Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Investigation into the combustion kinetics and spontaneous ignition of sweet sorghum as energy resource

Investigation into the combustion kinetics and spontaneous ignition of sweet sorghum as energy... Thermogravimetric analysis (TGA) has received Introduction immense attention for studying the combustion kinetics The priority of utilizing biomass has attracted much of biomass. Many studies have used TGA to profile its global attention to address environmental problems, characteristic temperatures (i.e., ignition temperature, such as global warming, due to the excessive use of fos- burnout temperature, and peak temperature(s)), kinetic sil fuels. Those concerns are reflected in IRENA’s 1.5  °C parameters (i.e., activation energy and pre-exponential scenarios on climate change, where bioenergy represents factor), and reactivity parameters (i.e., ignition index 25% of the total estimated primary supply or equals 153 and combustion index) in an oxidative environment. EJ by 2050 (IRENA 2021). Hence, the exploration of vari- The traced thermal degradation reveals the response of ous potential biomass for the power generation industry combustion reaction to various operating conditions, will be intensively carried out to meet the needs. such as biomass composition, heating rate, and others. For years, studies of combustion kinetic and sponta- The experimental study using olive residues was done by neous ignition have been proposed to understand the Magalhaes et  al. (2017). The characteristic temperatures essential characteristics of biomass during utilization were graphically identified from the TG and DTG curves and storage before being applied in large-scale power at 240–350, 340–700, 240–245, and 531–703  °C for industries. Direct combustion has become an efficient two combustion reaction stages, ignition, and burnout, technology for converting biomass into energy and has respectively. The apparent activation energies, which are been widely used in power plants with steam turbine sys- 44.2–47.7 and 9.1–13.4 kJ/mol, were presented by assum- tems. Thermal behavior is crucial information to explain ing reaction mechanisms and calculating them using the the ignition and burnout of fire which influences boiler model-fitting method of Coats–Redfern. operation, energy efficiency, and emissions (Cao et  al. Meanwhile, spontaneous ignition can be predicted in a 2017). Therefore, understanding combustion kinetic can biomass stockpile by considering its combustion kinetic help better knowledge of design and optimization in bio- parameters and applying the mathematical theory of mass combustion systems. Meanwhile, spontaneous igni- Frank-Kamenetskii. The theory expresses that the igni - tion may arise from the exothermic process due to the tion arises based on solid heat conduction in specific susceptibility of stockpiled materials to self-heating. The geometric shapes. Boonmee and Pongsamana (2017) physical, biological, and chemical heating mechanisms performed a laboratory-scale experiment using bagasse. are addressed to contribute to self-heating events (Sheng A cylindrical- or rectangular-shaped stockpile with any and Yao 2022). There were many reports of fire incidents radius or length stored with asymptotic heights below 10, during storage. For instance, a fire started from a large 7.8, and 6 m is considered safe from fire hazards at ambi - biomass pile of 500 tons at Advanced Agro-Power Plant ent temperatures of 40, 45, and 50 °C. in Thailand on 12 March 2017 due to accumulated heat. The current study focused on a comprehensive investi - For that reason, the knowledge of spontaneous ignition gation of combustion kinetics and spontaneous ignition can give a fundamental evaluation and mitigation of fire using sweet sorghum. Sorghum was chosen to represent hazards in biomass storage. L uthfi et al. Bioresources and Bioprocessing (2022) 9:49 Page 3 of 12 the primary marginal crop, which has many superior experiment was conducted 2 times under the same con- characteristics compared to other crops, such as grow- ditions to verify the reproducibility. A small sample mass ing in a short period of about 4–6 months and being able and low heating rate were applied to avoid the transport to withstand an environment with little water and high effect and enhance the resulting signal of slower kinetic soil salinity (Xie et al. 2018). TGA was performed at dif- events. ferent heating rates and in an oxidative environment to assess its response to reaction stages, characteristic tem- Combustion kinetic analysis peratures, kinetic parameters, and reactivity parameters. Determination of reaction stages Subsequently, the spontaneous ignition was predicted The combustion reaction region (exothermic reaction) by substituting the existing kinetic parameters into the was qualitatively determined by observing the gain por- steady-state solution of the Frank-Kamenetskii equation. tion of the DTA curve (see Fig.  1a). Then, by following The prediction was presented regarding the relationship the method of Pickard et  al. (2013) or tangent-bisection between ambient temperature and safe silo size. method (TBM), the reaction stages of combustion were identified on a DTG curve with the assumption that they Materials and methods are non-competing, first-order, and single-stage reac - Characteristics and preparation of sweet sorghum samples tions obeying the Arrhenius law (see Fig.  1b). Tangent Sweet sorghum was collected locally from farm areas lines were drawn to the edges of the leading and trailing around Mie University, Japan. Sorghum was appropri- curves. Bisection lines were extended at each tangent ately washed to eliminate dirt and was air-dried (± 20 °C) intersection until they reached the DTG trace and labeled for several days to reduce moisture content and inhibit as the start or end of the reaction stages. decay. Dried sorghum samples were ground using a high- speed blender, YKB (AS ONE Corp.), with a rotation speed of approximately 28,000 rpm for 1 min. All samples Determination of characteristic temperatures were sieved to a size range of 250–500 μm. According to The characteristic temperatures were determined using a Wilen et al. (1996), the proximate and ultimate character- TG and DTG curve, based on the report of Lu and Chen istics of sweet sorghum are summarized in Table 1. (2015) (see Fig.  2). Peak temperatures were observed at the maximum mass loss rate of each reaction stage. Then, TGA experimental approach the intersection method (IM) was employed to observe A thermal analyzer, EXSTAR 6000 TG/DTA 6200 (Seiko the ignition and burnout temperatures. Horizontal lines Instruments Inc.), was used to conduct the thermo- were drawn at points B and D, where the TG curve gravimetry (TG), derivative thermogravimetry (DTG), became steady after the evaporation and combustion and differential thermal analysis (DTA) experiments. reactions were complete. Afterward, tangent lines were Air was fed into the furnace from the pump and was drawn at points A and C through the horizontal lines of regulated to 100  ml/min for each experiment. Sam- points B and D, respectively. Points A and C are defined ples with a mass of 6  mg were used in each experiment as cross points at which vertical lines from