Optimal Integrated Plant for Production of Biodiesel from Waste

Research Article pubs.acs.org/journal/ascecg

Optimal Integrated Plant for Production of Biodiesel from Waste Borja Hernández and Mariano Martín* Departamento de Ingeniería Química, Universidad de Salamanca, Plaza de los Caídos 1-5, 37008 Salamanca, Spain ABSTRACT: In this work, we have optimized an integrated facility for the production of fully renewable biodiesel from waste via anaerobic digestion. On the one hand, the biogas produced is dry reformed to obtain syngas and methanol. On the other hand, the digestate remaining is used as a nutrient to grow algae. Algae oil is transesterified using internally produced methanol. The problem is formulated as a large non-linear programming problem (NLP) to compute the optimal operating conditions of the various reactors as well as the biogas composition and the content of nutrients in the digestate. Optimal heat exchanger and water networks are designed afterward. A total of 3110 kt/yr of manure is processed for the production of 90 Mgal/yr of biodiesel and 46.7 kt/yr of methanol. The process uses a fraction of the biogas to provide the energy required at the reformer, but the process still needs 1.5 MJ/gal of fatty acid methyl ester and 1.22 gal of water per gal of biodiesel. The production cost of biodiesel is competitive with previous work, 0.31 €/gal, but the investment is higher due to the need to process the waste into biogas, 193 M€. A scaled-down study is also carried out to evaluate the production and investment cost as a function of the farm size. KEYWORDS: Biogas, Sludge, Manure, Circular economy, Mathematical optimization

■

INTRODUCTION Current developed societies generate large volumes of waste. The major types are the sludge produced in water treatment, urban residues, forest residues, and manure. Apart from the volume, their composition is complex and dangerous. Anaerobic digestion represents a technology to stabilize the waste while generating added value in the form of methane and digestate that can be used as fertilizer. For decades, the aim of processing waste focused on power production. León and Martı ́n1 optimized a facility for the use of manure to produce power using a gas and a steam turbine. The facility was only profitable if the digestate was used a fertilizer and sold as such. However, biogas is an interesting carbon source, not only because of the methane, but also due to the amount of CO2 available. Hernández and Martı ́n2 evaluated the possibility of using the biogas as a source for the production of methanol, taking into account that its composition already contains all the species required for dry reforming of methane. The process consisted of producing syngas out of biogas. Next, methanol was synthesized. Lately, Hernández et al3 evaluated the use of different wastes toward the production of power and chemicals. They found that the price of digestate as an asset determined the selection of the waste or the mix of wastes. Algae production requires a source of carbon and nutrients. We realize that both can be obtained from the processes presented above. Digestate contains nutrients that can allow algae growth. Xin et al.4 evaluated the effect of the nutrient concentration on the growth rate of algae. Furthermore, in the production of methanol, CO 2 is captured while the composition of the syngas is adjusted before feeding the © 2017 American Chemical Society

reactor. Even if there is not enough CO2, we have all the ingredients for the production of biodiesel with no need for fossil-based chemicals and with the opportunity of additional CO2 capture. This facility can be compared to previous work where various integrated designs were presented for the production of biodiesel with no fossil-based raw material, such as the work of Martı ́n and Grossmann,5 where algae was used to produce both the alcohol, ethanol, and the oil needed for biodiesel production. Recently, Martı ́n and Grossmann6 suggested the simultaneous production of the alcohol and the oil by using switchgrass and algae. Furthermore, other alternatives proposed in the literature involve the use of solar energy and CO2 as the raw materials for the production of both the algae and methanol using electrolytic hydrogen and CO2.7 In both cases, nutrients for the algae were bought, but no fossil raw material was required since the oil and the methanol came from renewable sources. Process integration has provided the means to mitigate the lack of resources and improve the economics or the environmental impact by putting together chemical complexes from renewables such has solar, wind, and biomass8,9 by integrating nuclear energy and water management10 or upgrading sections of refineries with utilities plants.11 Thus, in this work we propose and optimize an integrated facility that produces biodiesel from waste. Waste is digested to produce biogas and digestate. While digestate is used for algae growing, providing the required nutrients, biogas is used to Received: April 3, 2017 Revised: July 3, 2017 Published: July 6, 2017 6756

DOI: 10.1021/acssuschemeng.7b01007 ACS Sustainable Chem. Eng. 2017, 5, 6756−6767

Research Article

ACS Sustainable Chemistry & Engineering

Figure 1. Integrated process for biodiesel production from biomass wastes.

Figure 2. Detail of the biogas production section of the flowsheet.

recovered by distillation, and the oil is sent to transesterification. We use the methanol produced from the biogas to produce biodiesel, fatty acid methyl ester (FAME). In this way we have an integrated facility from biomass waste to biodiesel production with no need for purchasing methanol to be compared with other integrated facilities where methanol was also produced from renewable-based resources.6,7 Figure 1 shows a scheme of the process. In the Modeling Issues section, we present the modeling issues related to the process.

produce the methanol required for the transesterification of the algae. This facility does not need to use nonrenewable intermediates such as methanol or nutrients via the proper integration of the subprocesses involved, avoiding ethical issues such as the use of fossil alcohol to produce a biofuel or environmental problems typically related to the discharge of the digestate in already nitrogen- or phosphorus-rich areas. We organize the paper as follows. First we provide a brief description of the process. Next the modeling issues related to each section of the process are presented. Subsequently, the results of the optimal operation of the facility are shown, together with an economic evaluation. Finally, we draw some conclusions.

