nevertheless, Silla69 stated that although estimations may show a potential profitability, it is possible that the capital

IP T

requirements for investment and operation “strain the financial capabilities” of the company. Therefore, the evaluation of the economic feasibility of a process must take into account the total capital (CAPEX) and operating (OPEX) expenditures.

SC R

CAPEX and OPEX correspond to the costs associated with initial expenses (e.g. purchasing and installing industrial equipment, acquiring land for construction, etc.) and operational

investments (e.g. raw material, utilities supply, etc.), respectively70. Thus, the estimation of such expenses is conveniently useful for assisting in decision-making tasks, especially with regard to

U

the selection of different production routes and plant configurations.

N

The economic analysis was carried out by calculating both the respective processes’ CAPEX

A

and OPEX and specifying a hypothetical payback period of 3 years11,15,17. The TAC for each

CAPEX  OPEX Payback period

(5)

ED

TAC 

M

technology was then determined from Eq. (5).

The CAPEX for each process was calculated with the aid of AspenTech’s Process Economic Analyzer (PEA) V8.8 with basis on the equipment specifications provided by the reference

PT

authors and updated currencies. Moreover, as exercized by Zhai et al.17, a penalty of 10% to both the DWC and RD column was assumed since the construction and installation of such

CC E

equipment may be more complex than the conventional columns. For the utility plant, the economic evaluation was performed by considering that pumps and compressors were of centrifugal type, whereas feed tanks, filter, deaerator and ionic beds were

A

assumed to be vertical pressure vessels. Cooling water tower and boiler costs were estimated with basis on the Leeper’s book71 and PEDCo Environmental, Inc.’s report72, respectively. High pressure heater was assumed to be of shell and tube type. All construction materials were interpreted as stainless steel 304. Regarding the OPEX, some other assumptions were taken into consideration, as shown in Table 5. From the TAC results and by estimating the processes’ revenue due to cumene production, as well as to fuel and hps credits, we could determine the gross annual profits (GAPs) of each technology. However, comparing the processes’ economics by only considering such variable

22 would not be the most appropriate strategy since they have different production capacities. Hence, in order to establish a more categorical comparison, a specific production cost indicator (SPCi) was used, which could be determined by dividing the TAC of each design and by their respective total annual production rate.

Plant operating time

8,000 h/year

________

Internal Rate of Return

15%

________

Project lifetime

20 years

________

Benzene cost

93.43 $/kmol

Sharma et al.21

Propylene cost

56.50 $/kmol

Sharma et al.21

Electricity cost

16.80 $/GJ

Natural gas cost

3.71 $/GJ

Cooling water cost

1.04 $/GJ

Chilled water cost

4.43 $/GJ

High pressure steam credit

12.33 $/GJ

Turton et al.13

Fuel gas and DIPB credit

6.00 $/GJ

Luyben15

Cumene selling priced

198.73 $/kmol

Sharma et al.21

9e

Turton et al.13

44,315 $/year

Turton et al.13

Supervisors per shift

2

Turton et al.13

Supervisor Cost (SC)

53,178 $/year

Turton et al.13

Number of shifts

5

Turton et al.13

Laboratory labor

20% of (OC + SC)

Silla69

Maintenance Labor (ML)

2% of CAPEX

Silla69

Operating costs

10% of ML

Sinnott70

Maintenance

2%/year of CAPEX

Sinnott70

Overhead expenditures

80% of labor costs

Silla69

Taxes and insurance

2%/year of CAPEX

Silla69

Depreciation

10%/year of CAPEX

Turton et al.13

Unscheduled equipment

10% of CAPEX

________

Operators per shift

CC E

Supply

PT

ED

Labor

M

Operator Cost (OC)

A

Other

IP T

Reference

SC R

Process

Value

U

Guidelines

Variable

A

Basis

N

Table 5. OPEX estimation assumptions.

Turton et al.13 Aspen Plus Basis Turton et al.13 Turton et al.13

Sales and R&D 3% of the total costs Sinnott70 General d All cumene product streams are approximately 99.9 mol % pure. Therefore, the same selling price can be assigned for all six technologies. e The number of operators per shift in the RD plant was assumed to be 5 since its configuration is much more compact than the other processes’.

5. Eco-efficiency Comparison Index The joint evaluation of several different eco-indicators not always results in a simple and easy way to rule out the least eco-efficient processes, since assuming which indicators are more

23 important for each scenario is not always possible. This difficulty was overcome by Pereira and coworkers47, who developed a methodology called “Eco-efficiency Comparison Index (ECI)”. Such methodology was applied by Mangili et al.45 to compare the eco-efficiencies of three different acetone-methanol separation processes. The strategy consists in normalizing the environmental and economic indicators by dividing them by the highest value of their respective category. The normalized indicators are then arranged in a radar type chart to form polygonshaped diagrams corresponding to the processes to be compared. The shape area (ST) is then

IP T

determined by summing up the areas (Sa) of the n minor triangles formed, which are calculated through the Law of Sines shown in Eq. (6), where lA and lB correspond to the adjacent sides A and B of the minor triangle, respectively. θ refers to the angle formed between lA and lB, and

SC R

corresponds to 2π divided by the number n of indicators evaluated, since all axis are equidistant from one another and set apart in a 360° circle45,47.

l A  lB  sen 2

(6)

U

Sa 

The ECI is calculated as described in Eq. (7), where ST* refers to the area of the largest shape

N

formed in the chart.