the DTG and were heated from ambient temperature to 550 °C at curve peaks of stages I and II cross the TG curve. The three different heating rates of 2, 5, and 10  °C/min. The temperatures corresponding to the intersections of the Table 1 Proximate and ultimate characteristics of sweet sorghum ( Wilen et al. 1996) db Proximate analysis Value (%) Method Moisture content 7.04arb DIN 51,718 Ash content 4.74 DIN 51,719 Volatile matter 77.20 DIN 51,720 Fixed carbon 18.06 – db Ultimate analysis Value (%) Method Carbon 47.30 – Hydrogen 5.80 – Oxygen 41.67 By difference Nitrogen 0.40 – Sulphur 0.09 ASTM D 4239 Luthfi et al. Bioresources and Bioprocessing (2022) 9:49 Page 4 of 12 tangent and horizontal lines were labeled as the ignition and burnout temperatures, respectively. Determination of reactivity parameters In support of the reactivity data, the ignition and combustion indices were calculated for all samples to provide a more accurate measure of the conver- sion over time and temperature. According to Fraga et  al. (2020b), the ignition and combustion indices are defined in Eq. 1 and Eq. 2, respectively: (dm/dt) max D = , (1) t t max ig (dm/dt) (dm/dt) max avg S = , (2) T T ig where D and S are the ignition (in %/min ) and combus- 2  3 tion indices (in %/min °C ). (dm/dt) and (dm/dt) max avg are the maximum and average mass loss rates (in %/min), respectively. T and T are the ignition and burnout tem- ig b peratures (in °C). t and t are the times corresponding ig max to the ignition and maximum combustion rates (in min). Determination of kinetic parameters The Friedman method, which obeys the Arrhenius law, was used to estimate the kinetic data from a TG curve. Fig. 1 Schematic for determining a the combustion reaction region Since the amount of airflow is much greater than the and b combustion reaction stages (e.g., at a heating rate of 5 ºC/min; 1st iteration sample) sample mass used, the oxidation process is assumed not to depend on the oxygen concentration. Therefore, the combustion reaction was modeled as first-order kinetics. According to Yao et al. (2008) and Huang et al. (2016), the combustion of solid biomass is described by the following: dx = kf (x), (3) dt where dx/dt is the conversion rate (in 1/s), k is the reac- tion rate constant (in 1/s), and f(x) is the reaction func- tion. In addition, x is the degree of conversion. The expression for x is defined as: m − m i t x = , (4) m − m Fig. 2 Schematic for determining the peak, ignition, and burnout where m , m , and m are the initial mass of the sample, i t f temperatures (e.g., at a heating rate of 5 ºC/min; 1st iteration sample) mass of the sample at a specified time, and final mass of the sample (in kg), respectively. Furthermore, the expres- sion of k is defined by the Arrhenius equation: −E/RT (5) k = Ae , L uthfi et al. Bioresources and Bioprocessing (2022) 9:49 Page 5 of 12 where A is the pre-exponential factor (in 1/s), E is the Results and discussion activation energy (in J/mol), R is the universal gas con- Combustion reaction and characteristic temperatures stant (in J/molK), and T is the temperature of the sample of sweet sorghum (in K). Based on the graphical analysis of DTA, sweet sorghum Subsequently, substituting Eq. 5 into Eq. 3 gives the fol- was thermally degraded in two distinct regions of exo- lowing equations: thermic and endothermic reactions due to evapora- tion (region I) and combustion (region II). As shown dx −E/RT in Fig.  1a, the first gain in heat flow permits the identi - = Af (x)e . (6) dt fication of these two regions. The gain temperatures were identified at 226–223  °C for all heating rates. The The model-free method of Friedman is defined in Eq.  7 negative-valued region asserts that sorghum samples by taking a logarithm on both sides of Eq.  6 without absorbed heat to release surface and inherent moisture. assuming the reaction model and the reaction function. In contrast, the positive-valued region corresponds to The following equation is hence obtained: heat liberation due to lignocellulosic pyrolysis (i.e., dehy- dx E dration and decarboxylation of hemicellulose, cellulose ln = ln Af (x) − . (7) dt RT (> 400–450 °C), and lignin) and burning volatile and char (Rahib et  al. 2019; Basu 2013). The heating rate affected By plotting the left-hand side of Eq.  7 against 1/T at the onset of heat release by lowering the gain tempera- different heating rates followed by taking a linear regres - ture by 3 °C at higher heating rates and vice versa. It took sion, the activation energy (E) can be obtained from the place at approximately 226  °C for heating rates of 2  °C/ slope. Meanwhile, the intercept may be used to deter- min and 223 °C for rates of 5–10 °C/min. This is because mine the pre-exponential factor (A) if the reaction func- the heat given at a low rate provided enough time for tion is modeled beforehand (for first-order kinetics, it is water to diffuse or even break their bonds from vessels f(x) = (1-x)). and fibers completely (Penvern et  al. 2020). The force holding might derive from hydrogen bonds among the Spontaneous ignition analysis water in vessels or between the water molecules and The safe size of the stockpile and ambient temperature hydrophilic sites via functional groups, such as hydroxyls, under critical condition were evaluated using the dimen- phenolic, and carboxylic acids (Khare and Baruah 2014). sionless equation of Frank-Kamenetskii. The equation is The difference in the mass loss proves the presence of a expressed by Eq.  8, which represents the ratio of chemi- water concentration gradient in the sample. The smaller cal energy to thermal conduction (Boonmee and Pongsa- the mass loss is, the more water remains. TG curve shows mana 2017; Fisher and Goetz 1993): that the mass loss decreased from 82.51 to 86.80% at the transition point as the rate increased. �HEr AC E The combustion region, profoundly, was identified as δ = exp − , (8) RT RT c having multi-stage reactions, as depicted by the occur- rence of slopes and peaks on the TG and DTG curves where δ is the critical Damkohler number, and T is the c c (see Fig.  3). The tangent-bisection and intersection critical ambient temperature (in K). r is the characteristic methods were performed on the DTG curve to enhance dimension of a pile with origin at the center (in m), which confidence in quantifying the major reaction stages and is half of the total depth (r = 1/2r ) or equal to the total characteristic temperatures (see Fig.  1b and Fig.  2). For radius (r = r ). ΔH is the heat reaction (in J/kg), which is each temperature parameter obtained (i.e., reaction assumed to be equivalent to the calorific value (Q) mul - temperatures and characteristic temperatures) from all tiplied by the conversion degree (x) at the ignition tem- heating rates, extrapolations were subsequently done at perature (ΔH = x Q). C and λ are the bulk density (in kg/ ig 0 0 °C/min to exhibit the material nature and eliminate the m ) and thermal conductivity of the sample (in W/(mK)). linear influence of the heating procedure. On the other n is the kinetic reaction order. E and A are the activation hand, the remaining masses, showing the material nature, energy (in J/mol) and pre-exponential factor of reac- were determined by simply averaging the three corre- tion (in 