Biomass cogeneration plants integrated into poultry slaughterhouses for reducing industry costs with energy

A technical and economic feasibility analysis was performed concerning biomass cogeneration to supply the thermal and electricity demands of poultry slaughterhouses. The analysis considers measured data referring to the annual energy consumption from an existing industry as well as the characteristics of equipment available in the Brazilian market. The cogeneration plant is equipped with a water tube steam generator and a condensing-extraction steam turbine in a Rankine cycle. Four different configurations were evaluated, including impulse and reaction turbines at two steam pressure/temperature levels (43 bar / 450 °C and 68 bar / 520 °C). A steady state full load operation is considered at cogeneration mode on the weekdays and at Rankine power plant mode on the weekends, when there is no process steam consumption. The technical analysis pointed out the reaction turbine at 68 bar / 520 oC as the best alternative, leading to the highest overall efficiency. In addition, this plant configuration showed economic advantages represented by an Internal Rate of Return (IRR) of 21%, a Net Present Value (NPV) of US$ 10.93 million, and a payback time of 6 years, enabling a reduction on the industrial cost with energy in the slaughterhouse to 19 US$/ton of product (-30% in comparison to the base case). Finally, the calculated LCOE of 73 US$/MWh was lower than the current price of the electricity in the market, indicating potential economic feasibility of the proposed concept.


Introduction
The meat industries require a large amount of process steam and electricity to drive motors, compressors, conveyors and lighting (Bueno, Rossi, Souza, Teruel, & Bueno, 2015;Feliciano, Rodrigues, Gonçalves, Santos, & Leite, 2014;Nunes, Silva, Andrade, & Gaspar, 2016;Ramírez, Patel, & Blok, 2006;Rocha, Bajay, & Gorla, 2010). In Brazil, wood chips of Eucalyptus or Pinus are used in low pressure saturated steam generators, while the electricity is imported from local distributors or from the open electricity market. Although in Brazil the cogeneration technology is consolidated, its use in the meat industries still does not occur; therefore, this is the motivation of this work.
Cogeneration power plants have been widely used in the world for industrial, commercial and district heating or cooling applications, using any liquid, gaseous or solid fuels, as well as industrial process waste gases. Typical prime movers are steam turbines, gas turbines, reciprocating engines and combined cycles. By using cogeneration plants, the industry revenues could be increased by exporting surplus electricity to the local grid (Alves, Ponce, Silva, & Ensinas, 2015). In addition, the industry expenses with electricity, heating and cooling can be reduced. For example, Silva, Higa and Silva (2019) identified the opportunity of reducing the costs with energy by up to 72% in a payback period of 24 months in a soluble coffee industry in Brazil, by integrating a cogeneration plant equipped with a condensing-extraction steam turbine and an ammoniawater absorption cooling system.
Previous works have been reported focusing on cogeneration applied to the meat industry. As examples, Bianchi, Cherubini, De Pascale, Peretto and Elmegaard (2006) proposed the use of poultry industry wastes' energy content to run an indirectly fired gas turbine cogeneration plant application. Huang et al. (2015) presented a technical and economic analysis of producing biochar and generating electricity and heat from poultry litter. Compton, Willis, Rezaie and Humes (2018) presented several energy saving measures (including the use of cogeneration plants) to be applied in the food industries in US. Furthermore, Philipp et al. (2018) evaluated different energy supply structures for industrial food processing sites in different countries, showing the advantages of cogeneration. Regarding cogeneration plants integrated into poultry slaughterhouses, Sordi, Souza, Galdino and Oliva (2002) proposed the use of a biomass-fueled plant equipped with a back-pressure steam turbine in order to produce the own electricity and to supply the process with saturated steam. It's important to observe, nevertheless, that the use of a back-pressure steam turbine linked the generation of electricity to the intermittent consumption of process steam, what potentially leads to a reduction of plant's lifetime (Azimov, Shkret, & Garievskii, 2016). Thus, further investigation is demanded. Different plant configurations are possible and might be tested to find more economic and reliable concepts applied to the Brazilian scenario.
In this regard, the objective and novelty of this work consists on providing the reliable use of wood chips on cogeneration plants as an alternative for supplying the thermal and electricity demands of the poultry meat industries, considering the application of condensing-extraction steam turbines (CEST) to provide the full-time operation of plant and electricity surplus export to the grid. Four different configurations were evaluated, considering impulse and reaction steam turbines, for two levels of steam pressure/temperature (43 bar / 450°C and 68 bar / 520°C). The simulations consider the corresponding data carried out on an existing poultry slaughterhouse during the year 2017, as well as performance and budget data provided by equipment manufacturers in Brazil. The obtained results showed that the industry cost with energy can be significantly reduced by applying the proposed concept.