(7)

M

A

 S  ECI  1  T*  100%  ST 

The environmental performances of the cumene production processes were quantitative and

ED

qualitatively compared by means of the ECI method. Since all processes were optimized by the reference authors with basis on economic parameters, a fair and categorical analysis can be performed. Moreover, such evaluation strategy is singularly convenient since it allows the

PT

determination of the environmental impacts of industrial processes without the undue burden of re-designing the technologies to be evaluated. Hence, it is possible to appropriately compare

CC E

high-capacity processes (e.g. Flegiel et al.12’s) with lower-capacity designs (e.g. other technologies), since relative metrics are used. The same required procedure and formulae described by Mangili and coworkers45 were applied in this work.

A

6. Results and Discussion The simulation of the cumene production processes provided the required variables for calculating their respective indicators. The results are shown in Figures 1, 2, 3, 4, 5 and 6 which show that our results are indeed comparable to the authors’ and, therefore, demonstrate the plausibility of our study. Some differences regarding energy requirements were found when compared to the data provided by the reference authors. Such deviations might be due to the

24 different software used. With exception of Flegiel et al.12, which simulated the two-flash energy-integrated technology in Aspen HYSYS, all authors used early versions of Aspen Plus. One should note that, since the cumene product stream of all six processes is approximately 99.9 mol % pure, the indicators were equally determined in accordance with the methodology described in Section 3. The results for the environmental indicators are given in Table 6. Table 6. Eco-indicator results. Flegiel et al.12 38.36

Zhai et al.17 12.11

Pathak et al.11 11.96

Norouzi et al.19 12.24

Tofal feed flow rate, tmp/h

12.02

12.16

40.22

12.92

12.25

12.01

RMCi, t/t

1.0761

1.0227

1.0485

1.0669

1.0242

0.9812

Fuel flow rate, tfuel/h

0.69

0.63

0.49

FCi, tfuel/t

0.0618

0.0530

Energy demand, GJ/h

33.40

EUi, GJ/t CO2 emissions, tCO2/h

Production rate, t/h

H2O/t

Wastewater flow,

m3ww/h

0.0128

0.0289

0.0251

0.0327

30.62

26.97

16.86

14.31

19.45

2.9902

2.5753

0.7031

1.3922

1.1965

1.5891

1.87

1.72

0.1674

0.1447

12.05 1.0788 12.19

U

WCi, m

0.40

1.40

0.94

0.80

1.09

0.0365

0.0776

0.0669

0.0891

10.40

14.19

11.56

8.34

12.34

0.8747

0.3699

0.9546

0.6973

1.0082

10.53

15.43

12.57

8.36

12.70

0.4022

1.0380

0.6990

1.0376

M

3

H2O/h

0.30

N

Make-up flow rate, m

3

0.35

A

CO2Ei, tCO2/t

WGi, m3ww/t

SC R

Luyben15

IP T

11.17

Maity et al.16 11.89

Process

1.0913

0.8856

ED

6.1. Raw Material and Fuel Consumption The total amount of raw material required to feed each process was simply determined by

PT

summing up fresh benzene and fresh propylene’s flow rates. The eco-indicator was then calculated by dividing the result by their respective total flow rate of cumene. With regard to fuel consumption, in turn, the simulation of the conventional and transalkylation technologies

CC E

provided the fired heater’s energy demands, while the simulation of the utilities plant provided the boiler’s energy requirements. Hence, the fuel consumption eco-indicator could be determined for each process through Eq. (4).

A

Table 6 shows the results for each process, from which it can be observed that all modifications in relation to the conventional technology provided a decrease in both the raw materials consumption and fuel consumption indicators. In fact, the higher fuel consumption of the conventional and transalkylation processes is due to the fact that they not only have a boiler to generate utility steam but also a fired heater to pre-heat the reactor’s inlet stream.

25

6.2. Energy Use and CO2 Emissions The total energy requirements of each process were divided by their respective cumene product’s flow rate in order to calculate their energy use indicators. Their equivalent CO2 emissions were then estimated by following the procedure described by Mangili et al.45. It can be seen that Flegiel et al.12’s process consumes much less energy and emits less carbon dioxide per ton of product than the other technologies. Such advantage is due to both the energy integration between C1’s reboiler and the recompressed C2’s top product and the use of the

IP T

electricity generated by the reactor product’s gas expander in C2’s top product compressor.

The low energy consumption of Pathak et al.11’s technology, in turn, can be associated with the RD column, which integrates the process by allowing the utilization of the heat of reaction to

SC R

assist in the colum’s separation mechanism and, thus, increases overall conversion. As for Zhai

and coworkers17’s process, the low energy requirements are due to liquid/vapor interconnection As previously highlighted, the electricity-CO2 emissions factor of 0.0227 tCO2/GJ was used

U

since the process plants were assumed to be installed in Brazil, whose energy mix mainly

N

consists of hydroelectric plants. In order to show how such factor affects the CO2 emissions depending on the energy generation source, different locations with different energy mixes were

A

evaluated. The results are shown in Figure 9, in which the bars refer to the left y-axis and the data-labeled curves relate to the right y-axis. We note that Hong Kong has the highest emission

M

values, followed by South Korea and United States. Hong Kong’s energy mix is mainly based on coal power stations (70%), which results in an electricity-CO2 emissions factor as high as

ED

0.2138 tCO2/GJ. South Korea’s energy mix comprises 40% coal plants and 30% nuclear stations, thus having a factor of 0.1533 tCO2/GJ, whereas United States’ consists mainly of

A

CC E

PT

natural gas thermal plants (30%), resulting in a factor of 0.1385 tCO2/GJ72.