1/s), respectively, which is obtained from kinetic sponding heating rates due to intrinsic property (Yao analysis at the ignition temperature. R is the universal gas et al. 2008). The obtained data are summarized in Table  2 constant (in J/(molK)). and Table 3. The robust methods divided the combustion reaction into two stages, which occurred at 131–336 and 336– 475  °C for stages I and II, respectively. The first stage is Luthfi et al. Bioresources and Bioprocessing (2022) 9:49 Page 6 of 12 while the chars remained (Sami et al. 2001). In a sufficient proportion, the mixed gases and/or some solids were predicted to give off flames at 215  °C due to oxidation, when cellulose became the dominant mechanism (Lu and Chen 2015). The pyrolysis reaction consumed 56.87% of the sorghum mass, with the reactivity rate increas- ing sharply at 264 °C (T ). Meanwhile, the second stage p,I corresponds to the char oxidation within the diffusion- controlled phase via mixed gas (Magalhaes et  al. 2017). In  situ oxidation of volatiles and chars might proceed in parallel due to overlapping. The transition from the burn - ing of volatile to char can be notified by the change of flame into an ember (Rahib et al. 2019). After the volatile matter in the atmosphere was exhausted, oxygen-rich air diffused further into the interior of the char and burned out the rest to ashes (Sami et  al. 2001). The ember was found to extinguish at 433 °C (T ). Overall, the oxidation reaction converted 24.44% of the sorghum mass with the highest reactivity at 405 °C (T ). p,II Following the preliminary identification, the appear - ance of the shoulder (less pronounced) and head peaks on the DTG curve at around 180–200 and 270–290  °C for all heating rates, respectively, indicates the pseudo- hemicellulose and pseudo-cellulose within the pyrolytic stage (see Fig. 3b). Both could also be marked at the first and second slopes of the TG curve, excluding the slope of water evaporation (see Fig.  3a). In contrast, the pseudo- lignin peak, which is believed to be indistinct due to its notoriously broad reaction, might fall around 320–420 °C according to the third TG slope. Meanwhile, the char oxi- dation reaction could be notified by the prominent DTG Fig. 3 a TG and b DTG signals with plots of the characteristic peak at 410–450 °C or the fourth TG slope. Similar loca- temperatures for all heating rates (1st iteration sample) tions of pseudo-lignocellulose were reported by Carvalho et  al. (2015) on sweet sorghum in an inert atmosphere using the deconvolution method. Accordingly, Jayaraman et al. (2017) confirmed that pyrolysis and char oxidation attributed to the pyrolysis and volatile oxidation con- occurred in similar ranges by investigating the evolu- trolled by chemical reactions (Magalhaes et  al. 2017). tion of typical gases from various biomass. A significant Slow pyrolysis (< 50 °C/min) produced volatiles and chars amount of water vapor (H O), carbon monoxide (CO), majorly through dehydration and decarboxylation of hydrogen (H ), and aromatic compounds (i.e., meth- hemicellulose (127–235  °C), cellulose (227–350  °C), and ane (CH ), benzene (C H ), and so on) was detected in 4 6 6 lignin (127–527  °C) (Basu 2013; Carvalho et  al. 2015), the range of 150–400  °C, showing the typical gaseous wherein the volatiles then diffused into the ambient air Table 2 Reaction temperatures and characteristic temperatures of the combustion reaction under different heating rates Heating rate (°C/min) Reaction temperature (°C) Characteristic temperature (°C) Stage I Stage II Tp,I Tp,II Tig Tb - (material nature) 131–336 336–475 264 405 215 433 2 143–348 348–475 268 412 218 444 5 168–362 362–486 281 435 229 467 10 197–391 391–487 292 452 236 495 L uthfi et al. Bioresources and Bioprocessing (2022) 9:49 Page 7 of 12 Table 3 Mass changes including reaction durations of the combustion reaction under different heating rates Heating rate (°C/min) Remaining mass, % (standard deviation) Mass loss, % (reaction duration, s) Maximum mass- loss rate (%/min) Stage I Stage II Stage I Stage II Tp,I Tp,II - (material nature) 91.01 (1.26)–34.14 (1.42) 34.14 (1.42)–9.70 (0.22) 56.87 24.44 – – 2 91.95–32.55 32.55–9.80 59.40 (6057) 22.75 (3846) 1.42 0.66 5 91.50–35.28 35.28–9.45 56.22 (2265) 25.83 (1506) 3.38 1.86 10 89.59–34.59 34.59–9.84 55.00 (1110) 24.74 (585) 6.94 17.69 products of dehydration, decarboxylation, and scission from 3846 to 585  s during stage II. This implies that the in pyrolysis reactions. In contrast, carbon dioxide (C O ) combustion reaction proceeded faster at higher heating was released more in char oxidation between 400–750 °C rates without any obvious change in the amount of mass along with a small release of CO, representing the partial converted. oxidation in char. The maximum mass-loss rates at the corresponding Furthermore, increasing the heating rate affected the stages rose from 1.42 to 6.94%/min and from 0.66 to TG and DTG curves to shift to higher temperatures, 17.69%/min as the heating rate increased, agreeing with marked by a significant delay of 10–50  °C for both tem - the mass loss to reaction duration ratio (see Table  3). perature parameters as shown in Fig.  3 and Table  2. The values define the rate at which the non-condensed The significant linear influence is a consequence of the and condensed phases of biomass decompose to gases. decrease in the time required for heat transfer to cross As can be observed, stage II, which was dominated by the biomass interior before the surface temperature char, had lower mass-loss rates than stage I due to the increases (Elorf et  al. 2021). A major deterrent to the high energy bond of the carbon–carbon bond. Cor- higher heating rate is the low thermal conductivity of sor- respondingly, the susceptibility of sorghum charac- ghum by 0.13 W/mK (Fennell and Boldor 2014). For this terized by a high fuel ratio (> 2.0) of volatile matter to reason, even though the heating rate was high, the sor- fixed carbon (4.27) exerted influences on the results of ghum interior was heated at a considerably slower. the early stage (Lu et  al. 2013; Wilen et  al. 1996), such as low ignition temperature (200–300  °C) and higher Combustion reactivity of sweet sorghum maximum mass-loss rate during devolatilization (see Changes in heating rate exhibit unique effects on the Table  1) (Basu 2013). Comparing the rate magnitude mass changes and reaction durations. Based on Fig. 3 and of the two, stage II rose approximately ten times higher Table  3, the mass losses fluctuated during combustion under 5–10  °C/min. In contrast, stage I increased with a tendency to decrease insignificantly (1–4%) as the merely twofold at the same level of heating rate. It may heating rate increased. Stage I experienced a decrease in be caused by the ability of oxygen to overcome mass the mass loss by 59.40–55.00% at a heating rate of 2 to transfer resistance and diffuse at high heating rates 10 °C/min. Meanwhile, the mass loss at stage II increased (Islam et  al. 2016). Unlike the devolatilization process from 2 to 5 °C/min by 22.75–25.83% and decreased from which is only sensitive to the particle temperature, the 5 to 10  °C/min by 25.83–24.74%. Fluctuating results reactivity of char oxidation can accelerate due to tem- were also obtained in the report of Fraga et  