A survey of technical literature [4], [5], [11], [20], [24], [25] indicated straw addition rate (A), straw moisture (µ), and pressure of boiler-generated superheated steam (p) as the most influential parameters in the thermodynamic performance of cogeneration systems operating in sugar and alcohol distilleries. The variation range p was defined based on technical (as a domain of cogeneration technology at high pressures) and economic-strategic (future perspectives for the bioelectricity market) criteria, specific for the São Paulo sugar and alcohol sector. Reports by Guerra et al. [24] and Moore et al. [29] were used for this purpose.

The upper limit of A was set at 50%_w/w of the straw generated in the field because of two factors: additions over this level might compromise the physical integrity of the boiler due to fouling and corrosion, and avoid damage to the soil. According to Kiatkitipong et al. [14], the fouling arises from the silicates existing in the straw, whereas corrosion may be caused from chlorides that are also part of its composition. Specialists in sugarcane cultivation suggest that at least half of the sugar straw should remain in the field to protect the arable land from actions such as erosion, wear and compaction, which can affect crop yield [30]. The lower end of the range (A = 10%) refers to the minimum amount of biomass capable of being transported without economic losses along the average displacement distance between the crop and the industrial plant in the state of São Paulo (32 km) [25].

The application of µ = 50% was an attempt to study the thermodynamic behavior of the Rankine cycle in situations in which straw moisture equaled the standard value of the bagasse parameter (µ_b = 50%). The other situations, at μ < 50%, were tested with the aim of verifying energy and environmental effects where straw was dried prior to use. In these cases, water evaporation from the biomass was assumed as occurring naturally without any auxiliary consumption energy source.

Table 1 exhibits the values (and ranges) for each parameter selected for the study. The analysis scenarios were established by the combination of the system operating conditions. The independent character of the variables led to 125 possibilities. However, a prior analysis of the energy-environmental performance of that collection regrouped the alternatives, giving rise to the twelve scenarios indicated in Table 2.

In this new conception, each scenario considers maximum and/or minimum values of two of the analyzed parameters, at the same time as the third variable was explored to the full extent of its variation levels. This approach allowed for the more precise identification of the effect of each parameter on the final result, since the others parameters are always considered in limit situations.

Regarding the thermodynamic behavior it was admitted, for any of the scenarios under analysis, that the cogeneration unit operates according to Rankine cycle with reheating, producing superheated steam, which expands in an extraction-condensing turbine. The system consumes all the bagasse obtained from sugarcane milling activities, in addition to straw, which is baled in the field and taken to the plant.

A Rankine cycle consists of main equipment (boiler, turbine, condenser, and electric generation) and auxiliary equipment (deaerator, pumps, desuperheater, pipes and steam trap). Modeling of this equipment was assisted by the Engineering Equation Solver (EES)^® software, which takes into account the principles of mass and energy conservation and is based on the 1^st and 2^nd Laws of Thermodynamics and provides solutions for an extensive set of linear equations with a high degree of precision.

The mass and energy balances for a steady-state system are expressed by eq. (1) and eq. (2), respectively:

${\displaystyle \sum _i{\dot{m}}_i={\displaystyle \sum _e{\dot{m}}_e}}$ (1)

${\displaystyle \sum _i{E}_i+\dot{Q}={\displaystyle \sum _e{\dot{E}}_e+\dot{W}}}$ (2)

where variables ${\dot{m}}_i$ and ${\dot{m}}_e$ are the mass flow rates of the working fluid entering and leaving the system, ${\dot{E}}_i$ and ${\dot{E}}_e$ are the energy rates that also cross the boundaries from control volume, $\dot{Q}$ expresses the heat rate provided to the system, and $\dot{W}$ refers to the work transferring rate.

The entropy balance concerning a steady-state system is indicated in eq. (3):

${\displaystyle \sum _j\frac{\dot{Q}}{T_j}+{\displaystyle \sum _i{\dot{m}}_i{s}_i+{\dot{S}}_{ger}={\displaystyle \sum _e{\dot{m}}_e{s}_e}}}$ (3)

where ${\dot{S}}_{ger}$ is the entropy generation rate and T_j is the temperature of the control volume, S_i and S_e are the specific entropy rates of input and output flows circulating throughout the boundaries of the system.