Figure 9. CO2 indicator for different locations.

26 Canada, Brazil and France’s energy mixes, in turn, are mostly comprised of either hydroelectric or nuclear power plants, which justifies the lower CO2 emissions per GJ of energy generated. In fact, Canada and France’s factors are as low as 0.0456 tCO2/GJ and 0.0163 tCO2/GJ, respectively72. Although the locations evaluated have significantly different electricity-CO2 conversion factors, the carbon dioxide indicators do not vary greatly. This is due to the low electricity consumption of the cumene processes studied. As we can see, the energy consumed by pumps and compressors only accounts for a maximum of 1% of the processes’ total energy

6.3. Water Consumption and Wastewater Generation

IP T

requirements.

The water consumption eco-indicators were calculated with basis on the procedure described by

SC R

Mangili et al.45, that is, only water and steam losses were taken into account. Besides Flegiel et al.12’s HE1 and C2’s condenser, which use chilled water, all condensers and coolers operate

with cooling water, whereas tubular reactors are BFW-cooled. Regarding heaters and reboilers,

U

LPS, MPS and HPS were assumed to be supplied deending on the equipment’s temperature requirements, in accordance with the data provided by Mangili et al.45. The required make-up

N

water for the utility plant was given by simulation and used to determine the indicators. From

A

the results given in Table 6 we observe that Flegiel et al.12’s process proved to be the most water consumption-efficient technology, which is, again, due to C1’s reboiler energy

M

integration.

Finally, the wastewater generation indicator from each process was determined by dividing the

ED

the total wastewater flow rate produced – sum of all process losses and equipment blowdown – by their respective production rate. Again, Flegiel and coworkers12’s two-flash energy-

PT

integrated technology showed to be the best alternative since it produces less wastewater per ton of product than the other processes. It is important to highlight that the wastewater generation and water consumption indicators are interconnected, since the lower the water and steam

CC E

requirements, the lower the process losses.

6.4. Economic Analysis

A

From the assumptions described in Section 4 it was possible to determine the capital and operating expenditures for all six cumene manufacturing processes. In addition, by calculating the processes’ TACs and credits we estimated the gross annual profit (GAP) associated with each technology. The results are shown in Table 7, from which it can be verified that the intensified processes are not only more environmentaly friendly but also economically attractive. The economic results provided in this paper are different from those calculated by the reference authors due to the differences regarding the methodology applied. None of the authors took into

27 accounts the utility plant costs to estimate CAPEX and other general expenses such as the number of shifts and operators, overhead costs, maintenance costs and taxes and insurance, among others, to estimate OPEX. In addition, the values found in this study were determined with basis on Aspen PEA’s estimation method and updated currencies/prices. Table 7. Gross profit results. Flegiel et al.12 17.87

Zhai et Pathak Norouzi al.17 et al.11 et al.19 8.22 5.90 8.09

OPEX (10 $/year)

132.71

140.31

469.41

143.24

140.11

138.97

Payback period (years)

3

3

3

3

3

3

TAC (106 $/year)

135.15

143.02

475.37

132.11

131.54

141.67

1.71

________

0.77

0.01

________

0.02

6

6

HPS credit (10 $/year)

0.92

1.02

2.69

SC R

CAPEX (106 $)

IP T

7.33

Maity et al.16 8.13

Luyben15

Process

6

0.03

0.03

0.06

6

Cumene revenue (10 $/year)

147.70

157.24

507.32

160.47

158.19

158.98

GAP (106 $/year)

13.50

15.27

34.70

15.74

16.11

18.11

U

Fuel/DIPB credit (10 $/year)

As previously discussed, although the intensified processes present higher annual profits, such

N

parameter is not appropriate for comparing the processes’ economics. In fact, according to the

A

results for the specific production cost indicators shown in Table 8, we note that Zhai et al.17’s technology has the lowest costs per production rate, followed by Pathak et al.11’s process. We

M

verify, however, that although Flegiel and coworkers12’s design has the highest GAP, its production cost indicator is the lowest. This indicates a lower economic performance when

ED

compared to the other alternatives.

Process

PT

Table 8. SPCi results.

TAC, 106 $/year

3

CC E

Production rate, 10 t/year SPCi, 103 $/t

135.15

Maity et al.16 143.02

Flegiel et al.12 475.37

Zhai et al.17 132.11

Pathak Norouzi et al.11 et al.19 131.54 141.67

89.36

95.12

306.88

96.88

95.68

97.92

1.5124

1.5036

1.5490

1.3636

1.3748

1.4468

Luyben15

6.1 Eco-efficiency Comparison

A

Following Mangili and coworkers45 ECI methodology, the cumene production processes were compared in terms of their respective environmental performances. Since each eco-indicator falls into different categories, they were normalized by dividing them by the highest value present in the same group, as presented in Table 9. We can verify that the results vary from 0 to 1, that is, from the most eco-efficient to the least eco-efficient in that category. The normalized eco-indicators were then plotted in a radar-type chart to provide a qualitative analysis of ther eco-efficiency, as illustrated in Figure 10. The ECI diagram resulted in

28 heptagon-formed shapes since 7 indicators were evaluated. It should be pointed out, however, that the same relevance (weight) was assumed for all categories. Table 9. Normalization of eco-indicators. Zhai et al.17 1.0669 0.0289 1.3922 0.0776 0.9546 1.0380 1.3636