al. (2020a) perature elevation and oxidizer concentration (Li et  al. and Jayaraman et  al. (2017) using various biomasses 2016), which may affect the present findings in the sim - and broader ranges of heating rates. Fraga et  al. (2020a) ilar way. deduced this to be a form of randomness. Additionally, Supporting the given trends in the maximum mass- the insignificance of the mass changes was expressed in loss rate, the change in heating rate shows a directly pro- small standard deviations (SD) of the remaining masses, portional relationship with the ignition and combustion namely 1.24 (1.38%; relative standard deviation (RSD)), indices. Both indices signify the decomposition over time 1.42 (4.16%), and 0.22% (2.27%) for the onset of stage I, and temperature within combustion stages. The ignition transition point, and offset of stage II, respectively. This and combustion indices were obtained to increase from 4 4 3 8 implies that the mass change was probably an intrinsic 1.21 × 10 to 192.17 × 10 %/min and from 3.37 × 10 to 8 2 3 property and not affected by the heating rate (Yao et  al. 180.14 × 10 %/min °C , respectively, under 2–10  °C/min 2008). Conversely, the increase in heating rate alone heating rates (see Fig. 4). Substantially, the ignition index significantly affected the reaction durations, denoted suggests that sorghum caught fire faster with magni - by the decrease from 6057 to 1110  s during stage I and tudes of approximately 13–12 times higher as the heating Luthfi et al. Bioresources and Bioprocessing (2022) 9:49 Page 8 of 12 rate increased. This index is assigned by concerning the amount of volatile evolution within the first combustion stage or pyrolysis reaction. Meanwhile, the combustion index exhibits the activity of the substance consumption, starting when the sorghum caught fire until the chars burned out into ashes, signifying the predominance of oxidation. Following the same previous causal relation- ship, the entire combustion reaction proceeded rapidly by the increment of about 5- to 10-fold (Wnorowska et al. 2021; Fraga et al. 2020b). Kinetic behavior of sweet sorghum The Friedman method was employed to evaluate the kinetic data of the overall combustion reaction by linearly Fig. 4 Eec ff ts of different heating rates on the ignition and fitting the term ln(dx/dt) to 1/T in Eq.  7 for a series of combustion indices conversion degrees at different heating rates. The conver - sion degree in the range of 0.1–0.9 was used in this study with segmentation of 0.1. Figure 5 shows that fitted lines were nearly parallel in three conversion ranges of 0.1–0.5, 0.6–0.7, and 0.8–0.9, denoting a single reaction or uni- fication of several reactions (Yao et  al. 2008). Averaging over the corresponding parallel ranges, then, will give meaningful activation energy. To avoid overestimation, the relative standard deviation (RSD) was determined to be no more than 10% (see Table  4). The correlation coefficient values demonstrated the coherence of kinetic parameters generated by the Friedman method. Fraga et  al. 2020b stated that the first-order kinetic model can be considered to fit well if the coefficient is more than 0.90 (R > 0.90). Because the correlation coefficients of this study ranged from 0.9893 to 0.9999 for the whole conversion degree, these results indicate the adequacy of the kinetic parameters to represent good linear fitted plots and reaction mechanism (i.e., model and function). Fig. 5 Linear plots of ln(dx/dt) against 1/T for various conversion The distribution of activation energies and pre-expo - degree in Friedman method representing the entire combustion reaction nential factors in the function of conversion degree is Table 4 Activation energies, pre-exponential factors, and correlation coefficients (R ) for the entire combustion reaction using Friedman method Conversion (mg/ Friedman method Conversion (mg/ Friedman method mg) mg) E (kJ/mol) A (1/s) R E (kJ/mol) SD Relative SD avg 0.1 129.57 5.25 × 10 0.9998 0.1–0.5 132.91 8.60 6.47 0.2 143.38 7.30 × 10 0.9999 0.6–0.7 79.40 7.21 9.07 0.3 140.19 2.14 × 10 0.9992 0.8–0.9 169.44 2.45 1.45 0.4 132.49 2.21 × 10 0.9994 0.5 118.92 6.00 × 10 0.9961 0.6 72.20 1.26 × 10 0.9918 0.7 86.61 6.30 × 10 0.9994 0.8 166.99 7.97 × 10 0.9893 0.9 171.89 1.41 × 10 0.9993 L uthfi et al. Bioresources and Bioprocessing (2022) 9:49 Page 9 of 12 It is important to note that satisfying activation energy is required to represent the entire reaction. According to Yao et  al. (2008) and Luo et  al. (2016), fractions below 0.5–0.7, showing parallel plots at the beginning, generally provide meaningful information instead of the whole process, especially for the case of self-heating and combustion. The activation energy, which encompasses the ignition process, determines the external energy to heat the fuel reactor since the subsequent reaction will become self-sustaining above this temperature, related to exothermic chain reac- tions. Therefore, the apparent activation energy of the combustion reaction in this study was 132.91  kJ/mol regardless of the heating rate. Fig. 6 Apparent activation energy and pre-exponential factor as a function of conversion degree calculated by Friedman method Spontaneous ignition of stockpiled sweet sorghum The dimensionless parameter of Frank-Kamenetskii or critical Damkohler number (δ ), expressed by Eq.  8, was shown in Fig.  6. Converting the conversion degree of used to predict spontaneous ignition based on the energy 0.1–0.5 to temperature range from 218 to 317 °C, which equation at steady state. The thermal explosion model has the exact position in the applicable heating rates of was derived by assuming that the Biot number is equiva- 2, 5, and 10  °C/min, the unified reaction was allegedly lent to infinity (Bi = ∞). The number means that reactant suspected to be conjoint pyrolysis of hemicellulose and temperature is distributed along the body in a parabolic cellulose predominantly. The mean activation energy manner, governed by internal conduction. Thus, the was found to be 132.91  kJ/mol, while the pre-expo- critical Damkohler number depends specifically on the 9 7 −1 nential factor ranged from 5.25 × 10 to 6.00 × 10  s , geometry of reacting systems (Fisher and Goetz 1993). indicating the frequency of molecular collision for the Current work studied the relationship between criti- reaction to occur (see Table 4). Thus, the lower the acti - cal ambient temperature and silo dimension in sorghum vation energy is, the easier it is for a reaction to com- storage. Since feedstock is usually piled up in a non- mence. Considering the same way to elucidate the data, geometric shape, the stockpile was assumed to comply lignin-dominated pyrolysis underwent in the second with two common geometric shapes in the market: (1) a half of the conversion degree (0.6–0.7), settling in the cylinder with dimensions of diameter (d) and height (h); temperature from 304 to 395  °C. The required activa - and (2) a rectangular box with dimensions of height (h), tion energy decreased to 79.40 kJ/mol, presumably due length (l), and width (w). If the critical ambient tempera- to the least presence of hemicellulose and cellulose ture is set within a specific range, and the Damkohler remaining at the high temperature. At a conversion number for a given geometry