■

MODELING ISSUES Biogas Production. Figure 2 shows the detail for the section of biogas production and gas cleanup. The anaerobic digestion of waste consists of a series of reactions, including hydrolysis, acidogenesis, acetogenesis, and methanogenesis, that break carbohydrates, lipids, and proteins of the feedstock into methane and CO2.12 Following the results of previous papers, we consider the operation of the digester under thermophilic conditions for 15−20 days. However, the composition of the biogas required for further stages may change the actual operating conditions. Apart from methane and CO2, nitrogen, H2S, and NH3 are produced.13 Thus, the biogas needs cleanup before further use. The yield from waste to biogas depends on the type of raw material. 14−16 Thus, in order to compute the biogas composition, a mass balance is performed considering the composition of the different manure sources such as pig and cattle slurry:1

■

PROCESS DESCRIPTION Biogas and digestate are produced via anaerobic digestion of wastes. The digestate is used as nutrients for algae growing. The biogas is further processed for methanol production. After its cleanup, it is reformed with the CO2 already available in the gas and steam, if necessary. Next, the syngas composition is adjusted for methanol production, where CO2 has to be partially captured before feeding the synthesis reactor. In order not to emit this greenhouse gas, we use it as the carbon source for algae production. Algae are grown using CO2 as a carbon source and the nutrients provided by the digestate. Water is added to the ponds. Most of it can be recycled or wastewater can also be used, but there is a fraction lost by evaporation from the ponds. The oil accumulated in the algae is extracted using a solvent, cyclohexane, and mechanical action. The solvent is 6757


Research Article

ACS Sustainable Chemistry & Engineering MWdry − biogas =

∑ Ya ′/biogas− dry MWa ′ a′

6 ≤ R C − N/CS ≤ 20 (1)

0.005 ≤ w′Nam /CS ≤ 0.047

The gas composition is bounded by typical results in the

0.005 ≤ w′Norg /CS ≤ 0.036

literature6 and it exits the reactor saturated with water. The

0.008 ≤ w′P/CS ≤ 0.013

composition is computed as per eqs 2−7

0.033 ≤ w′K/CS ≤ 0.1

0.7 ≤ YCH4 ≤ 0.5

3 ≤ R C − N/CS ≤ 10

0.3 ≤ YCO2 ≤ 0.5

0.005 ≤ w′Nam /PS ≤ 0.095

0.02 ≤ YN2 ≤ 0.06

0.005 ≤ w′Norg /PS ≤ 0.030

0.005 ≤ YO2 ≤ 0.16

0.019 ≤ w′P/PS ≤ 0.022

YH2S ≤ 0.002

0.0039 ≤ w′K/PS ≤ 0.083

9 × 10−5 ≤ YNH3 ≤ 1 × 10−4

ybiogas =

MWH2O MWbiogas − dry

Pv(T ) P − Pv(T )

(2)

(10)

wC/biomass + wNorg /biomass + wNam /biomass + wP/biomass + wK/biomass + wrest/biomass (11)

=1

(3)

fc(C)digestate = w′C/biomass wMS/biomassFbiomass

Fbiogas = ρbiogas [w′SV/cattle wMS/cattleFcattleVbiogas/cattle + w′SV/pork wMS/porkFporkVbiogas/pork ] fc(H 2O)biogas = ybiogas ∑ fc(a′)biogas a′

− fc(CH4) (4)

(12)

fc(Norg)digestate = w′Norg /biomass wMS/biomassFbiomass (5)

− fc(N2)

fc(a′,Bioreactor,Compres1)/MWa ′ Ya ′ /biogas − dry = [F(Bioreactor,Compres1) MWbiogas − dry − fc(H 2O,Bioreactor,Compres1)] MWbiogas ∑ a

xa /biogas MWa

=

∑ xa /biogas a

MWN MWN2

(13)

fc(N)digestate = w′N/biomass wMS/biomassFbiomass − fc(NH3) (6)

(7)

MWN MWNH3

(14)

fc(P)digestate = w′P/biomass wMS/biomassFbiomass

(15)

fc(K)digestate = w′K/biomass wMS/biomassFbiomass

(16)

fc(rest)digestate = w′rest/waste wMS/wasteFwaste − fc(CH4)biogas

Where the lower and upper limits for the potential generation

2MWO 4MWH − fc(CO2 )biogas MWCH4 MWCO2

of biogas are given by eq 8:8 0.20 ≤ Vbiogas/biomass ≤ 0.50

− fc(NH3)biogas

0.10 ≤ wMS/biomass ≤ 0.20 0.50 ≤ wVS/biomass ≤ 0.80

MWC MWC − fc(CO2 ) MWCH4 MWCO2

3MWH − fc(H 2S)biogas MWNH3

− fc(O2 )biogas

(8)

(17)

fc(H 2O)digestate = (1 − wMS/biomass)Fbiomass − fc(H 2O)biogas

A mass balance determines the composition of the digestate

(18)

consisting of the residue once the biogas is produced.12 Defra’s17 studies show the typical composition of the digestate

The energy balance to the digester is as follows:

for the two manure types (cattle and pig, subscript CS and PS

Q digestor = ΔHreaction − Fcp(Tdigestor − Tin)

in equations below) considered. The digestate composition is

ΔHreaction =

computed with eqs 9−18: w′C/ k = R C − N/ k(w′Norg / k + w′Nam / k )

∑ ΔHcombus − ∑ prod

reactants

ΔHcombus (19)

Digestate is sent to the ponds and, if needed, more water is added.

(9) 6758


Research Article


Figure 3. Biogas reforming stage.

Figure 4. Syngas composition adjustment.