Pathak et al.11 1.0242 0.0251 1.1965 0.0669 0.6973 0.6990 1.3748

Norouzi et al.19 0.9812 0.0327 1.5891 0.0891 1.0082 1.0376 1.4468

0.9444 0.8333 0.8629 0.8824 0.8148 0.8165 0.9677

0.9722 0.1667 0.2341 0.2353 0.3426 0.3670 1.0000

0.9907 0.5000 0.4682 0.4706 0.8889 0.9541 0.8774

0.9537 0.3333 0.4013 0.4118 0.6481 0.6422 0.8839

0.9074 0.5000 0.5318 0.5294 0.9352 0.9541 0.9355

IP T

Flegiel et al.12 1.0485 0.0128 0.7031 0.0365 0.3699 0.4022 1.5490*

N

RMCi 1.0761* FCi 0.0618* EUi 2.9902* CO2Ei 0.1674* WCi 1.0788* WGi 1.0913* SPCi 1.5124 Normalized eco-indicators RMCi 1.0000 FCi 1.0000 EUi 1.0000 CO2Ei 1.0000 WCi 1.0000 WGi 1.0000 SPCi 0.9742 * Highest values by category.

Maity et al.16 1.0227 0.0530 2.5753 0.1447 0.8747 0.8856 1.5036

SC R

Luyben15

U

Eco-indicator

The normalized eco-indicators were then plotted in a radar-type chart to provide a qualitative

A

analysis of ther eco-efficiency, as illustrated in Figure 10. The ECI diagram resulted in

M

heptagon-formed shapes since 7 indicators were evaluated. It should be pointed out, however,

A

CC E

PT

ED

that the same relevance (weight) was assumed for all categories.

Figure 10. Qualitative ECI for all six cumene production technologies.

29 Although we can visually infer that some processes are less eco-efficient than others due to the differences of their polygon areas, some technologies present similar shapes and sizes. Therefore, it is necessary to develop a quantitative evaluation with basis on the indicators. The areas of each minor triangle inside the heptagon-shaped polygons were calculated through Eq. (6) and subsequently summed up in order to determined their respective heptagon areas45, through Eq. (7). The quantitative ECI for each production plant was then determined, as presented in Table 10.

IP T

We can note that the conventional process is the least eco-efficient due to its bigger ECI

heptagon area, followed by the transalkylation technology, which is 22.94% more eco-efficient than the former. The intensified processes, in turn, proved to be much more environmentaly

SC R

efficient, having the two-flash energy integrated technology the highest performance (79.20% more than the conventional process). This substantiates the relevance of energy integration in

improving process economics and reducing ecological impacts, even more than two of the most

U

promising process intensification technologies (DWC and RD).

Table 10. Quantitative ECI for all six cumene production technologies.

1.00 1.00 1.00 1.00 1.00 0.97 1.00 1.00 1.00 1.00 0.97 1.00 1.00 1.00 0.97 1.00 1.00 0.97 1.00 0.97 0.97 20.85 8.15

A

CC E

PT

ED ______

Flegiel et al.12 0.16 0.23 0.23 0.33 0.36 0.97 0.04 0.04 0.06 0.06 0.17 0.06 0.08 0.09 0.23 0.08 0.09 0.24 0.13 0.34 0.37 4.34 1.70 79.20

N

RMCi x FCi RMCi x EUi RMCi x CO2Ei RMCi x WCi RMCi x WGi RMCi x SPCi FCi x EUi FCi x CO2Ei FCi x WCi FCi x WGi FCi x SPCi EUi x CO2Ei EUi x WCi EUi x WGi EUi x SPCi CO2Ei x WCi CO2Ei x WGi CO2Ei x SPCi WCi x WGi WCi x SPCi WGi x SPCi Sum Heptagon area (ST)45,47 ECI (Eq. (7))45,47

Maity et al.16 0.79 0.81 0.83 0.77 0.77 0.91 0.72 0.74 0.68 0.68 0.81 0.76 0.70 0.70 0.84 0.72 0.72 0.85 0.67 0.79 0.79 16.05 6.28 22.94

A

Luyben15

M

Indicator x indicator

Zhai et al.17 0.50 0.46 0.47 0.88 0.95 0.87 0.23 0.24 0.44 0.48 0.44 0.22 0.42 0.45 0.41 0.42 0.45 0.41 0.85 0.78 0.84 11.19 4.38 46.32

Pathak et al.11 0.32 0.38 0.39 0.62 0.61 0.84 0.13 0.14 0.22 0.21 0.29 0.17 0.26 0.26 0.35 0.27 0.26 0.36 0.42 0.57 0.57 7.65 2.99 63.29

Norouzi et al.19 0.45 0.48 0.48 0.85 0.87 0.85 0.27 0.26 0.47 0.48 0.47 0.28 0.50 0.51 0.50 0.50 0.51 0.49 0.89 0.87 0.89 11.85 4.63 43.14

30 As previously discussed, the ECI method consists of a joint evaluation of indicators, thus representing an appropriate strategy to compare the environmental performance of processes, since the determination of single indicators is not sufficient to assert which technology is the most eco-efficent. However, when the assessment lacks sufficient data to establish a composite analysis, the assessor may perform a rough comparison in terms of single, yet adequate indicators. Although practical, the raw materials consumption indicator may be deemed unessential either