is specified, the response degree of 0.8–0.9, equivalent to a temperature range of the selected silo dimension to the critical temperature from 386 to 453 °C, char oxidation was presumed to be can be calculated. According to Fisher and Goetz (1993), the reaction to occur. The minimum energy for oxida - Damkohler numbers for finite cylinder (heat loss one tion to react increased drastically to 169.44 kJ/mol as a end) and rectangular box are shown in Eq.  9 and Eq.  10, consequence of the gradual deposition of carbon con- respectively. stituents. Additionally, both possible reactions were disclosed to have pre-exponential factors spreading δ = 2.0 + 0.195(d/h) , (9) 3 3 9 from 1.26 × 10 to 6.30 × 10 and from 7.97 × 10 to 10 −1 1.41 × 10  s , respectively. Table 5 Calculation parameters used in Frank-Kamenetskii model Parameter Value Parameter Value Parameter Value 5 3 T (K) 273.15–473.15 ΔH (J/kg) 17.79 × 10 E (J/mol) 129.57 × 10 3 5 9 C (kg/m ) 38.86 Q (J/kg) 173.91 × 10 A (1/s) 5.25 × 10 λ ( W/mK) 0.13 xig 0.102 R (J/molK) 8.314 Luthfi et al. Bioresources and Bioprocessing (2022) 9:49 Page 10 of 12 Table 6 Geometric shapes and dimension ratios used in Frank-Kamenetskii model Geometric shape Dimension ratio Geometric shape Dimension ratio d/h Note h/l h/w Note Cylinder 1/10 (0.1) (h > d) Box 1/2 (0.5) 1/2 (0.5) (h < l,w) (l > w) 1/3.3 (0.3) 1/4 (0.25) 1/2 (0.5) 1/2 (0.5) 1/8 (0.125) 1/2 (0.5) 1/1.4 (0.7) 1/4 (0.25) 1/4 (0.25) 1/1.1 (0.9) 1/8 (0.125) 1/4 (0.25) 1/16 (0.0625) 1/4 (0.25) 2 2 δ = 0.825(1.067 + (h/l) + (h/w) ). (10) Regarding the calculations, the critical ambient temperature (T ) was set between 0–200  °C (273.15– 473.15  K) to represent the relative reference that changed in the input (see Table  5). A series of diam- eter and height were generated as the selected dimen- sions of cylindrical and box silos, which are equivalent to twofold the characteristic dimension (r). Meanwhile, other storage dimensions were converted into ratios such as d/h, h/l, and h/w, so that Eq. 9 and Eq. 10 could be calculated mathematically and substituted into Eq. 8 (see Table 6). Concerning the rest parameter in Table 5, the thermal conductivity (λ) was cited from the work of Fennel and Boldor (2014), while the bulk density (C ) and calorific value (Q ) were, respectively, experimented based on the method of Bhagwanrao and Singaravelu (2014) and the standard of adiabatic calorimeter (O.S.K 150, Ogawa Sampling Co., Japan). Upon extrapolating at 0 °C/min, the conversion degree of ignition (x ) was ig multiplied by calorific value to determine the heat reac - tion (ΔH). Meanwhile, the activation energy (E) and pre-exponential factor (A) were fitted to the linear plots in Fig.  5, which satisfied the conversion degree of igni - tion of 0.102. Observing Fig.  7, the resulting curves suggest that at certain levels of temperature and silo size, the areas under the curves are appropriate for storing sorghum piles safely. The opposite meaning applies to the area above the curve. For both geometric shapes of cylinder and box, the curves exhibit vertical (va) and horizon- Fig. 7 Relationship between the critical ambient temperature and tal asymptotes (ha) around the coordinates y = 8  km; va selected silo dimension for a cylindrical and b rectangular-boxed x = 60 °C and y = 6 km; x = 60 °C when the selected va ha ha shapes, including the enlarged curves to show the effect of different dimensions were increased or decreased to near infinity dimension ratios while keeping the respective ratios. The findings con - firm the report of Murasawa et al. (2013) on storage of soy sauce residue and fishmeal, in which similar asymp - 10 m for both geometric shapes to study the spontane- totic responses were detected. ous ignition of sorghum on a relevant scale (see Fig. 7a The market designs were employed to present and Fig. 7b). These values were chosen because they are enlarged curves around diameter and height of 15 and L uthfi et al. Bioresources and Bioprocessing (2022) 9:49 Page 11 of 12 Abbreviations commonly used in large-scale industries and are decent TGA : Thermogravimetric analysis; TG: Thermogravimetry; DTG: Derivative to apply to several dimension ratios in the current thermogravimetry; DTA: Differential thermal analysis; TBM: Tangent-bisection study. Based on the cylindrical silo design (d = 15  m), method; IM: Intersection method. the spontaneous ignition of sorghum will not occur at Acknowledgements ambient temperature below 83–84 °C, within the incre- Not applicable. ment of d/h ratio from 0.1 to 0.9. It is possible to notify Author contributions that the shift due to various d/h ratios was only 1  °C. The contributions of each author are as follows: NL performed the experi- Similar findings were found in the box-shaped silo ments, analyzed the data obtained, and wrote the manuscript. TO performed design (h = 10  m). The critical temperature showing the experiments and analyzed the data obtained. Y T administered the project and acquired funding. TF supervised the study. KT conceptualized and super- a safe limit for spontaneous ignition ranged between vised the study. All authors read and approved the final manuscript. 84–87  °C. The temperature shifted by 3  °C as the h/l and h/w ratios rose from 0.0625 to 0.5 and from 0.25 Funding This study was funded by Chubu Electric Power Co., Inc., Aichi, Japan. to 0.5, respectively. The combination of the two ratios is shown in Table  6. In addition to finding the critical Availability of data and materials temperatures for each design, although not significant, All data analyzed during this study are included in this article. it can be seen that the smaller the resized dimension to the ratio is, the greater the ambient temperature for Declarations a fire to ignite is required, owing to the immense heat Ethics approval and consent to participate dissipation. Not applicable. Consent for publication Conclusions Not applicable. The combustion kinetics and spontaneous ignition of sweet sorghum have been investigated successfully using Competing interests The authors declare that they have no competing interests. TGA and the Frank-Kamenetskii theory. The findings are highly related to sorghum utilization as a fuel in a com- Author details bustion reactor and to safety storage as a feedstock from Department of Environmental Science and Technology, Graduate School of Bioresources, Mie University, 1577 Kurimamachiyacho, Tsu, Mie 514-8507, fire hazards. Investigation shows that activation energy of Japan. Staff Service, 85 Kandaneribeicho, Chiyoda-ku, Tokyo 101-0022, Japan. 132.91  kJ/mol was required to undergo the combustion Department of Life Sciences, Graduate School of Bioresources, Mie University, in a reactor, fitted well by the first-order model. The com - 1577 Kurimamachiyacho, Tsu, Mie 514-8507, Japan. bustion of sweet sorghum occurred in the temperature Received: 5 December 2021 Accepted: 7 April 2022 range of 131–475  °C and comprised two different ther - mal stages, corresponding to pyrolysis and char oxida- tion. The flame was predicted to give off at 215  °C, and the sorghum was almost consumed completely at 433 °C. References The effect of heating rate on reactivity suggests that reac - Basu P (2013) Biomass gasification, pyrolysis and torrefaction, 2nd edn. Else - vier, London tor operation was preferably at 10  °C/min due to time Bhagwanrao SV, Singaravelu M (2014) Bulk density of biomass and particle effectiveness and no distinct mass change. The ignition density of their briquettes. Int J Agric Eng 7(1):221–224 and combustion indices were evidenced to rise to 12 and Boonmee N, Pongsamana P (2017) Spontaneous ignition of bagasse stockpiles in Thailand: a fire safety concern. Eng J 21(3):37–50. https:// doi. org/ 10. 