Energy balances corresponding to the auxiliary equipment did not consider the energy losses originated in these units, since previous analyses on Rankine cycles with reheating and straw reuse indicate that these parameters remained below 1.2% [24].

Consumption and emission estimates have been performed for anhydrous ethanol production (99.5%_w/w) in autonomous distilleries with milling capacities of 2.5 Mt sugarcane per crop [33]. Table 3 depicts others technical indexes for that process, which were adopted for designing the models applied herein

Energy performance is depicted by the eq. (4):

$I_{e,i}={\left(\frac{E_{exp}}{RB}\right)}_i$ (4)

where E_exp is the total of electricity exported, and RB is the intrinsic energy of the biomass (both bagasse and straw) consumed in the cogeneration. The specific amount of RB is calculated for each scenario from eq. (5):

$R{B}_i=\left(m_b\times PC{I}_b\right)+\left(m_{s,i}\times A_i\times PC{I}_{s,i}\right)$ (5)

where m_b concerns the mass of bagasse delivered into the boiler, PCI_b represents the LHV for this biomass for a moisture content µ_b = 50% (Table 2), m_s,i is the amount of straw corresponding to the total sugarcane to be converted into anhydrous ethanol in the distillery, and PCI_s,i is the LHV of straw, whose values were calculated for each scenario by eq. (6) [34]:

$PC{I}_{s,i}\left[J{\text{g}}^{\text{-1}}\right]=\left\{\frac{PC{S}_s-\left[600\times \left(\raisebox{1ex}{${\mu }_i$}\!\left/ \!\raisebox{-1ex}{$100$}\right.+9\times \left(\raisebox{1ex}{$H$}\!\left/ \!\raisebox{-1ex}{$100$}\right.\right)\right)\right]}{\left[1+\left(\raisebox{1ex}{${\mu }_i$}\!\left/ \!\raisebox{-1ex}{$100$}\right.\right)\right]}\right\}\times 4.18$ (6)

where PCS_i is the high heat value of straw and makes up 4,060 cal g⁻¹ in the circumstances to be verified in the study [5], ${\mu }_i$ [%] is the moisture of the straw for each scenario, and H [%] is the hydrogen content in the elemental composition of the straw to be burned. For the analyzed situation, [34] suggests assuming H = 6.4% PCI_s,i, and m_s,i values for the range of straw moisture variation applied in the study are given in Table 4.

This study follows the guidelines and requirements described in ISO 14040 [35] and 14044 standards [36]. The product system encompasses the stages described in Figure 1. The agricultural model considers the application of chemical fertilizers – ammonia, urea, ammonium nitrate, Monoammonium Phosphate (MAP), Single Superphosphate (SSP) and potassium chloride, to supply macronutrient needs (N, P and K). The use of limestone to correct soil acidity and agrochemicals to control plagues and crop diseases is also carried out in this same system. Moreover, the dosage of industrial by-products (vinasse, filter cake, and ashes) complements sugarcane nutritional requirements. The use of machinery is also considered, with consequent diesel consumption, in soil preparation, sowing, treatment and harvesting. The sugarcane harvest is fully mechanized, eliminating any action of biomass burning in the field.

Figure 1.

Scheme of anhydrous ethanol production linked to bioelectricity cogeneration (superheated steam production at 67 bar)

A productivity of 83.3 L C₂H₆O/t_c is obtained by fermenting sugar juice at controlled temperature (26-32 °C) and acidity (4.5 < pH < 5.5). The ethanol generated by this process is distilled, rectified, and dehydrated to reach 99.5%_v/v [33].

The environmental analysis was carried out by applying an attributional LCA with a ‘cradle-to-gate’ approach for a Reference Flow (RF) of 10 t anhydrous ethanol (99.5%_v/v). The setting of this parameter renders sugarcane consumption and bagasse generation for cogeneration in the distillery invariable, thus eliminating any interference from this biomass source on the results.

Primary data are used to represent industrial processes, whereas agricultural activities were modeled from secondary data. Data collection considered the technological pattern practiced in the state of São Paulo between 2008 and 2016 [12], [15], [24], [25].