IP T

when the reaction yields of different processes are similar or when no reactions take place,

which is the case of Mangili et al.45’s work. The fuel consumption, in turn, cannot be estimated for applications where no fuel consumers are used, or when the utilities are provided by an

SC R

external source instead of a utility plant, which is assumed by most authors (including the

references studies here). When the utilities plant is considered, particular care must be taken when investigating the other indicators, especially in relation to the energy consumption, which is associated with all other metrics. The energy results obtained from simulation are useful in

U

determining the amount of fuel consumed in the boiler to generate steam for providing the

N

required heat to reboilers and heaters. Also, such results are used to estimate the CO2 emissions due to electricity and fuel consumption, as well as the amount of water necessary in coolers and

A

condensers. In this regard, the water consumption indicator is also of particular importance

M

since it is directly related to the wastewater generation, which is calculated from the utility plant heuristics that are mainly based on the recirculating water flow rate. The energy and water

ED

requirements are then utilized in the estimation of the production costs. Hence, determining the energy use and production cost indicators may be deemed sufficient for carring out a rough comparison, especially when it comes to process intensification techniques,

PT

since they are aimed at fostering the development of more energy-efficient technologies In fact, most authors compare process plants in terms of energy consumption and economics, which

CC E

then provides a partial, preliminary assessment to assist decision-making procedures. Nevertheless, attention must be drawn to the fact that the energy consumption is sufficient only when no off-gas stream is combusted (e.g. in a flare, heat recovering unit, etc.), or when vent streams directly release CO2 to the environment. In Mangili et al.45’s work, for instance, no such

A

streams are present, whereas in this paper they are assumed to be exported as fuel credits. On one hand, for applications where a flare or a heat recovering system is considered, the carbon dioxide produced from combustion must be taken into consideration, which is the case of Guillén-Gozálbez et al.40’s toluene hydrodealkilation processes. The authors stated that the methane purge streams are combusted in oder to recover the heat. On the other hand, if CO2 is directly vented to the atmosphere its flow rate must be accounted, as in the case of Jia et al.36’s

31 ethanol production processes via ethylene and via cellulose. These aspects would affect greatly the CO2 indicator, which would then be required in the analysis.

7. Conclusion This study compared six cumene production processes, including their utilities plants, with regard to their respective economics and ecological performance. The environmental analysis was carried out with basis on the calculation of six different equally-weighted categories of eco-

IP T

indicators, while the economic evaluation was performed with the aid of Aspen PEA and equipment specifications provided by the reference authors in order to determine the processes’specific production cost indicator. The environmental performances of all

SC R

technologies were compared through the Eco-efficiency Comparison Index. The comparison

showed that the intensified processes are both more economically and environmentaly attractive in relation to the conventional and transalkylation processes. The two-flash energy-integrated technology proved to be the best alternative, even when compared to the DWC and RD designs,

U

since it is the most eco-efficient (79% more than the conventional process). Such conclusion is

N

mainly based on the energy integration between column C1’s reboiler and the recompressed

A

column C2’s top product.

It was, therefore, possible to demonstrate the significance of process intensification in terms of

M

not only reducing process’ costs but also minimizing environmental impacts from the industry. In light of these results, it is possible to develop future studies based on eco-indicators to

ED

analyze the feasibility of a thermal integration in the conventional and transalkylation technologies. A future evaluation regarding the use of a dividing wall reactive column could

PT

also be carried out with basis on both Pathak et al.11 and Zhai et al.17’s configurations.

Symbols Used

A

CC E

Symbols C1

Column 1

C2

Column 2

C3

Column 3

Cbenzene

Benzene composition

Ccumene

Cumene composition

Cpropylene

Propylene composition

D

Vessel diameter

E

Required energy

E1

Expander

32

Heat exchanger

H1

Fired heater

HE1

Cooler 1

HE2

Cooler 2

HX1

Heater 1

HX2

Heater 2

K1

Compressor

L

Vessel length

ṁng

Natural gas mass flow rate

n

Thermal efficiency

ng

Natural gas

NCVng

Natural gas net calorific value

P1

Pump 1

P2

Pump 2

P3

Pump 3

R1

Tubular reactor

R2

Transalkylator

V1

Vessel 1

V2

Vessel 2

xBz

Benzene molar fraction

xC3

Propane molar fraction

xC3=

Propylene molar fraction

xCu

Cumene molar fraction p-Diisopropyl benzene molar fraction

CC E

Subscripts i

Equipment (fuel consumer)

Abbreviations

A

SC R U N

A M

ED

PT

xDIPB

IP T

FEHE

ANOVA

Analysis of Variance

BFP

Boiler Feed Pump

BFW

Boiler Feed Water

CAPEX

Capital Expenditure

CHW

Chilled Water

CW

Cooling Water

33

p-Diisopropyl benzene

DWC

Dividing Wall Column

ECI

Eco-efficiency Comparison Index

ESCAP

Economic and Social Commission for Asia and the Pacific

GHG

Greenshouse Gas

GAP

Gross Annual Profit

GWP

Global Warming Potential

HPS

High Pressure Steam

LCA

Lice Cycle Assessment

LPS

Low Pressure Steam

ML

Maintenance Labor

MOO

Multi Objective Optimization

MPS

Medium Pressure Steam

NPV

Net Present Value

OPEX

Operating Expenditure

PEA

Process Economic Analyzer

PEPA

Process Environmental Performance Assessment

RAPID

Rapid Advancement in Process Intensification Deployment

RD

Reactive Distillation

TAC

Total Annualized Cost

WMO

World Meteorological Organization

SC R U N

A

M

ED

PT

References

WMO – World Meteorological Organization. WMO Statement on the State of the Global Climate in 2016. WMO (2017). (Available at https://library.wmo.int/opac/doc_num.php?explnum_id=3414).