10 times higher. 4186/ ej. 2017. 21.3. 37 Regarding sorghum storage, the stockpile was assumed Cao W, Li J, Lue L (2017) Study on the ignition behavior and kinetics of com- to comply with typical cylindrical and box silos in the bustion of biomass. Energy Procedia 142:136–141. https:// doi. org/ 10. 1016/j. egypro. 2017. 12. 022 large-scale industries, having diameter and height of 15 Carvalho WS, Oliveira TJ, Cardoso CR, Ataide CH (2015) Thermogravimetric and 10  m, respectively. Based on d/h ratios of 0.1–0.9, analysis and analytical pyrolysis of a variety of lignocellulosic sorghum. sorghum in the designated cylindrical silo (d = 15  m) Chem Eng Res Des 95:337–345. https:// doi. org/ 10. 1016/j. cherd. 2014. 11. had critical temperature showing a safe limit between Elorf A, Kandasamy J, Belandria V, Bostyn S, Sarh B, Gokalp I (2021) Heating rate 83–84  °C. Meanwhile, the spontaneous ignition is not effects on pyrolysis, gasification and combustion of olive waste. Biofuels possible to occur at ambient temperature below 84–87 °C 12(9):1157–1164. https:// doi. org/ 10. 1080/ 17597 269. 2019. 15945 98 Fennell LP, Boldor D (2014) Dielectric and thermal properties of sweet for the designated box silo (d = 10  m), within the incre- sorghum biomass. J Microw Power Electromagn Energy 48(4):244–260. ment of h/l and h/w combination ratios between 0.0625– https:// doi. org/ 10. 1080/ 08327 823. 2014. 11689 888 0.5 and 0.25–0.5. Fisher HG, Goetz DD (1993) Determination of self-accelerating decomposi- tion temperatures for self-reactive substances. J Loss Prev Process Ind 6(3):183–194. https:// doi. org/ 10. 1016/ 0950- 4230(93) 85008-9 Fraga LG, Silva J, Teixeira S, Soares D, Ferreira M, Teixeira J (2020) Influence of operating conditions on the thermal behavior and kinetics of pine wood Luthfi et al. Bioresources and Bioprocessing (2022) 9:49 Page 12 of 12 particles using thermogravimetric analysis. Energies 13(11):2756. https:// Polym Degrad Stab 93(1):90–98. https:// doi. org/ 10. 1016/j. polym degra doi. org/ 10. 3390/ en131 12756dstab. 2007. 10. 012 Fraga LG, Silva J, Teixeira S, Soares D, Ferreira M, Teixeira J (2020b) Thermal conversion of pine wood and kinetic analysis under oxidative and non- Publisher’s Note oxidative environments at low heating rate. Proceedings 58(1):23. https:// Springer Nature remains neutral with regard to jurisdictional claims in pub- doi. org/ 10. 3390/ WEF- 06921 lished maps and institutional affiliations. Huang S, Jia H, Sheng JJ (2016) Research on oxidation kinetics of tight oil from wolfcamp field. Pet Sci Technol 34(10):903–910. https:// doi. org/ 10. 1080/ 10916 466. 2016. 11747 14 IRENA (2021) World energy transitions outlook: 1.5 °C pathway. International Renewable Energy Agency, Abu Dhabi. https:// irena. org/ publi catio ns/ 2021/ Jun/ World- Energy- Trans itions- Outlo ok Islam MA, Auta M, Kabir G, Hameed BH (2016) A thermogravimetric analysis of the combustion kinetics of karanja (pongamia pinnata) fruit hulls char. Bioresour Technol 200:335–341. https:// doi. org/ 10. 1016/j. biort ech. 2015. 09. 057 Jayaraman K, Kok MV, Gokalp I (2017) Combustion properties and kinetics of different biomass samples using TG-MS technique. J Therm Anal Calorim 127:1361–1370. https:// doi. org/ 10. 1007/ s10973- 016- 6042-1 Khare P, Baruah BP (2014) Thermogravimetric analysis of perhydrous Indian coals. Energy Sour 36(7):774–782. https:// doi. org/ 10. 1080/ 15567 036. 2010. 547919 Li J, Paul MC, Younger PL, Watson I, Hossain M, Welch S (2016) Prediction of high-temperature rapid combustion behavior of woody biomass parti- cles. Fuel 165:205–214. https:// doi. org/ 10. 1016/j. fuel. 2015. 10. 061 Lu JJ, Chen WH (2015) Investigation on the ignition and burnout temperatures of bamboo and sugarcane bagasse by thermogravimetric analysis. Appl Energy 160:49–57. https:// doi. org/ 10. 1016/j. apene rgy. 2015. 09. 026 Lu L, Devasahayam S, Sahajwalla V (2013) Evaluation of coal for metallurgical applications. In: Osborne D (ed) The coal handbook: towards cleaner production. Woodhead Publishing, Cambridge, p 352 Luo Q, Liang D, Shen H (2016) Evaluation of self-heating and spontaneous combustion risk of biomass and fishmeal with thermal analysis (dsc-tg) and self-heating substances test experiments. Thermochim Acta 635:1–7. https:// doi. org/ 10. 1016/j. tca. 2016. 04. 017 Magalhaes D, Kazanc F, Riaza J, Erensoy S, Kabakli O, Chalmers H (2017) Combustion of Turkish lignites and olive residue: experiments and kinetic modelling. Fuel 203:868–876. https:// doi. org/ 10. 1016/j. fuel. 2017. 05. 050 Murasawa N, Koseki H, Iwata Y, Gao L (2013) Study on spontaneous ignition of stored food waste to be used for recycling. Fire Mater 37:520–529. https:// doi. org/ 10. 1002/ fam. 2144 Penvern H, Zhou M, Maillet B, Murias DC, Scheel M, Perrin J, Weitkamp T, Bardet S, Care S, Coussot P (2020) How bound water regulates wood drying. Phys Rev Appl 14:054051. https:// doi. org/ 10. 1103/ PhysR evApp lied. 14. Pickard S, Daood SS, Pourkashanian M, Nimmo W (2013) Robust extension of the Coats–Redfern technique: reviewing rapid and reliable reactivity analysis of complex fuels decomposing in inert and oxidizing thermo- gravimetric analysis atmosphere. Energy Fuels 27(5):2818–2826. https:// doi. org/ 10. 1021/ ef400 477u Rahib Y, Elorf A, Sarh B, Bonnamy S, Chaoufi J, Ezahri M (2019) Experimental analysis on thermal characteristics of argan nut shell (ans) biomass as a green energy resource. Int J Renew Energy Res 9(4):1606–1615. https:// doi. org/ 10. 20508/ ijrer. v9i4. 9724. g7763 Sami M, Annamalai K, Wooldridge M (2001) Co-firing of coal and biomass fuel blends. Prog Energy Combust Sci 27(2):171–214. https:// doi. org/ 10. 1016/ S0360- 1285(00) 00020-4 Sheng C, Yao C (2022) Review on self-heating of biomass materials: under- standing and description. Energy Fuels 36:731–761. https:// doi. org/ 10. 1021/ acs. energ yfuels. 1c033 69 Wilen C, Moilanen A, Kurkela E (1996) Biomass feedstock analyses. VTT Techni- cal Research Centre of Finland, Espoo. https:// cris. vtt. fi/ en/ publi catio ns/ bioma ss- feeds tock- analy ses Wnorowska J, Ciukaj S, Kalisz S (2021) Thermogravimetric analysis of solid biofuels with additive under air atmosphere. Energies 14(8):2257. https:// doi. org/ 10. 3390/ en140 82257 Xie Q, Tang S, Wang Z, Chen C, Xie P (2018) The prospect of sweet sorghum as the source for high biomass crop. J Agric Sci Bot 2(3):5–11. https:// doi. org/ 10. 35841/ 2591- 7897.2. 3.5- 11 Yao F, Wu Q, Lei Y, Guo W, Xu Y (2008) Thermal decomposition kinetics of natu- ral fibers: activation energy with dynamic thermogravimetric analysis. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Bioresources and Bioprocessing Springer Journals

Investigation into the combustion kinetics and spontaneous ignition of sweet sorghum as energy resource

Download PDF
12 pages

Loading next page...