All the detected multifunctionality situations were treated by the allocation procedure. The first, which occurs in the agricultural stage between sugarcane and straw, was addressed by mass criterion. This methodological decision conditioned the energy and environmental load estimates for each one of these products to the figures set in the scenario for A_i and µ_i of straw (Table 1).

The second situation, also addressed by mass criterion, takes place in the distillery, for ethanol (4.9%), returning condensate (26%), vinasse (67%) and cake filter (1.9%). The last of the multifunctionality approaches concerns cogeneration among the exported electricity and the electricity consumed for ethanol synthesis, low-pressure steam and ashes. In this case, an allocation based on the energy criterion was performed. Hence, ashes do not accumulate environmental loads, whereas the allocation factors for the other coproducts were determined based on electricity rates exported and used by the process.

In contrast to what occurred with the energetic dimension, the environmental performances for the assorted operational conditions of the system were assessed directly, by the application of ReCiPe, midpoint (H), v 1.12 method [37]. Those analyses were restricted to the CC impact category. This decision was made because of energy planning reasons. In recent years, the profile of the Brazilian electrical matrix (grid BR) has been changing, with hydropower being replaced by thermoelectricity from non-renewable sources such as natural gas and coal [38]. In this context, the generation of ‘low carbon’ electricity ‒ as in principle must occur with biomass ‒ represents a strategic differential, especially in view of the commitments assumed by the Brazilian government during COP 21, to reduce GHG emissions by 43% by 2030, compared to 2005 [29]. Thus, holding low impacts of CC is a fundamental condition to consolidate the participation of bioelectricity in the grid BR. Under this perspective, the other environmental impacts associated with it, mainly due to the cultivation of sugar cane, become essential elements for the next level of the decision-making process where this issue will converge.

The LCA’s were assisted by SimaPro^®, a software commonly used in analyzes of this nature. Secondary data that supported the modeling of agricultural stage were collected from the Ecoinvent Database^®, v 3.2 [39]. However, even in this situation, the life cycle inventory ‘sugarcane, at farm/BR U’ was remodeled to reflect one of the most important conditions of the study, which, as mentioned in Section ‘Life cycle modelling’, consider exclusively mechanized harvesting of sugarcane.

Figures 2-4 describe the effects of p, A and µ variations on the energy performance of the system, expressed in terms of (I_e,i), for the analysis conditions. In the S1 → S4 scenarios where I_e,i = f(ln p) for the sake of scale, it has been generally noted that increases in superheated steam output pressure improve system energy performance, despite the moisture content and straw addition rates considered (Figure 2).

As might be assumed beforehand, scenarios considering minimum moisture rates (S1 and S2) projected I_e,i values higher than those in which the amount of water in the biomass was at maximum (S3 and S4). It is also not surprising that S2 accumulates the best energetic performances of the whole series, ranging from 0.111 < I_e,S2 < 0.193 as the pressure increases from 20 → 100 bar. This effect can be explained because the scenario simultaneously evaluates together μ_min and A_max values. It should be noted, however, that as the operating pressure of the boiler rises the advantage of S2 over S1 tends to decrease, remaining at around 6.0% from 80 bar. This suggests that the gain by increasing the straw addition rate becomes secondary in cases where the Rankine cycle is subjected to extreme pressures.

Figure 2.

Effect of pressure variation (Δp) on the energy performance of the system (I_e,i)

Another way of identifying this same effect is at the pressure level to be reached by a scenario in order to compare I_e to the next value. In situations where straw is at maximum moisture, I_e,S3 at 20 bar will be equated with S4 only when the boiler operates at about 30 bar.

If the same analysis were to be performed between S3 and S1, with the latter operating once again at 20 bar, the dichotomies imposed by μ and A in each situation would lead to superheated steam leaving the boiler in S3 at 26 bar. On the other hand, it has been noted that the I_e,S2 measured at 20 bar is equal to I_e,S1 if p_S1 = 28 bar at the boiler outlet. This indicates that, at lower pressure levels, S1 distances itself from S2 enough to reverse the p-reduction trends established previously.