The traces of NH3 and H2S in the biogas are removed using packed bed columns that operate at 50 °C. A mixed bed made of Fe2O3 and zeolites is assumed, since it is up to the zeolites to adsorb the NH3, while the Fe2O3 removes the H2S. In this last case, the following chemical reaction takes place:18

Q prod =

⎛

∑ fc(i ,Reformer,Mix2)⎜ΔHf + ∫ ⎝

i

Tout

Tref

⎞ Cp dT ⎟ ⎠ (20)

Q reac =

Fe2O3 + 3H 2 → Fe2S3 + 3H 2O

⎛

∑ fc(i ,Reformer,Mix2)⎜ΔHf + ∫ ⎝

i = in

Tref

Tin

⎞ Cp dT ⎟ ⎠ (21)

While the removal of ammonia is modeled just assuming an efficiency of 100%, the removal of H2S is modeled following the stoichiometry of the reaction assuming complete conversion of the sour gas. The bed of Fe2O3 can be regenerated using oxygen:18

Q (Methane reformer) = Q prod − Q reac

(22)

fc(CH4,Spl3,Reformer) × LHVgas = Q (Methane reformer) + Q (He1)

(23)

We can oversize the biogas production section to produce the biogas needed for the production of biodiesel. However, economically this is not even promising due to the large cost of the digester. Simple energy balances are used to model the heat exchanger, and the compressors are modeled as polytropic. The model of the methane reformer is based on the chemical equilibria given by eqs 24−26

2Fe2S3 + 3O2 → 2FeO3 + 6S

Biogas Reforming. The production of methanol is carried out from syngas that is obtained via dry or hybrid dry and steam reforming of the biogas. This is an endothermic process. Therefore, part of the biogas is required to provide the energy to heat up and maintain the operating temperature of the furnace; see Figure 3. The fraction of biogas burned is computed by assuming that the energy required to heat up and to perform the reforming reaction isothermally is provided as the lower heating value of the methane within the biogas. 6759

CH4 + CO2 ↔ 2CO + 2H 2

(24)

CH4 + H 2O ↔ CO + 3H 2

(25)

CO(g) + H 2O(g) ↔ CO2 (g) + H 2(g)

(26)


Research Article


Zeolite 5A or 13X. This bed operates at 25 °C and 4.5 bar. Thus, the gas is cooled down and compressed before being fed to the PSA system. The water condensed, if any, is removed from the system. The bed adsorbs 0.1 kg of CO2 per kg of zeolite. We assume a removal efficiency of 95%; see Figure 5.22,23 The unprocessed syngas is mixed with the stream exiting the adsorbing system.

Out of the three equations, only two are independent. Since the first equilibrium depends on the pressure, we use compressor 1 to adjust the pressure for the reaction. We model the reactor to compute the syngas composition based on the following atomic balances nCH4 + nCO2|in = nCH4 + nCO + nCO2|out 4nCH4 + 2n H 2O|in = 4nCH4 + 2n H 2 + 2n H 2O|out 2nCO2 + nCO + n H 2O|in = n H2O + nCO + 2nCO2|out

(27)

and on the chemical equilibrium, where the equilibrium constants (kp) are given by eqs 28−30 as follows19,20 kp = e[31.447 − 29580/ T ] =

PCOPH2 PCH4PCO2

kp = 10[−11650/ T + 13076] =

kp = 10[1910/ T − 1784] =

(28)

PCOPH2 3 PCH4PH2O

Figure 5. CO2 removal stage (based on ref 2).

(29)

PCO2PH2 PCOPH2O

Methanol Synthesis. Figure 6 presents the detail of the synthesis loop for methanol production. The syngas is fed to a reactor that uses a catalyst comprised of CuO−ZnO−AlO to produce methanol. The reactor considers the equilibria described by the following reactions:

(30)

The traces of hydrocarbons are removed using pressure swing absorption (PSA); see Figure 4. A bed of silica gel is selected for this task. The operating conditions are 25 °C and 4.5 bar. Two units operate in parallel to secure continuous operation of the downstream process. One of the beds is operating and the second one is in regeneration. To adjust the pressure and temperature of the raw syngas, the gas stream is compressed and cooled down. As before, the compressors are modeled as polytropic ones and the heat exchangers assume saturated gas after water condensation. Thus, the condensed water is removed. We use Antoine correlations to compute the water that remains with the gas after condensation. The pressure drop across the bed is assumed to be 10% of the operating pressure as a design condition based on previous work.1 The composition of the syngas may need to be further tuned in terms of the H2 to CO ratio for methanol production. We consider the use of three technologies. Two of them can operate in parallel simultaneously. The first one is a water gas shift reactor. We model this reactor by assuming molar mass balances and the equilibrium as given by eq 26. The second technology is a bypass, and the third one consists of a hybrid membrane/PSA for H2 recovery. The bed of this adsorbent unit is made of oxides. For this technology, the gas is recompressed up to 4.5 bar. The adsorbent bed operates at 25 °C. The hydrogen separation section is modeled by assuming that 100% of the hydrogen in the stream that is processed through this system can be recovered by means of using a palladium membrane that is permeable to hydrogen alone and the PSA system to retain any other gas that may accompany it. The split fraction depends on the performance of the reformer, which may make this composition adjustment redundant. See the work of Hernández and Martı ́n2 for further details. The last gas purification stage consists of the removal of CO2. However, the role of the CO2 in the synthesis of methanol results in the need to maintain a certain amount in the syngas, typically from 2% to 8% in volume.21 Therefore, a bypass system is also considered. A fraction of the syngas is processed through a PSA system to remove the excess of CO2 using

CO + 2H 2 ↔ CH3OH CO2 + H 2 ↔ CO + H 2O

(31)

The reaction is favored by low temperatures and high pressures, but lately pressures about 50−100 bar and temperatures in the range of 200−300 °C are the most common. For the reactor to operate properly there are a few constrains in the composition of the feed, such as the ratio between hydrogen and CO, given by eq 3224 1.75 ≤

H2 ≤3 CO

(32)

as well as the fact that the concentration of CO2 should be between 2% and 8%21,25 and the ratio of the syngas components involving CO224,26 given by eq 33 1.5 ≤

H 2 − CO2 ≤ 2.5 CO + CO2

(33)

The operation of the reactor is modeled by considering the mass balance to the species involved (eqs 34−36) and the equilibrium constants given by the experimental results27,28 with P in bars and T in K as presented by eqs 37 and 38 H: 2n H2 + 2n H2O|in − (2n H2 + 2n H2O + 4nCH3OH)|out = 0 (34)

C: nCO + nCO2|in − (nCO + nCO2 + nCH3OH)|out = 0

(35)

O: nCO + 2nCO2 + n H 2O|in − (nCO + 2nCO2 + n H 2O + nCH3OH)|out =0 6760

(36) DOI: 10.1021/acssuschemeng.7b01007 ACS Sustainable Chem. Eng. 2017, 5, 6756−6767

Research Article


Figure 6. Methanol synthesis section.