 
/lp/springer-journals/investigation-into-the-combustion-kinetics-and-spontaneous-ignition-of-whIBOlzmDg

References (39)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2022
eISSN
2197-4365
DOI
10.1186/s40643-022-00536-0
Publisher site
See Article on Publisher Site

Abstract

Thermogravimetric analysis (TGA) has received Introduction immense attention for studying the combustion kinetics The priority of utilizing biomass has attracted much of biomass. Many studies have used TGA to profile its global attention to address environmental problems, characteristic temperatures (i.e., ignition temperature, such as global warming, due to the excessive use of fos- burnout temperature, and peak temperature(s)), kinetic sil fuels. Those concerns are reflected in IRENA’s 1.5  °C parameters (i.e., activation energy and pre-exponential scenarios on climate change, where bioenergy represents factor), and reactivity parameters (i.e., ignition index 25% of the total estimated primary supply or equals 153 and combustion index) in an oxidative environment. EJ by 2050 (IRENA 2021). Hence, the exploration of vari- The traced thermal degradation reveals the response of ous potential biomass for the power generation industry combustion reaction to various operating conditions, will be intensively carried out to meet the needs. such as biomass composition, heating rate, and others. For years, studies of combustion kinetic and sponta- The experimental study using olive residues was done by neous ignition have been proposed to understand the Magalhaes et  al. (2017). The characteristic temperatures essential characteristics of biomass during utilization were graphically identified from the TG and DTG curves and storage before being applied in large-scale power at 240–350, 340–700, 240–245, and 531–703  °C for industries. Direct combustion has become an efficient two combustion reaction stages, ignition, and burnout, technology for converting biomass into energy and has respectively. The apparent activation energies, which are been widely used in power plants with steam turbine sys- 44.2–47.7 and 9.1–13.4 kJ/mol, were presented by assum- tems. Thermal behavior is crucial information to explain ing reaction mechanisms and calculating them using the the ignition and burnout of fire which influences boiler model-fitting method of Coats–Redfern. operation, energy efficiency, and emissions (Cao et  al. Meanwhile, spontaneous ignition can be predicted in a 2017). Therefore, understanding combustion kinetic can biomass stockpile by considering its combustion kinetic help better knowledge of design and optimization in bio- parameters and applying the mathematical theory of mass combustion systems. Meanwhile, spontaneous igni- Frank-Kamenetskii. The theory expresses that the igni - tion may arise from the exothermic process due to the tion arises based on solid heat conduction in specific susceptibility of stockpiled materials to self-heating. The geometric shapes. Boonmee and Pongsamana (2017) physical, biological, and chemical heating mechanisms performed a laboratory-scale experiment using bagasse. are addressed to contribute to self-heating events (Sheng A cylindrical- or rectangular-shaped stockpile with any and Yao 2022). There were many reports of fire incidents radius or length stored with asymptotic heights below 10, during storage. For instance, a fire started from a large 7.8, and 6 m is considered safe from fire hazards at ambi - biomass pile of 500 tons at Advanced Agro-Power Plant ent temperatures of 40, 45, and 50 °C. in Thailand on 12 March 2017 due to accumulated heat. The current study focused on a comprehensive investi - For that reason, the knowledge of spontaneous ignition gation of combustion kinetics and spontaneous ignition can give a fundamental evaluation and mitigation of fire using sweet sorghum. Sorghum was chosen to represent hazards in biomass storage. L uthfi et al. Bioresources and Bioprocessing (2022) 9:49 Page 3 of 12 the primary marginal crop, which has many superior experiment was conducted 2 times under the same con- characteristics compared to other crops, such as grow- ditions to verify the reproducibility. A small sample mass ing in a short period of about 4–6 months and being able and low heating rate were applied to avoid the transport to withstand an environment with little water and high effect and enhance the resulting signal of slower kinetic soil salinity (Xie et al. 2018). TGA was performed at dif- events. ferent heating rates and in an oxidative environment to assess its response to reaction stages, characteristic tem- Combustion kinetic analysis peratures, kinetic parameters, and reactivity parameters. Determination of reaction stages Subsequently, the spontaneous ignition was predicted The combustion reaction region (exothermic reaction) by substituting the existing kinetic parameters into the was qualitatively determined by observing the gain por- steady-state solution of the Frank-Kamenetskii equation. tion of the DTA curve (see Fig.  1a). Then, by following The prediction was presented regarding the relationship the method of Pickard et  al. (2013) or tangent-bisection between ambient temperature and safe silo size. method (TBM), the reaction stages of combustion were identified on a DTG curve with the assumption that they Materials and methods are non-competing, first-order, and single-stage reac - Characteristics and preparation of sweet sorghum samples tions obeying the Arrhenius law (see Fig.  1b). Tangent Sweet sorghum was collected locally from farm areas lines were drawn to the edges of the leading and trailing around Mie University, Japan. Sorghum was appropri- curves. Bisection lines were extended at each tangent ately washed to eliminate dirt and was air-dried (± 20 °C) intersection until they reached the DTG trace and labeled for several days to reduce moisture content and inhibit as the start or end of the reaction stages. decay. Dried sorghum samples were ground using a high- speed blender, YKB (AS ONE Corp.), with a rotation speed of approximately 28,000 rpm for 1 min. All samples Determination of characteristic temperatures were sieved to a size range of 250–500 μm. According to The characteristic temperatures were determined using a Wilen et al. (1996), the proximate and ultimate character- TG and DTG curve, based on the report of Lu and Chen istics of sweet sorghum are summarized in Table 1. (2015) (see Fig.  2). Peak temperatures were observed at the maximum mass loss rate of each reaction stage. Then, TGA experimental approach the intersection method (IM) was employed to observe A thermal analyzer, EXSTAR 6000 TG/DTA 6200 (Seiko the ignition and burnout temperatures. Horizontal lines Instruments Inc.), was used to conduct the thermo- were drawn at points B and D, where the TG curve gravimetry (TG), derivative thermogravimetry (DTG), became steady after the evaporation and combustion and differential thermal analysis (DTA) experiments. reactions were complete. Afterward, tangent lines were Air was fed into the furnace from the pump and was drawn at points A and C through the horizontal lines of regulated to 100  ml/min for each experiment. Sam- points B and D, respectively. Points A and C are defined ples with a mass of 6  mg were used in each experiment as cross points at which vertical lines from the DTG and were heated from ambient temperature to 550 °C at curve peaks of stages I and II cross the TG curve. The three different heating rates of 2, 5, and 10  °C/min. The temperatures corresponding to the intersections of the Table 1 Proximate and ultimate characteristics of sweet sorghum ( Wilen et al. 1996) db Proximate analysis Value (%) Method Moisture content 7.04arb DIN 51,718 Ash content 4.74 DIN 51,719 Volatile matter 77.20 DIN 51,720 Fixed carbon 18.06 – db Ultimate analysis Value (%) Method Carbon 47.30 – Hydrogen 5.80 – Oxygen 41.67 By difference Nitrogen 0.40 – Sulphur 0.09 ASTM D 4239 Luthfi et al. Bioresources and Bioprocessing (2022) 9:49 Page 4 of 12 tangent and horizontal lines were labeled as the ignition and burnout temperatures, respectively. Determination of reactivity parameters In support of the reactivity data, the ignition and combustion indices were calculated for all samples to provide a more accurate measure of the conver- sion over time and temperature. According to Fraga et  al. (2020b), the ignition and combustion indices are defined in Eq. 1 and Eq. 2, respectively: (dm/dt) max D = , (1) t t max ig (dm/dt) (dm/dt) max avg S = , (2) T T ig