A

CC E

(1)

IP T

DIPB

(2)

European Commision. CO2 time series 1990-2015 per region/country. European Commission, Joint Research Centre (JRC)/PBL Netherlands Environmental Assessment Agency. Emission Database for Global Atmospheric Research (EDGAR) (2017). (Availabe at http://edgar.jrc.ec.europa.eu/overview.php?v=CO2ts1990-2015&sort=des9).

(3)

A.I. Stankiewicz and J.A. Moulijn. Process intensification: transforming chemical engineering. Chem. Eng. Prog. 96, 22–34 (2000).

(4)

H. Wang, A. Mustaffar, A. N. Phan, V. Zivkovik, D. Reay, R. Law and K. Boodhoo. A review of process intensification applied to solids handling. Chem. Eng. Process. 118, 78–107 (2017).

34 European Commision. Rejuvenating the European chemical sector: Sustainable chemistry as a catalyst for change. European Union (2009). (Availabe at https://ec.europa.eu/research/industrial_technologies/pdf/chemical-sectorbrochure_en.pdf).

(6)

U.S. Department of Energy. Energy Department Announces American Institute of Chemical Engineers to Lead New Manufacturing USA Institute. U.S. Department of Energy (2016). (Available at https://energy.gov/articles/energy-department-announcesamerican-institute-chemical-engineers-lead-new-manufacturing).

(7)

IHS Markit. Cumene: Chemical Economics Handbook. HIS Markit (2016). (Available at https://ihsmarkit.com/products/cumene-chemical-economics-handbook.html).

(8)

Grand View Research, Inc. Cumene Market Analysis By Production, By Region and Segment Forecasts, 2014-2025. Grand View Research, Inc. (2017) (Available at https://www.grandviewresearch.com/industry-analysis/cumene-market).

(9)

S. H. Mcallister, J. Anderson and E. F. Bullard. Production of cumene by vapor-liquidphase alkylation process. Chem. Eng. Prog. 43, 189–196 (1942).

SC R

IP T

(5)

U

(10) A. Chudinova, A. Salischeva, E. Ivashkina, O. Moizes and A. Gavrikov. Application of cumene technology mathematical model. Procedia Chem. 15, 326–334 (2015).

A

N

(11) A. S. Pathak, S. Agarwal, V. Gera, N. Kaistha. Design and control of a vapor-phase conventional process and reactive distillation process for cumene production. Ind. Eng. Chem. Res. 50, 3312–3326 (2011).

M

(12) F. Flegiel, S. Sharma and G. P. Rangaiah. Development and multiobjective optimization of improved cumene production processes. Mater. Manuf. Processes 30, 444–457 (2015).

ED

(13) R. Turton, R. C. Bailie, W. B. Whiting, J. A. Shaelwitz. Analysis, synthesis and design of chemical processes, 2nd ed. Prentice Hall: Upper Saddle River, NJ (2010).

PT

(14) M. Han, X. Li and S. Lin. Intrinsic kinetics of the alkylation of benzene with propylene over ß zeolite catalyst. Kinet. Catal. 42, 533–538 (2001). (15) W. L. Luyben. Design and Control of the Cumene Process. Ind. Eng. Chem. Res. 49, 719–734 (2010).

CC E

(16) D. Maity, R. Jagtap and N. Kaistha. Systematic top-down economic plantwide control of the cumene process. J. Process Control 23, 1426–1440 (2013).

A

(17) J. Zhai, Y. Liu, L. Li, Y. Zhu, W. Zhong and L. Sun. Applications of dividing wall column technology to industrial-scale cumene production. Chem. Eng. Res. Des. 102, 138–149 (2015). (18) Z. Lei, C. Li, J. Li and B. Chen. Suspension catalytic distillation of simultaneous alkylation and transalkylation for producing cumene. Sep. Purif. Technol. 34, 265–271 (2004). (19) H. R. Norouzi, M. A. Hasani, B. Haddade-Sisakht and N. Mostoufi. Economic design and optimization of zeolite-based cumene production plant. Chem. Eng. Commun. 201 (10), 1270–1293 (2014).

35 (20) IPCC – International Panel on Climate Change. Climate Change 2014: Mitigation of Climate Change. Contribution of Working Group III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Edenhofer, O., R. Pichs-Madruga, Y. Sokona, E. Farahani, S. Kadner, K. Seyboth, A. Adler, I. Baum, S. Brunner, P. Eickemeier, B. Kriemann, J. Savolainen, S. Schlömer, C. von Stechow, T. Zwickel and J.C. Minx (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA (2014).

IP T

(21) S. Sharma, C. L. Zi and G. P. Rangaiah. Process design for economic, environmental and safety objectives with an application to the cumene process. In Multi-Objective Optimization in Chemical Engineering: Developments and Applications. Wiley: New York, NY (2013). (22) J. A. Weinfeld, S. A.Owens and R. B. Eldridge. Reactive dividing wall columns: A comprehensive review. Chem. Eng. Process. 123, 20–33 (2018).

SC R

(23) I. Dejanović, Lj. Matijašević and Ž.Olujić. Dividing wall column–a breakthrough towards sustainable distilling. Chem. Eng. Process. 6, 559–580 (2010).