Figure 7. Algae oil section.

come from Xin et al.,4 but also results from the work by Amini et al.29 are also included. The total phosphorus (TotP) and the total nitrogen (TotN) are given in milligrams per liter.

[PCH3OH] [PCO][PH2]2 −3

= 10[3971/ T − 7.492logT + 1.77 × 10

T − 3.11 × 10−8T 2 + 9.218]

(37)

⎛ g ⎞ growth ⎜ 2 ⎟ = 0.418528 × TotP + 0.52762 × TotN ⎝m d⎠

⎡ 5639.5 = exp⎢13.148 − − 1.077 ln T ⎣ [PCO2][PH2] T

[PCO][PH2O]

− 5.44 × 10−4T + 1.125 × 10−7T 2 49170 ⎤ + ⎥ T2 ⎦

+ 0.225013 × TotN × TotP − 0.20754 × TotP2 − 0.03026 × TotN2

(39)

Apart from nutrients, CO 2 is also consumed. The consumption rate depends on the growth rate of the algae, as given by eq 40:30

(38)

The unreacted gases are separated in a flash, recycling the syngas and recovering methanol and water. Methanol may need to be further purified depending on the performance of the reactor. In such a case, a system of molecular sieves or a distillation column, in an orange rectangle in Figure 6, can be used to remove the water formed in the synthesis. A fraction of the methanol will be used for the transesterification of the oil while the rest is sold to the market. Alternatively, if no enough methanol is produced out of the biogas, it has to be injected in the system at a cost. Oil Production. Figure 7 presents the details of the algaegrowing section of the flowsheet. The algae growth depends on a number of parameters, such as carbon source, water, nutrients, and light. In particular, the concentration of nutrients in terms of total nitrogen and total phosphorus determines the growth rate. We first develop a correlation for the effect of the nutrients’ concentration on algae growth using data from several studies. Most of the data

⎛ m3 ⎞ ⎛ g ⎞ CO2 ⎜ ⎟ = 0.6565 × growth ⎜ 2 ⎟ + 5.0784 ⎝m d⎠ ⎝ d ⎠

(40)

The amount of water required can be provided by two sources. The digestate from the bioreactor carries water, but an additional source may also be needed to meet an algae concentration of 0.006 kg per kg of biomass.31 The energy consumed for the operation of the ponds is computed following Sazdanoff’s data.30 Next, algae are harvested and dried up to 5% moisture content using Univenture’s design. The energy consumption of this stage is 40 W for 500 L/h of flow.32 We assume that the water can be recycled to the pond so that water is lost only by evaporation. The nutrients are provided by the liquid digestate. The oil, representing 55% of the dry biomass, is extracted using a solvent, cyclohexane, and a mechanical press, and the 6761


Research Article


Solution Procedure. The problem for the conceptual design of the integrated facility that produces biodiesel with no need for fossil intermediates is formulated as a large non-linear programming problem (NLP) with 5000 variables and 3500 equations writen in GAMS. The objective function includes the credit from biodiesel and that of the methanol in the case where an excess is produced, and we need to pay for utilities such as power or steam for certain heat exchangers. Note that HX3 is also minimized as a cost; however, the energy for the reformer and for this heat exchanger comes from the use of part of the biogas. The idea of adding a cost for methanol is simply in case the system cannot produce enough methanol, to reduce the external need for it.

solvent and the oil are separated in a vacuum distillation column. Starch is considered to be 35% of the algae dry mass and the rest protein. Next, the solvent and the oil are separated in a vacuum distillation column so that the bottoms are below 350 °C to avoid oil decomposition. The oil is sent to transesterification and the solvent is recycled. Biodiesel Synthesis and Transesterification. Martı ́n and Grossmann16 showed that the use of heterogeneous catalysis was promising due to the easier product separation. There is no need for biodiesel washing, reducing consumption of process water. Furthermore, the use of a heterogeneous catalyst provides flexibility to use different oil sources. A surface response model from previous work is used to compute the yield of the reaction (eq 41). Table 1 shows the lower and upper bounds of the operating conditions.

Z = fc(FAME) − C Power ∑ WCompres_i − CSteam ∑ Q HX_j i

yield = −73.6 + 2.5TTrans + 24.9 × Cat + 8.8(ratio methanol) − 0.01TTrans 2

j

− CMeOH × fc(MeOHadded) 2

The problem is solved to optimality using a multistart optimization approach with CONOPT 3.0 as the preferred solver. The model is solved in a Windows 7 professional machine, with an Intel Core i7. Twenty-five CPU minutes where needed to find convergence. Next, we design the heat exchanger network using Yee and Grossmann’s model33 and a water network.34

2

− 1.29 × Cat − 0.39(ratio methanol) − 0.26TTransCat

(41)

Table 1. Range of Operation of the Variables for Heterogeneous Catalysis variable

lower bound

upper bound

temp (T, °C) ratio methanol (mol/mol) amount catalyst (Cat, %)

40 6 1

60 12 4

(42)