When carrying out a similar check for the upper p limits, however, another trend is observed. S3 should operate at 53 bar so that its energy performance equals that achieved by S4 at 100 bar ($I_{e,S4}^{100bar}=I_{e,S3}^{53bar}$). Following the same trend, $I_{e,S3}^{100bar}=I_{e,S1}^{73bar}$ and $I_{e,S1}^{100bar}=I_{e,S2}^{82bar}$. Therefore, at this extreme of the p-scale the better energy performance scenarios must be subjected to increasing pressures so that their I_e are similar to those of their immediate predecessors operating at 100 bar, as the difference between the pressures from consecutive scenarios is always shrinking.

Figure 3 associates I_e,i = g(A_i) obtaining, as in previous cases, linear relations for all evaluated scenarios. S5 and S6 exhibited gains in terms of energy performance with increasing straw addition, motivated by the fact that straw moisture contents in these cases were minimal. S7 and S8, on the other hand, displayed inverse trends, since biomass in those cases presented μ_max. If I_e,S6 always exceeded its counterparts in any of the analysis situations, the effects of A_i on such a performance were discrete. This is justified by the fact that the trend line that guides the relationship between I_e,i and A_i presented the lowest angular coefficient [tg(φ_S6) = 0.0003] within all representations of this set.

A comparison between this value and its corresponding for S5 [tg(φ_S5) = 0.0004] corroborates the previous conclusion that straw addition does not significantly affect the energy performance of the system, to the extent that this effect becomes secondary at pressures higher than 80 bar.

Figure 3.

Effect of straw addition variation (ΔA) on the energy performance of the system (I_e,i)

In scenarios were A_i had a negative effect on I_e, the alternative of using high vapor pressures resulted in gains, some significant, in energy performance with I_e,S8 > I_e,S7. In fact, as I_e,S8 was always higher than I_e,S5 for any A_i value, it can be concluded that pressure oscillations influence the energy efficiency of the system more than the moisture introduced by the straw. According to Figure 4, the relationship I_e,i = h(μ_i) also presents linear profiles for all evaluated scenarios. However, it is worth noting that the ordering of energy performance-based scenarios is conditioned by μ_i, to the point that two preference inversions are observed as moisture evolves from minimum to maximum values.

Figure 4.

Effect of straw moisture variation (Δµ) on the energy performance of the system (I_e,i)

For μ = 10%, it became apparent that I_e,S12 > I_e,S10 > I_e,S11 > I_e,S9, a condition which only confirms earlier findings that, in such circumstances, Rankine cycle operating pressures are more influential to cogeneration energy performance than straw addition. The review of these sequencings for an arrangement where I_e,S10 > I_e,S12 > I_e,S9 > I_e,S11 is also expected when μ = 50%, as the high moisture content, albeit at this extreme of the scale, favors smaller straw dosages, and pressure remains a decisive parameter in energy performance composition.

By verifying the equations describing the trend lines for each scenario, it was possible to notice that a position inversion between S12 and S10 occurred at a level where μ = 21.4%, whereas the substitution of S11 by S9 was confirmed at μ = 30.1%. With regard to S12 and S10, the results indicate that, in systems where steam leaves the boiler at 100 bar, it is favorable in terms of energy performance to carry out straw additions of the order of 50% if this biomass comprises regulated moisture content lower than 21.4%. From this limit, reducing the rate of straw administration to 10% of the total volume generated in the field becomes a more conservative option. A similar reflection allows explaining the inversion between S11 and S9 for moisture content of 30.1% in the case where the cogeneration plant operates at 20 bar.

Figures 5-7 depict the environmental consequences of varying p, A and µ with regard to CC. An overview of the obtained results pointed out that the main contributions to this dimension come from losses of fossil carbon dioxide (CO_2,f) from diesel combustion by agricultural machines during the soil preparation and treatment and sugarcane harvest, and in transport activities throughout the production chain. The set of precursors is completed by air Dinitrogen monoxide (N₂O) emissions from land-use changes and Methane (CH₄) releases also from incomplete diesel burning in agricultural machinery.

Emissions from the decomposition of straw left in the field, mainly as biogenic methane (CH_4,b), did not represent a significant portion of the total impact for CC. On the other hand, releases of biogenic carbon dioxide (CO_2,b) that are derived from biomass burning in the boiler were disregarded. This is because the impact accounting model applied by ReCiPe follows strictly the guidelines established by the Intergovernmental Panel on Climate Change [40] of disregarding, the portions relating to air CO₂ fixation (CO_2,fix) and the emissions of CO_2,b.