Conceptual design of the cogeneration plant
In Brazil, approximately 60% of poultry production occurs in southern region, 20% in the southeast region and around 15% in the central-western region (Instituto Brasileiro de Geografia e Estatística [IBGE], 2019). In places where the poultry slaughterhouses are located, there is no natural gas infrastructure, the process steam demand is provided by low pressure saturated steam boilers using Eucalyptus or Pinus wood chips as fuels and electricity is purchased from the distributors. In this regard, it is proposed in this work the application of biomass Rankine cogeneration systems using condensing-extraction steam turbines ( Figure 1) to supply the heat and electricity demands of poultry slaughterhouses. Superheated steam is produced in the steam generator and is directed to the steam turbine (1). Part of steam mass flow is extracted at (2) and directed to a desuperheater (5) where it is mixed with liquid water (17). Then, the saturated steam (7) is directed to the process, while condensate is returned to the deaerator (9). The remaining steam mass flow in the turbine is expanded until the condenser (3), where condensation occurs by the action of the cooling water (18). The condensate is then pumped to the deaerator (12), while cooling water returns to the cooling tower (19). The steam extraction feeds the process demand and the steam (11) required by deaerator. Make-up water (10) is used to supply the losses in the plant. From the deaerator, feedwater is directed to the boiler (15).
The plant can be operated in cogeneration mode when there is poultry processing and in Rankine power plant mode when there is no process steam consumption. The continuous operation of the plant despite the process steam demand is possible by selling surplus electric energy to the Brazilian open or regulated markets.

Thermodynamic modelling
The thermodynamic performance of steam cycle was calculated by a component-wise modelling based on the first law and mass conservation equations followed by a simulation using the software EES®.

Steam generator
The performance of the steam generator is evaluated according to the energy balance described in Equation 1, is the thermal efficiency of the steam generator; ℎ ℎ and ℎ [kJ kg -1 ] are the specific enthalpy of superheated steam and feedwater, respectively; ̇ℎ and ̇ [kg s -1 ] are the superheated steam and feedwater mass flows; ̇ [kg s -1 ] is the biomass consumption and [kJ kg -1 ] the biomass low heat value.

Steam turbine
The net power output of a turbine stage is calculated according to Equation 2, where ̇ ( ℎ − ℎ ) is the energy output of the expanded steam [kW], , is the efficiency of mechanical system (bearing, coupling, and heat loss) and , is the electrical efficiency of the generator.
The output enthalpy of the exhaust steam is achieved considering the isentropic efficiency, , , of the turbine stage -Equation 3 (Shlyakhin, 2005), where the subscript denotes the output state for the isentropic process.

Deaerator
As it can be seen in Figure 1, the proposed cogeneration plant is equipped with a deaerator system to remove dissolved gases from the circuit to protect the plant from corrosion. Deaerator is modeled in this work considering Equation 4 (Borgnakke & Sonntag, 2019).

Pumps
The auxiliary electricity power required for a pump operation, ̇ [kW], is modelled according to Equation 5, where ̇ ( ℎ − ℎ ) represents the energy gain of the pumped fluid [kW], , is the efficiency of mechanical system (bearing, coupling, and heat loss), and , is the efficiency of the electric motor.
The output enthalpy of pumped fluid is achieved considering the isentropic efficiency of pump, , , as it is stated by Equation 6 (Borgnakke & Sontagg, 2019).

Condenser and cooling tower
It is considered that the cogeneration plant is equipped with a condenser and cooling tower system (see Figure 1). The condenser and the cooling tower are designed for the maximum turbine exhaust steam mass flow. The temperatures involved in condenser and cooling tower systems operation are indicated in Figure 2.
The energy balance in the condenser is performed to calculate the cooling water mass flow demand, ̇ [kg s -1 ], as stated by Equation (7), where ̇ [kg s -1 ] is the turbine exhaust steam mass flow, ℎ , and ℎ , [kJ kg -1 ] are the inlet and outlet steam enthalpies, and ℎ , and ℎ , [kJ kg -1 ] are the inlet and outlet cooling water enthalpies.
The cooling tower uses atmospheric air in countercurrent with the cooling water to reduce its temperature. Equation 8 represents the energy balance in the tower, where ̇ [kg s -1 ] is the air mass flow, , and , [kg kg -1 -kilogram of water per kilogram of dry air] are the absolute humidity of air at the inlet and outlet of the tower, ℎ , , and ℎ , , [kJ kg -1 ] are the inlet and outlet enthalpies of dry air, and ℎ , , and ℎ , , [kJ kg -1 ] are the inlet and outlet enthalpies of air moisture content.
Equation 9 is used to determine the mass flow rate of make-up water for losses by evaporation and by dragging of droplets and purges. The make-up water mass flow is normally around 2 to 5% of the total flow (Lora & Nascimento, 2004). Droplet drag and purge losses are considered to be 1.3% of the mass flow rate of cooling water. ̇( − ) +̇ 1,013 = ̇ (9) Figure 2. Condenser and cooling tower operation scheme. Adapted from Lora and Nascimento (2004).