■

RESULTS Mass and Energy Balances. The plant processes 100 kg/s of waste, 50% cattle manure and 50% pig manure, producing 88 Mgal/yr of biodiesel and 46.7 kt/yr of methanol, typical for algae-based biodiesel facilities in the literature.5 Note that the farm size to produce such waste is quite large, which involves a logistic problem that is not considered in this work. A total of 45% of the methane needs to be used to provide energy for the reforming process. As a result, the recommended biogas

The excess of methanol is recovered using a distillation column. Short cut methods are used to model this column, taking into account the presence of two liquid phases, a polar and a nonpolar one. To avoid the bottom reaching 150 °C, a temperature that can decompose glycerol, the column must operate under vacuum. A variable reflux ratio from 2 to 3 is allowed on the basis of offline simulations using CHEMCAD. Next, polar and nonpolar phases are separated at 40 °C. The glycerol phase may not need further purification, but biodiesel requires further treatment before being sold. A distillation column operating under vacuum is used to avoid reaching a temperature above 250 °C in the distillate. Figure 8 shows a flowsheet for this section of the facility.

Table 2. Main Operating Variables of Selected Units unit/oper number N(mg/L) P (mg/L) T (°C)

Pond 15000 56 45 ambient

Figure 8. Biodiesel production section (based on ref 39). 6762


Research Article

ACS Sustainable Chemistry & Engineering Table 3. Main Operating Variables of Selected Units (T, top; B, bottom) unit/oper condition T (°C)

dry reform 1000

transesterification reactor 60

P (bar) H2/CO2 MeOH/Oil CH4 fuel (%)

1.084

4

oil column T: 31 B:350 135/760

MeOH column T: 41.5 B:150 300/760

biodiesel column T: 243 B:274 45/760

MeOH purification T: 63 B: 101 1

MeOH synthesis 200 50 2.5

6.541 45

composition is 67.5% methane, 30% CO2, and the rest nitrogen, H2S, ammonia, and oxygen. Table 4 shows the Table 4. Power Consumption of the Units Involved unit

W (kW)

Ponds Belt Press Compres1 Compres2 Compres3 Compres4 Compres5 Compres6

1500 838 3813 564 −605 1146 58 1666 691

Figure 9. T−Q curve for the integrated facility.

Table 5. Thermal Energy Involved in the Units unit

kW

unit

kW

Digester HX1 HX3 HX4 HX5 HX6 HX7 HX8 HX9 HX10 HX11 HX17 HX18 HX19

9.460 −562 4.623 −5.245 −435 54 −360 −688 464 −2.782 −403 −6.270 −177 −2.318

HX20 HX21 HX22 HX23 HX24 HX25 HX26 HX27 HX28 HX29 Reformer RWGS RMeOH TransReact

680 −1.059 1.748 −7.773 5.455 −1.983 −95 343 −7905 3983 6.978 −118 −2.983 4.276

consumption would drop to 0.8 galwater/galbiodiesel. Furthermore, the process captures 3 kg of CO2 per kg of biodiesel produced. Tables 2 and 3 show the operating conditions of the major equipment such as reactors, including the ponds, the furnace for the production of syngas, the methanol synthesis reactor, and the transeterification reactor. In all cases, the temperatures and pressures are within typical operating limits. For instance, the transesterification reactor requires a molar ratio of methanol to oil close to 6, and the operating temperature is 60 °C, close to the ones presented in previous works.32 For the production of methanol, low temperatures and pressures are suggested, 200 °C and 50 bar, to reduce the costs related to preparing the feed before synthesis. However, in this case, the fraction of methane required to provide energy for the reforming is larger than usual, in the range of 10−25%. It is related to the high hydrogen to CO ratio required and the added value that the methanol brings to the process. Economic Evaluation. For the evaluation of the investment cost we use the factorial method.35 The Matches Web page36 is used to estimate the cost of the chemical processing equipment. The correlation for estimating the cost of the digester can be found in the Appendix below and the ones for the rest of the units are available in the Supporting Information of Almena and Martı ́n.37,38 Figure 10 shows the distribution of the cost among the three major sections of the process. We see that biogas production, and in particular the digester, represents the major contribution to the investment cost. Other sections, such as methanol and biodiesel, show a similar share, around 11−17%. Biodiesel involves algae growing and harvesting as well as oil transesterification, while the methanol section includes reforming, gas cleanup, and composition adjustment, followed by synthesis and purification stages. To estimate the investment cost, the installed equipment represents 1.5 times the equipment cost. Piping, insulation, instrumentation, and

power requried by the units across the flowsheet, while Table 5 presents the thermal energy invovled before integration. Note that the reformer actually uses a fraction of the biogas as fuel. Furthermore, the amount of energy required, after heat integration and the development of a HEN, is 4.25 MJ/s, 1.5 MJ/gal of FAME. In terms of cooling, 27 MJ/s is required, together with the energy result in the consumption of 1.22 galwater/galbiodiesel by developing the corresponding water network, where we consider a boiler and a cooling tower as units that require water and generate also wastewater as presented in previous work.34 Figure 9 shows the T−Q curve for the integrated system. A total of 17 MW was integrated by means of the formulation and the development of the heat exchanger network. There is enough energy to provide for the needs at the digester, but control challenges may require the use of steam. Thus, we do no integrate them in the T−Q curve. However, if that energy is used to provide the energy for the digester, 18 MW of cooling would be necessary and the water 6763