U

(24) N. Asprion and G. Kaibel. Dividing wall columns: Fundamentals and recent advances. Chem. Eng. Process. 48, 139–146 (2010).

N

(25) S. Tututi-Avila, L. A. Domínguez-Díaz, N. Medina-Herrera, A. Jiménez-Gutiérrez and J. Hahn. Dividing-wall columns: Design and control of a kaibel and a satellite distillation column for BTX separation. Chem. Eng. Process. 114, 1–15 (2017).

M

A

(26) M. Rodríguez, P. Z. Li and I. Díaz. A control strategy for extractive and reactive dividing wall columns. Chem. Eng. Process. 113, 14–19 (2017).

ED

(27) S. Y., Hwang and S. S. Chen. Cumene review. Kirk-Othmer Encyclopedia of Chemical Technology 8, 147–157 (2010). (28) R. Agrawal. Multieffect distillation for thermally coupled configurations. AIChE J. 46 (11), 2211–2224 (2000).

CC E

PT

(29) X. Gao, J. Chen, J. Tan, Y. Wang, Z. Ma and L. Yang. Application of mechanical vapor recompression heat pump to double-effect distillation for separating n,ndimethylacetamide/water mixture. Ind. Eng. Chem. Res. 54 (12), 3200–3204 (2015).

A

(30) R. Palacios-Bereche, A. V. Ensinas, M. Modesto and S. A. Nebra. Double-effect distillation and thermal integration applied to the ethanol production process. Energy 82, 512–523 (2015). (31) L. C. B. A. Bessa, M. C. Ferreira, E. A. C. Batista and A. J. A. Meirelles. Performance and cost evaluation of a new double-effect integration of multicomponent bioethanol distillation. Energy 63, 1–9 (2013). (32) X. You, I. Rodriguez-Donis and V. Gerbaud. Reducing process cost and CO2 emissions for extractive distillation by double-effect heat integration and mechanical heat pump. Appl. Energ. 166, 128–140 (2016).

36 (33) G. G. Vinning and S. M. Kowalski. Statistical methods for engineers, 3th ed. Brooks/Cole, Cengage Learning: Boston, MA (1998). (34) IPCC – International Panel on Climate Change. Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Core Writing Team, R.K. Pachauri and L.A. Meyer (eds.)]. IPCC, Geneva, Switzerland, 151 pp. (2014). (35) Water.org. The water crisis. Water.org (2017). (Availabe at https://water.org/ourimpact/water-crisis/).

IP T

(36) X. Jia, F. Han and X. Tan. Integrated environmental performance assessment of chemical processes. Comput. Chem. Eng. 29 (1), 243–247 (2004).

SC R

(37) C. Koroneos, A. Dompros and G. Roumbas. Hydrogen production via biomass gasification—A life cycle assessment approach. Chem. Eng. Process. 47 (8), 1261–1268 (2008).

(38) M. Brito and F. Martins. Life cycle assessment of butanol production. Fuel 208, 476–482 (2017).

N

U

(39) P. Vaskan, G. Guillén-Gonzálbez, M. Turkay and L. Jiménez. Multiobjective optimization of utility plants under several environmental indicators using an MILPbased dimensionality reduction approach. Ind. Eng. Chem. Res. 53 (50), 19559–19572 (2014).

M

A

(40) G. Guillén-Gozálbez, J. A. Caballero, L. J. Esteller and M. Gadalla. Application of life cycle assessment to the structural optimization of process flowsheets. Comput. Aided Chem. Eng. 24, 1163–1168 (2007).

ED

(41) L. Jacquemin, P. Pontalier and C. Sablayrolles. Life cycle assessment (LCA) applied to the process industry: a review. Int. J. Life Cycle Assess. 17 (8), 1028–1041 (2012)

PT

(42) G. J. Ruiz-Mercado, R. L. Smith and M. A. Gonzalez. Sustainability indicators for chemical processes: I. Taxonomy. Ind. Eng. Chem. Res. 51 (5), 2309–2328 (2012).

CC E

(43) The Brittish Standards Institution. Environmental management — Eco-efficiency assessment of product systems — Principles, requirements and guidelines (ISO 14045:2012). BSI Standards Limited (2012).

A

(44) ESCAP – Economic and Social Commission for Asia and Pacific. Eco-efficiency Indicators: Measuring Resource-use efficiency and the Impact of Economic Activities on the Environment. United Nations: Bangkok (2009). (Available at https://sustainabledevelopment.un.org/content/documents/785eco.pdf). (45) P. V. Mangili, Y. P. D. M. Souza, D. Q. F. Menezes, L. S. Santos and D. M. Prata. Ecoefficiency evaluation of acetone-methanol separation processes using computational simulation. Chem. Eng. Process. 123, 100–110 (2018). (46) IEA – International Energy Agency. Tracking Industrial Energy Efficiency and CO2 Emissions. International Energy Agency: Paris (2007). (Available at https://www.iea.org/publications/freepublications/publication/tracking_emissions.pdf).

37 (47) C. P. Pereira, D. M. Prata, L. S. Santos, L. P. C. Monteiro. Development of eco-efficiency comparison index through eco-indicators for industrial applications. Braz. J. Chem. Eng, 35 (1), 63–84 (2018). (48) T. Silalertruksa, S. H. Gheewala and P. Pongpat. Sustainability assessment of sugarcane biorefinery and molasses ethanol production in Thailand using eco-efficiency indicator. Appl. Energy 160, 603–609 (2015).