The results from Figure 5 indicate that S2 repeated the performance achieved in the energy dimension by registering the most expressive results from the entire series also in environmental terms (688 kg CO_2eq/MWh).

Regardless, S2 performance was always higher than its counterparts in all the analyzed pressures, as CC impacts are evaluated in specific terms (per MWh of exported electricity) and thus, the total electric energy generated when μ_S2 = 10% and A_S2 = 50% is able to cushion all GHG emissions that occur in the system, even if it operates at 20 bar, when energy production reaches minimum values. In addition, as the boiler operating pressure (p) rises, differences between the impacts generated by the scenarios for the same p_i value are reduced to the lowest values at 100 bar.

Figure 5.

Effect of pressure variation (Δp) on the environmental performance of the system

Such behavior confirms the expectation that the amount of energy contributed by the system at high pressures is even more significant than the GHG emissions that will be added due to increasing straw additions. Other evidence of this phenomenon lies in the pseudo-overlap between S1 (μ_S1 = 10%, A_S1 = 10%) and S4 (μ_S4 = 50%, A_S4 = 50%) performances for all evaluated pressures.

The worst performance achieved by the system was in S3. In this case, which accommodates low levels of electricity exports due to unfavorable relationships between straw moisture rates and biomass addition (μ_S3 = 50%, A_S3 = 10%), the successive increases in the Rankine cycle operating pressure did not go beyond reducing their differences in the other scenarios.

Figure 6 depicts an important aspect of system environmental performance. Although S5 achieves sufficient results only in terms of energy performance, the CC profile of that scenario appears as the best of the series that verifies the consequences of adding straw to the environmental bias. This divergence can be explained by the sharp interference that μ and A variations exercise on GHG emissions. The successive use of straw in the boiler raises the emission rates of CO_2,f, N₂O and CH₄ associated with this scenario, which originates from the agricultural stage. It should be noted that, in the logic practiced by the LCA, the straw left in the field is not characterized as a product of that specific processing and is therefore free of environmental loads.

Figure 6.

Effect of straw addition variation (ΔA) on the environmental performance of the system

Following a similar line of approach, the presence of increasing amounts of water in the straw is reverted to excess transport load, thus intensifying the CO₂ emissions resulting from this operation. In this context, although the total electricity generated in S5 is low (μ_S5 = 10%, p_S5 = 20 bar), GHG emissions are moderate enough to be attributed the lowest specific impacts of the set. S7 also exhibits low energy performance, and ranks second among the options, for operating at 20 bar but with μ_S7 = 50%. S6 and S8 recorded the worst rates of absolute impact for CC. On the other hand, by generating more expressive amounts of exported electricity (p_S8 = p_S6 = 100 bar) these effects are attenuated. Successive straw additions into the boiler lead to the S5 and S7 distancing of environmental performances. The apex of this discrepancy occurs at A = 50%. When admitting μ = 10%, S5 shows a downward trend profile due to the increase in straw addition [tg(φ_S5) = −6.15]. As S7 considers μ = 50%, its evolution manifests itself in the form of a more restrained trajectory [tg(φ_S7) = −2.27]. The same phenomenon can be observed between S8 and S6, where [tg(φ_S8) = −10.9] and [tg(φ_S6) = −3.58].

Finally, Figure 7 highlights the effects of moisture variation on the environmental profile of the electric cogeneration process from sugarcane biomass. By achieving the best specific CC performance indices, S12 corroborates previous findings. When operating with at A_S12 = 10% and p_S5 = 100 bar, this option achieves energy production levels high enough to absorb CC impact increases motivated by the same process conditions. The high operating pressures and high intake straw rates were shown to be more effective in terms of environmental performance than their procedural variants. In addition, unlike in the case of energy performance, no trend reversals associated with increased straw moisture were observed.

On the other hand, S12 and S11 are more sensitive to μ increments than their counterparts S10 and S9, from two circumstances: as straw addition rates increase, the amount of water introduced into the system with the biomass tends to reduce its potential for energy generation, as observed between S12 and S11, and if the Rankine cycle operates at low pressures, such as in S9 and S11, a situation in which production capacity is more restricted.

Figure 7.

Effect of straw moisture variation (Δµ) on the environmental performance of the system