Research Article


optimal source of energy and the operating conditions over a year for an average production of 60 Mgal/yr of biodiesel. The facility requires an investment of 120 M€ for a production cost of 0.79 €/gal (0.25 €/kg). However, that facility production capacity is smaller than the one in this study. To scale up the cost, we make use of the fact that the investment cost is computed using the factorial method.35 This method relies on the equipment cost, where the investment cost is computed using factors to account for installation, piping, isolation, etc. Therefore, if the cost estimation of the units is a function of the mass or energy flows and accounts for rules of thumb, such as the fact that if a flow is larger than standard, the unit must be duplicated, i.e., heat exchangers larger than 2000 m2 are not commercially available, the entire plant cost can estimate the effect of the scale. In our studies, the equipment cost is estimated using the correlations developed in previous work as a function of the production capacity.37 By increasing the production capacity, we easily recomputed the unit cost and the investment of the scaled-up facility. The procedure is explained in detail in a previous work.40 By scaling this facility to the same size as this work, the production cost decreases to 0.76 €/gal and the investment increases to 144 M€. The energy consumption is 3.2 MJ/gal biodiesel, capturing 4.05 kg of carbon dioxide per kg of biodiesel and consuming 1 L/L of water, including the operation of the cooling tower and the boiler. As we see, there are a number of trade-offs, since this process is cheaper than the one in this work and consumes less freshwater, but with a larger energy consumption. The second comparison consists of the use of switchgrass to produce methanol and algae for the biodiesel (FAME). A superstructure optimization approach is used to select the optimal integrated topology, gasification, and reforming technologies that provide the thermal energy and the methanol that biodiesel production requires. The excess of methanol can be sold as biofuel or it can be further processed to gasoline. The optimal integrated process involves indirect gasification followed by steam reforming based on the need for a high H2 to CO ratio to produce methanol and avoid the use of oxygen. The integrated process produces 776 ML/yr (205 Mgal/yr) of biofuels, 34% FAME, at 0.14 €/L (0.53 €/gal). The plant does not emit CO2, but captures 1.27 kg of CO2 per kg of methanol produced. The investment costs adds up to 180 M€. By scaling this size to the one in this work, the production costs goes up to 1 €/gal and the investment decreases to 134 M€.6 Finally, although it works with ethanol instead of methanol, the process proposed by Martı ́n and Grossmann5 produces ethanol from the starch in the algae and oil from the lipids. In that work, for the production of 90 Mgal/yr of biodiesel, FAEE, the production cost results in 0.35 $/gal, using enzymes, with energy and water consumption values of 4.00 MJ/gal and 0.59 gal/gal. The investment cost is close to the one in this work, 180 M$. Table 6 summarizes the comparison.

Figure 10. Breakdown of the contribution of the different sections to the overal CAPEX.

utilities represent 20%, 15%, 20%, and 10% of the equipment cost, respectively. The capital expenditure (CAPEX) adds up to 193 M€. To estimate the production cost of the fully renewable FAME, we consider labor, taxes, administration, utilities, and equipment maintenance.35 Figure 11 presents the breakdown of

Figure 11. Breadown of the production costs excluding credits.

the production costs. Note that most of the cost is related to amortization and maintenance (equipment in Figure 11). Utilities represent the second contribution. Cooling water and steam add up to 2.5 M€/yr. The total production cost is 40.5 M €/yr. By selling the excess of methanol at 0.3 €/kg we can obtain a credit of 13 M€/y. Thus, the production costs for the 88 Mgal/yr is $0.31/gal. Comparison with Other Integrated Processes. The first comparison is with a conceptual design by Martı ́n and Grossmann.7 That facility produces biodiesel using solar and/ or wind energy and carbon dioxide. Microalgae are grown to accumulate lipids that are extracted. The methanol needed for the transesterification is synthesized from carbon dioxide and electrolytic hydrogen. The electricity for the complex is produced using solar panels or wind turbines. The flowsheet is formulated as a multiperiod mixed integer nonlinear programming problem, the solution of which provides the

Table 6. Comparison among Self-Sufficient Processes for Biodiesel Production process

investment (M€)

production cost (€/gal)

kg CO2 captured per kg produced of biofuel

energy consumed (MJ/gal of biodiesel)

this work ref 7 ref 5

193 144 180

0.31 0.76 0.53

3 4.05 1.27

1.5 3.2 1.94

ref 5

170

0.33

2.81

4 6764

water consumed (gal/gal biodiesel) 1.22/0.8 1 not applicable; high cooling needs 0.59 DOI: 10.1021/acssuschemeng.7b01007 ACS Sustainable Chem. Eng. 2017, 5, 6756−6767

Research Article


Figure 12. Cost vs scale of the biodiesel facility.

growth rate as a function of the nutrients has also been developed. This framework exploits the integration of up to four different sections that are typically evaluated separately, such as biogas production, methanol production, algae growing, and biodiesel production, but that they are strongly linked. The production of biodiesel from manure is promising from the production cost and environmental points of view, since it is competitive with other integrated processes that simultaneously produce methanol and oil. The investment cost is larger, at least 30% more, due to the digestion and gas reforming processes. However, the large amount of feedstock may jeopardize the implementation of this kind of facility. Most of the biogas plants are actually built on a smaller scale than the one proposed here. In this case, the process can be also applied if the biogas and digestate are fed externally. Thus, it opens a new research topic about the supply chain of biogas and nutirents, where further studies can be carried out.

In all cases, the nutrients were bought, while the use of the digestate as a source for fertilizers reduces not only the cost but also the actual impact. Scale up and down of the Facility: Manure Availability. The facility evaluated tried to reach a production capacity of biodiesel comparable to typical sizes in the literature. However, the feedstock required represents the manure produced by around 300 000 animals.41 Therefore, in this section we evaluate the economics and scale down of the facility to 3000 animals, around 1 kg/s of manure. The same procedure presented above on plant scale up and down is used.35 Figure 12 shows the production costs as a function of the farm size required. We see that for the fuel to have a price below 1 €/gal, a farm above 35 000 animals is required. eq 43 shows the correlation for the production cost as a function of the number of animals. No good correlation as a function of the scale can be obtained from the results of the investment cost. biodiesel prod cost = 1160.3(no. of animals)−0.664

■

APPENDIX34 The correlation for estimating the cost of the digester is as follows:

(43)

However, even though in the United States there has been an increase in the typical farms size, most farms do not reach 10042 cows, resulting in the fact that a detail supply chain study is required to evaluate the actual production of fully renewable diesel. However, instead of transporting manure as fertilizer, recently Sampat et al.41 evaluated the production of struvite, an easy to transport solid fertilizer that provides an interesting option toward the use of manure as fertilizer.