IP T

(49) Q. Chen, M. Lv, D. Wang, Z. Tang, W. Wei and Y. Sun. Eco-efficiency assessment for global warming potential of ethylene production processes: a case study of China. J. Clean. Prod. 142 (4), 3109–3116 (2017). (50) W. Wang, D. Jiang, D. Chen, Z. Chen, W. Zhou and B. Zhu. A Material Flow Analysis (MFA)-based potential analysis of eco-efficiency indicators of China's cement and cement-based materials industry. J. Clean. Prod. 112 (1), 787–796 (2016).

SC R

(51) J. C. Sacramento-Rivero. A methodology for evaluating the sustainability of biorefineries: framework and indicators. Biofuels, Bioprod. Bioref. 6, 32–44 (2012).

U

(52) G. Al-Sharrah, A. Elkamel and A. Almanssoor. Sustainability indicators for decisionmaking and optimisation in the process industry: The case of the petrochemical industry. Chem. Eng. Sci. 65, 1452–1461 (2010).

A

N

(53) A. Zagmani, P. Buttol, P. L. Porta, R. Buonamici, P. Masoni, J.Guinée, R. Heijungs, T. Ekvall, R. Bersani, A. Biénkowska and U. Pretato. Critical review of the current research needs and limitations related to ISO-LCA practice. Co-ordination Action for innovation in Life-Cycle Analysis for Sustainability (2008).

ED

M

(54) J. C. Bare, P. Hofstetter, D. W. Pennington and H. U. de Haes. Life cycle impact assessment workshop summary – Midpoints versus endpoints: The sacrifices and benefits. Int. J. Life Cycle. Assess. 5 (6), 319–326 (2000).

PT

(55) M. Goedkoop and R. Spriensma. The Eco-indicator 99: A damage oriented method for life cycle impact assessment – methodology report, 2th ed. PRé Consultants B. V. (2000). (Available at https://www.pre-sustainability.com/download/EI99_annexe_v3.pdf). (56) H. S. Park and S. K. Behera. Methodological aspects of applying eco-efficiency indicators to industrial symbiosis networks. J. Clean. Prod. 64, 478–485 (2014).

CC E

(57) IEA – International Energy Agency. Energy Statistics Manual. OECD/IEA (2005). (Available at https://www.iea.org/publications/freepublications/publication/statistics_manual.pdf).

A

(58) EPA – Environmental Protection Agency. Project Type: Industrial Boiler Efficiency. NSCEP, EPA (2008). (Available at https://www.epa.gov/nscep). (59) W. Trinks, M. H. Mawhinney, R. A. Shannon, R. J. Reed and J. R. Garvey. Industrial Furnaces, 6th ed. John Wiley & Sons Inc.: Hoboken, NJ (2004). (60) R. H. Perry, D. W. Green and J. O Maloney. Perry’s chemical engineers’ handbook, 7th ed. McGraw-Hill: New York, NY (1997). (61) IPCC – Intergovernmental Panel on Climate Change. Guidelines for National Greenhouse Gas Inventories. Institute for Global Environmental Strategies (IGES): Hayama (2006). (Available at http://www.ipcc-nggip.iges.or.jp/public/2006gl/vol1.html).

38

(62) MCTIC – Ministério da Ciência, Tecnologia, Inovação e Comunicações. Fator de Emissão de CO2 pela geração de energia elétrica no Sistema Interligado Nacional (SIN). MCTIC: Brazil (2017). (Available at http://www.mct.gov.br/index.php/content/view/321144.html#ancora). (63) T. J. Tvedt and R. T. Holloway. Consensus on operating practices for the control of feedwater and boilerwater chemistry in modern industrial boilers (ASME I00367). NACE International (1997).

IP T

(64) W. D. Seider, J. D. Seader, D. R. Lewin, S. Widagdo. Product and Process Design Principles, 3rd ed. John Wiley & Sons, Inc.: New Jersey (2008).

SC R

(65) GE – General Electric. Handbook of Industrial Water Treatment. GE Water and Power: Water & Process Technologies. GE (2012). (Available at https://www.suezwatertechnologies.com/handbook/index.jsp). (66) M. Novak and P. Horvat. Mathematical modelling and optimisation of a waste water treatment plant by combined oxygen electrode and biological waste water treatment model. Appl. Math. Modelling 36 (8), 3813–3825 (2012).

U

(67) R. Smith. Chemical process design and integration. John Wiley & Sons, Inc.: West Sessex, England (2005).

A

N

(68) H. Silla. Chemical Process Engineering: Design and Economics. Taylor & Francis Group LLC: Hoboken, NJ (2003).

M

(69) R. K. Sinnott. Chemical Engineering Design, 3rd ed. Elsevier Butterworth-Heinemann: Jordan Hill, MA (2005).

ED

(70) S. A. Leeper. Wet Cooling Towers: Rule-of-thumb Design and Simulation. Idaho National Engineering and Environmental Laboratory, EG & G Idaho: Idaho Falls, ID (1981).

PT

(71) EPA – U. S. Environmental Protection Agency. Capital and operating costs for industrial boilers. Prepared by PEDCo Environmental, Inc.: Cincinnati, OH (1979).

A

CC E

(72) EASME – Executive Agency for Small and Medium-sized Enterprises. Electricity Map: Live CO2 emissions of the European electricity consumption. BUILD UP, European Commission (2017). (Available at https://www.electricitymap.org/?page=map&solar=false&remote=true&wind=false).