Process heat demand
In this work the process heat demand, ̇ [kW], is calculated according to Equation 10, where ̇ [kg s -1 ] process steam mass flow (Figure 1, point 7) and ℎ ℎ [kJ kg -1 ] the steam latent heat.  The concept here proposed is applied to an existing typical poultry slaughterhouse plant located in the state of Parana, Brazil. The plant has the capacity to slaughter 500 thousand poultries per day and is operated from Mondays to Fridays (24h day -1 ), with Saturdays and Sundays reserved for scheduled maintenance services.
Data related to process operation was collected in the plant, including final product output, electricity bills, electricity and steam mass flow demands, biomass consumption, etc. A sample of the collected data is presented in the Appendix A, while Appendix B presents the energy consumption and demand charges applied locally by the distributor during the period when this work was developed.
During the year of 2017, the final product output was 312.2 thousand tons. Regarding electricity consumption, it was 66,962 MWh at an average price of 111 US$ MWh -1 (with peak electricity demand limited to 12 MW).
The process steam consumption was on average equal to 14 t h -1 (at 10 bar) during operation days (16 t h -1 peak). The low-pressure steam generator was fueled with eucalyptus chips purchased at an average price of 45.19 US$ t -1 . In 2017 the biomass consumption was 19,833 tons, leading to the annual Levelized Cost of Heat (LCOH) of 21.87 US$/MWh.
In this regard, the industrial costs with thermal energy and electricity were 3.35 US$ ton -1 and 23.89 US$ ton -1 , respectively, leading to the total of 27.24 US$ ton -1 .

Cogeneration plant design
The plant design was performed considering the parameters presented in Table 1. The plant capacity was set to 12 MW, while steam extraction (Figure 1, point 2) mass flow was calculated to feed the process saturated steam demand (limited to 16 t h -1 , 10 bar, x=1) and deaerator. The plant is operated at constant power output in the cogeneration mode during the weekdays and in the Rankine mode on weekends or holidays. In both operation modes, electricity surplus is delivered to the grid. Four plant configurations are compared, involving two pressure and superheated steam temperature levels (43 bar / 450°C and 68 bar / 520°C), as well as the use of impulse or reaction turbines. The configurations are: SG1/ST1 (boiler 43 bar / 450°C -impulse turbine), SG1/ST2 (boiler 43 bar / 450°C -reaction turbine), SG2/ST1 (boiler 68 bar and 520°C -impulse turbine) and SG2/ST2 (68 bar and 520°C boiler -reaction turbine). These configurations were selected to represent an important range of commercially available solutions in the Brazilian market. Technical data and the Capital Expenditure (CAPEX) for steam generators and steam turbines are presented in Tables 2 and 3. Acta Scientiarum. Technology, v. 43, e50967, 2021   Figure 1); c Point 2 to 3 (Figure 1).
Additional parameters considered for economic analysis are presented in Table 4.