■

Cdigester (€) = 1475100 × [flow manure (kg/s)]0.82227

AUTHOR INFORMATION

Corresponding Author

■

*E-mail: [email protected]. ORCID

CONCLUSIONS AND FUTURE CHALLENGES In this work, we have developed an integrated process for the production of renewable diesel from manure via anaerobic digestion in search for making full use of the two main products, biogas and digestate. On the one hand, biogas is produced and further used to obtain renewable methanol via dry reforming. On the other hand, the digestate is used as nutrients to increase the algae yield. This point has two main implications: the need for input from other sources is reduced, and therefore the cost, but also the discharge of digestate can generate an environmental problem due to the high nitrogen and phosphorus content in farm regions. Thus, FAME is produced by transesterifying the algae oil with biogas-based methanol. The problem is formualted as a large NLP for the optimization of the operating conditions at the various reactors, such as the dry reformer, the methanol synthesis, the FAME production, and the algae growing, for which a model for the

Mariano Martín: 0000-0001-8554-4813 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors also thank Salamanca Research for the software licenses and the project DPI2015-67341-C2-1-R for financial support.

■ 6765

NOMENCLATURE a′ = CH4, CO2, NH3, H2S, O2, or N2 a = H2O, CH4, CO2, NH3, H2S, O2, or N2 b = C, Nam, Norg, P, K, or rest Cp = heat capacity (kJ/kgK) ΔHf = formation enthalpy (kJ/kg) fc = component mass flowrate (kg/s) F = total flowrate (kg/s) DOI: 10.1021/acssuschemeng.7b01007 ACS Sustainable Chem. Eng. 2017, 5, 6756−6767

Research Article

ACS Sustainable Chemistry & Engineering ni = molar flow rate of component i (kmol/s) Pi = partial pressure of component i (bar) Pv = vapor pressure (bar) P = total pressure (bar) Q = thermal energy (kJ/s) Vbiogas/k = biogas volume produced per unit of volatile solids (SV) (m3biogás/kgSV/i) RC−N/k = mass ratio between the carbon and nitrogen in the biomass wMS/k = dry mass fraction (kgMS/k/kg) w′SV/k = dry mass fraction of volatile solids out of the dry biomass (kgSV/k/kgMS/k) w′C/k = dry mass fraction of C in biomass (kgC/k/kgMS/k) w′Nam/k = dry mass fraction of ammonia nitrogen in biomass (kgNam/k/kgMS/k) w′Norg/k = dry mass fraction of organic nitrogen in biomass (kgNorg/k/kgMS/k) w′P/k = dry mass fraction of P in biomass (kgP/k/kgMS/k) w′K/k = dry mass fraction of K in biomass (kgK/k/kgMS/k) w′rest/k = dry mass fraction of the balance of the biomass (kgK/k/kgMS/k) Ya′/biogas_sec = molar fraction of the dry gas xa/biogas = mass fraction of the gas ybiogas = specific saturated moisture of biogas MW = molar mass (kg/kmol) T = temperature (°C or K) as specified in the text HEN = heat exchanger network FAME = fatty acid methyl ester LHV = lower heating value

■

(12) Al Seadi, T., Rutz, D., Prassl, H., Köttner, M., Finsterwalder, T., Volk, S., Janssen, R. Biogas Handbook; University of Southern Denmark Esbjerg: Esbjerg, Denmark, 2008. (13) Kaparaju, P.; Rintala, J. 17 - Generation of heat and power from biogas for stationary applications: boilers, gas engines and turbines, combined heat and power (CHP) plants and fuel cells. Biogas Handbook 2013, 404−427. (14) http://www.biogas-info.co.uk/biogas-biogas-yields.html. (15) Steffen, R.; Szolar, O.; Braun, R. Feedstocks for anaerobic digestion. http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1. 470.1914&rep=rep1&type=pdf (16) Gunaseelan, V. N. Anaerobic digestion of biomass for methane production: A review. Biomass Bioenergy 1997, 13 (12), 83−114. (17) Defra. Fertilizer manual. 2011 http://adlib.everysite.co.uk/adlib/ defra/content.aspx?id=2RRVTHNXTS.88UF9N65FWLCJ http:// adlib.everysite.co.uk/adlib/defra/content.aspx?doc=262994&id= 263068. (18) Ryckebosch, E.; Drouillon, M.; Vervaeren, H. Techniques for transformation of biogas to biomethane. Biomass Bioenergy 2011, 35, 1633−1645. (19) Luyben, W. L. (2014) Design and Control of the Dry Methane Reforming Process. Ind. Eng. Chem. Res. 2014, 53 (37), 14423−14439. (20) Roh, H.-S.; Lee, D. K.; Koo, K. Y.; Jung, U. H.; Yoon, W. L. Natural gas steam reforming for hydrogen production over metal monolith catalyst with efficient heat-transfer. Int. J. Hydrogen Energy 2010, 35 (3), 1613−1619. (21) Lee, S. Methanol Synthesis from Syngas. Handbook of Alternative Fuel Technologies; Taylor and Francis: Boca Raton, FL, 2007; Chapter 9, DOI: 10.1201/9781420014518.ch9. (22) Nexant Inc. Equipment Design and Cost Estimation for Small Modular Biomass Systems, Synthesis Gas Cleanup, and Oxygen Separation Equipment Task 2: Gas Cleanup Design and Cost Estimates−Black Liquor Gasification Subcontract; Report NREL/SR-510-39944, May 2006. (23) GPSA Engineering Data Book, FPS VERSION 21-10; GPSA, 2004. (24) Ribeiro, A.; Santos, J. C.; Rodrigues, A.E. PSA design for stoichiometric adjustment of bio-syngas for methanol production and co-capture of carbon dioxide. Chem. Eng. J. 2010, 163 (3), 355−363. (25) Tijm, P. J. A.; Waller, F. J.; Brown, D. M. Methanol technology developments for the new millennium. Appl. Catal., A 2001, 221, 275− 282. (26) Marechal, F.; Heyen, G.; Kalitventzeff, B. Energy Saving in Methanol Synthesis: Use of Heat Integration Techniques and Simulation Tools. Comput. Chem. Eng. 1997, 21, S511−S516. (27) Cherednichenko, V. M. Dissertation, Karpova, Physico Chemical Institute, Moscow, U.S.S.R., 1953. (28) Bissett, L. Equilibrium Constants for Shift Reactions. Chem. Eng. 1977, 84 (21), 155−156. (29) Amini, H.; Wang, L.; Shahbazi, A. Effects of harvesting cell density, medium depth and environmental factors on biomass and lipid productivities of Chlorella vulgaris grown in swine wastewater. Chem. Eng. Sci. 2016, 152, 403−412. (30) Sazdanoff, N. Modeling and Simulation of the Algae to Biodiesel Fuel Cycle. Undegraduate Thesis. The Ohio State University, Columbus, OH, 2006. (31) Pate, R. Biofuels and the Energy-Water Nexus AAAS/SWARM April 11, 2008 Albuquerque, NM. 2008. (32) Martín, M.; Grossmann, I. E. Simultaneous optimization and heat integration for biodiesel production from cooking oil and algae. Ind. Eng. Chem. Res. 2012, 51 (23), 7998−8014. (33) Yee, T. F.; Grossmann, I. E. Simultaneous optimization models for heat integration − II. Heat exchanger networks synthesis. Comput. Chem. Eng. 1990, 14, 1165−1184. (34) Ahmetović, E.; Grossmann, I. E. Global superstructure optimization for the design of integrated process water networks. AIChE J. 2011, 57 (2), 434−457. (35) Sinnot, R. K. Coulson and Richardson, Chemical Engineering, 3rd ed.; Butterworth Heinemann: Singapore, 1999.

REFERENCES

(1) León, E.; Martín, M. Optimal production of power in a combined cycle from manure based biogas. Energy Convers. Manage. 2016, 114, 89−99. (2) Hernández, B.; Martín, M. Optimal Process Operation for Biogas Reforming to Methanol: Effects of Dry Reforming and Biogas Composition. Ind. Eng. Chem. Res. 2016, 55, 6677−6685. (3) Hernández, B.; León, E.; Martín, M. Bio-waste selection and blending for the optimal production of power and fuels via anaerobic digestion. Chem. Eng. Res. Des. 2017, 121, 163−172. (4) Xin, L.; Hong ying, H.; Ke, G.; Ying-xue, S. Effects of different nitrogen and phosphorus concentrations on the growth, nutrient uptake, and lipid accumulation of a freshwater microalga Scenedesmus sp. Bioresour. Technol. 2010, 101, 5494−5500. (5) Martín, M.; Grossmann, I. E. Optimal engineered algae composition for the integrated simultaneous production of bioethanol and biodiesel. AIChE J. 2013, 59 (8), 2872−2883. (6) Martín, M.; Grossmann, I. E. Optimal Integration of Algae− Switchgrass Facility for the Production of Methanol and Biodiesel. ACS Sustainable Chem. Eng. 2016, 4, 5651−5658. (7) Martín, M.; Grossmann, I.E. Optimal integration of a self sustained algae based facility with solar and/or wind energy. J. Cleaner Prod. 2017, 145, 336. (8) Elsholkami, M.; Elkamel, A.; Vargas, F. Optimized integration of renewavle energy technologies into Alberta’s oil Sands industry. Comput. Chem. Eng. 2016, 90, 1−22. (9) Shahandeh, H.; Li, Z. Design of Bitumen Upgrading and utility plant through integrated optimization. Ind. Eng. Chem. Res. 2017, 56 (8), 2107−2126. (10) Martín, M.; Davis, W. Integration of wind, solar and biomass over a year for the constant production of CH4 from CO2 and water. Comput. Chem. Eng. 2016, 84, 313−325. (11) Betancourt-Torcat, A.; Elkamel, A.; Ricardez-Sandoval, L. Optimal integration of nuclear energy and water management into the oil sands operations. AIChE J. 2012, 58 (11), 3433−3454. 6766


Research Article

ACS Sustainable Chemistry & Engineering (36) Matches Index of Process Equipent. http://www.matche.com/ equipcost/EquipmentIndex.html. (37) Almena, A.; Martín, M. Techno-economic analysis of the production of epiclorhidrin from glycerol. Ind. Eng. Chem. Res. 2016, 55 (12), 3226−3238. (38) Cost of Anaerobic Digestion Plant. Very Approximate AD Plant Cost Estimates.http://anaerobic-digestion.com/anaerobic-digestionplants/cost-of-anaerobic-digestion-plant/. (39) Martín, M. Industrial Chemical Process Analysis and Design; Elsevier: Oxford, 2016. (40) Martín, M. Artifical vs natural reuse of CO2 for DME production. Are we getting any closer? Engineering 2017, 3, 166−170. (41) Sampat, A. S.; Martín, E.; Martín, M.; Zavala, V. M. Optimization Formulations for Multi-Product Supply Chain Networks. Comput. Chem. Eng. 2017, 104, 296−310. (42) Zoetis. Qué Pasa con el Tamaño de las Granjas Lecheras a Nivel Mundial? http://blog.especialistasennovillas.es/posts/tamano-granjaslecheras.aspx (accessed June 2017).

6767


Optimal Integrated Plant for Production of Biodiesel from Waste

Recommend Documents