Results and discussion
In Table 5, the results for design point operation of proposed concepts are presented. The SG2/ST2 configuration showed the highest thermal efficiency values, equal to 40.2 and 27.2% for the cogeneration and Rankine operation modes. In this regard, the SG2/ST2 configuration had the lowest biomass consumption, equal to 18.0 t h -1 during the week and 15.1 t h -1 at the end of the week. In the same way, this configuration led to the lowest cooling tower water consumption, equal to 45.0 t h -1 and 44.6 t h -1 for the cogeneration and Rankine operation modes. The best results of the SG2/ST2 are due to higher grade superheated steam and the higher isentropic efficiency of ST2 reaction turbine. It is also interesting to observe in Table 5, the great increment of plant efficiency when, for the same superheated steam parameters, a reaction turbine (with higher isentropic efficiency) is considered in place of an impulse model. This is further discussed in sequence considering the economic analysis figures. The process diagrams related to the design point SG2/ST2 plant operation in cogeneration and Rankine modes are presented in Figures 3 and 4, respectively. As it can be observed, the superheated steam mass flow is reduced when there is no process heat consumption and additional mass flow is expanded to the last stages of turbine. This provides the part load operation of steam generator (84% load), which might not lead to significant impact on equipment performance. Moreover, during the process operation days, 41% of steam mass flow is extracted at point 2 at 10 bar, while 59% is expanded up to condensation pressure. On weekends, on the other hand, 13% of steam mass flow is extracted at point 2 at 10 bar, while 87% is expanded up to condensation pressure. These figures might be considered when specifying the steam turbine.  The results for the annual operation of plants are presented in Table 6. As it can be seen, the concept SG2/ST2 presented the highest thermal efficiency due to the higher grade of superheated steam and the higher isentropic efficiency of reaction turbine. This led to reduced annual biomass and make-up water consumption. Furthermore, the reduced biomass consumption translates into less land area demanded for the biomass crop. Under the base case scenario where process steam is generated in a low-pressure boiler, the land area demand calculated according to Equations 12 and 13 is estimated at 944 ha, while in the case of SG2/ST2, it might be expanded to 6,576 ha (~7x more land area demand). In this regard, land area availability is an important aspect to be considered when proposing to supply the heat and electricity demands of poultry industries.
Results related to the economic analysis are presented in Table 7. The Capital Expenditures (CAPEX) include the investment cost of the steam generator, steam turbine, cooling tower and water treatment system. As it can be seen, no significant CAPEX variation between the concepts is observed. In case of SG2/TG2, despite the use of higher pressure/temperature steam parameters and reaction turbine type, the improved thermal efficiency of plant led to the specification of a cheaper reduced capacity steam generator. The Operational Expenditures (OPEX) include the annual operating and maintenance costs, consisting of labor, water treatment and equipment maintenance costs. Expenditure on the purchase of biomass is calculated considering the annual fuel consumption of each concept and the local biomass cost (45.19 US$ t -1 ). Regarding the avoided costs, they consist of the amount that would be paid for the electricity and for producing steam if base case system was used, while the revenue is related to the sale of surplus electricity to the grid. In turn, the annual expenses are calculated considering the cost of fuel, OPEX and costs related to electricity demand contract and electricity consumption. The demand contract is necessary to guarantee the supply of electricity in case the cogeneration plant is offline (See Appendix B). Moreover, the electricity consumption refers to the value consumed during the shutdown of the cogeneration plant for maintenance.
Regarding the economic analysis, the SG1/ST1 configuration presented negative NPV, that is, during the useful life of the plant the return is lower than the investment. On the other hand, the SG1/ST2, SG2/ST1 and SG2/ST2 configurations had positive NPV; therefore, the initial investment was paid and yielded 4.49, 1.72, and US$ 10.93 million, respectively.
The IRR and the discounted payback DP are shown in Figure 5. When the NPV is positive, the IRR is higher than the minimum acceptable rate of return and the discounted payback is less than the plant lifetime. In contrast, for a negative NPV, the IRR is lower than the minimum acceptable rate of return and the discounted payback is longer than the plant lifetime. The SG2/ST2 configuration presented positive NPV of US$ 10.93 million, IRR of 20.77% (which is higher than the 10% minimum acceptable rate of return) and discounted payback of six years. The LCOE versus plant thermal efficiency for the proposed concepts are presented in Figure 6. As it can be seen, the SG2/ST2 plant presented the best operating conditions, both for the efficiency of 37.2% and for the LCOE of 73 US$ MWh -1 . In all cases, nevertheless, the obtained LCOE results were higher than the 58 US$/MWh paid for surplus electricity. In fact, the feasibility of plants was determined by the avoided costs related to electricity purchasing and steam generation under the base case layout. In summary, it was possible to identify that the most feasible layout was based on the use of higher steam parameters and reaction turbine. This is in accordance with the trend observed in the Brazilian market of biomass power plants, where improved steam parameters are being implemented today -see Figure 7. Furthermore, regarding the influence of superheated steam parameters and turbine expansion type on the economic figures, it was possible to identify, for the scenarios here studied, that upgrading the plant configuration to a reaction turbine (SG1/ST1 to SG1/ST2) tented to be more profitable than upgrading steam parameters (SG1/ST1 to SG2/ST1). This is an important result that might be considered during the conceptual engineering phase, when designers must choose the most feasible solution. Finally, the industrial costs with thermal energy [US$ ton -1 ] and electricity [US$ ton -1 ] for the base case and cogeneration equipped plants are presented in Figure 8. In case of SG2/ST2 scenario, the industrial cost with energy is reduced to 19 US$ ton -1 , i.e. 30% lower when compared to the base case, indicating potential economic feasibility of the proposed concept.

Conclusion
The results obtained in this work indicate the technical and economic feasibility of cogeneration plants to meet the thermal and electrical energy demands of poultry slaughterhouses, which represent an important sector of the Brazilian agribusiness. In general, the improvement of superheated steam parameters and the use of reaction steam turbines lead to a higher thermal efficiency of the plants, reduced energy forest plantation area for the production of woodchips, and greater economic attractiveness. The results The hourly electricity and process steam demands for one selected day are presented in Figures A1 and A2, respectively. As it can be seen, the demand profiles are nearly stable along the day.