where D and S are the ignition (in %/min ) and combus- 2  3 tion indices (in %/min °C ). (dm/dt) and (dm/dt) max avg are the maximum and average mass loss rates (in %/min), respectively. T and T are the ignition and burnout tem- ig b peratures (in °C). t and t are the times corresponding ig max to the ignition and maximum combustion rates (in min). Determination of kinetic parameters The Friedman method, which obeys the Arrhenius law, was used to estimate the kinetic data from a TG curve. Fig. 1 Schematic for determining a the combustion reaction region Since the amount of airflow is much greater than the and b combustion reaction stages (e.g., at a heating rate of 5 ºC/min; 1st iteration sample) sample mass used, the oxidation process is assumed not to depend on the oxygen concentration. Therefore, the combustion reaction was modeled as first-order kinetics. According to Yao et al. (2008) and Huang et al. (2016), the combustion of solid biomass is described by the following: dx = kf (x), (3) dt where dx/dt is the conversion rate (in 1/s), k is the reac- tion rate constant (in 1/s), and f(x) is the reaction func- tion. In addition, x is the degree of conversion. The expression for x is defined as: m − m i t x = , (4) m − m Fig. 2 Schematic for determining the peak, ignition, and burnout where m , m , and m are the initial mass of the sample, i t f temperatures (e.g., at a heating rate of 5 ºC/min; 1st iteration sample) mass of the sample at a specified time, and final mass of the sample (in kg), respectively. Furthermore, the expres- sion of k is defined by the Arrhenius equation: −E/RT (5) k = Ae , L uthfi et al. Bioresources and Bioprocessing (2022) 9:49 Page 5 of 12 where A is the pre-exponential factor (in 1/s), E is the Results and discussion activation energy (in J/mol), R is the universal gas con- Combustion reaction and characteristic temperatures stant (in J/molK), and T is the temperature of the sample of sweet sorghum (in K). Based on the graphical analysis of DTA, sweet sorghum Subsequently, substituting Eq. 5 into Eq. 3 gives the fol- was thermally degraded in two distinct regions of exo- lowing equations: thermic and endothermic reactions due to evapora- tion (region I) and combustion (region II). As shown dx −E/RT in Fig.  1a, the first gain in heat flow permits the identi - = Af (x)e . (6) dt fication of these two regions. The gain temperatures were identified at 226–223  °C for all heating rates. The The model-free method of Friedman is defined in Eq.  7 negative-valued region asserts that sorghum samples by taking a logarithm on both sides of Eq.  6 without absorbed heat to release surface and inherent moisture. assuming the reaction model and the reaction function. In contrast, the positive-valued region corresponds to The following equation is hence obtained: heat liberation due to lignocellulosic pyrolysis (i.e., dehy- dx E dration and decarboxylation of hemicellulose, cellulose ln = ln Af (x) − . (7) dt RT (> 400–450 °C), and lignin) and burning volatile and char (Rahib et  al. 2019; Basu 2013). The heating rate affected By plotting the left-hand side of Eq.  7 against 1/T at the onset of heat release by lowering the gain tempera- different heating rates followed by taking a linear regres - ture by 3 °C at higher heating rates and vice versa. It took sion, the activation energy (E) can be obtained from the place at approximately 226  °C for heating rates of 2  °C/ slope. Meanwhile, the intercept may be used to deter- min and 223 °C for rates of 5–10 °C/min. This is because mine the pre-exponential factor (A) if the reaction func- the heat given at a low rate provided enough time for tion is modeled beforehand (for first-order kinetics, it is water to diffuse or even break their bonds from vessels f(x) = (1-x)). and fibers completely (Penvern et  al. 2020). The force holding might derive from hydrogen bonds among the Spontaneous ignition analysis water in vessels or between the water molecules and The safe size of the stockpile and ambient temperature hydrophilic sites via functional groups, such as hydroxyls, under critical condition were evaluated using the dimen- phenolic, and carboxylic acids (Khare and Baruah 2014). sionless equation of Frank-Kamenetskii. The equation is The difference in the mass loss proves the presence of a expressed by Eq.  8, which represents the ratio of chemi- water concentration gradient in the sample. The smaller cal energy to thermal conduction (Boonmee and Pongsa- the mass loss is, the more water remains. TG curve shows mana 2017; Fisher and Goetz 1993): that the mass loss decreased from 82.51 to 86.80% at the transition point as the rate increased. �HEr AC E The combustion region, profoundly, was identified as δ = exp − , (8) RT RT c having multi-stage reactions, as depicted by the occur- rence of slopes and peaks on the TG and DTG curves where δ is the critical Damkohler number, and T is the c c (see Fig.  3). The tangent-bisection and intersection critical ambient temperature (in K). r is the characteristic methods were performed on the DTG curve to enhance dimension of a pile with origin at the center (in m), which confidence in quantifying the major reaction stages and is half of the total depth (r = 1/2r ) or equal to the total characteristic temperatures (see Fig.  1b and Fig.  2). For radius (r = r ). ΔH is the heat reaction (in J/kg), which is each temperature parameter obtained (i.e., reaction assumed to be equivalent to the calorific value (Q) mul - temperatures and characteristic temperatures) from all tiplied by the conversion degree (x) at the ignition tem- heating rates, extrapolations were subsequently done at perature (ΔH = x Q). C and λ are the bulk density (in kg/ ig 0 0 °C/min to exhibit the material nature and eliminate the m ) and thermal conductivity of the sample (in W/(mK)). linear influence of the heating procedure. On the other n is the kinetic reaction order. E and A are the activation hand, the remaining masses, showing the material nature, energy (in J/mol) and pre-exponential factor of reac- were determined by simply averaging the three corre- tion (in 1/s), respectively, which is obtained from kinetic sponding heating rates due to intrinsic property (Yao analysis at the ignition temperature. R is the universal gas et al. 2008). The obtained data are summarized in Table  2 constant (in J/(molK)). and Table 3. The robust methods divided the combustion reaction into two stages, which occurred at 131–336 and 336– 475  °C for stages I and II, respectively. The first stage is Luthfi et al. Bioresources and Bioprocessing (2022) 9:49 Page 6 of 12 while the chars remained (Sami et al. 2001). In a sufficient proportion, the mixed gases and/or some solids were predicted to give off flames at 215  °C due to oxidation, when cellulose became the dominant mechanism (Lu and Chen 2015). The pyrolysis reaction consumed 56.87% of the sorghum mass, with the reactivity rate increas- ing sharply at 264 °C (T ). Meanwhile, the second stage p,I corresponds to the char oxidation within the diffusion- controlled phase via mixed gas (Magalhaes et  al. 2017). In  situ oxidation of volatiles and chars might proceed in parallel due to overlapping. The transition from the burn - ing of volatile to char can be notified by the change of flame into an ember (Rahib et al. 2019). After the volatile matter in the atmosphere was exhausted, oxygen-rich air diffused further into the interior of the char and burned out the rest to ashes (Sami et  al. 2001). The ember was found to extinguish at 433 °C (T ). Overall, the oxidation reaction converted 24.44% of the sorghum mass with the highest reactivity at 405 °C (T ). p,II Following the preliminary identification, the appear - ance of the shoulder (less pronounced) and head peaks on the DTG curve at around 180–200 and 270–290  °C for all heating rates, respectively, indicates the pseudo- hemicellulose and pseudo-cellulose within the pyrolytic stage (see Fig. 3b). Both could also be marked at the first and second slopes of the TG curve, excluding the slope of water evaporation (see Fig.  3a). In contrast, the pseudo- lignin peak, which is believed to be indistinct due to its notoriously broad reaction, might fall around 320–420 °C according to the third TG slope. Meanwhile, the char oxi- dation reaction could be notified by the prominent DTG Fig. 3 a TG and b DTG signals with plots of the characteristic peak at 410–450 °C or the fourth TG slope. Similar loca- temperatures for all heating rates (1st iteration sample) tions of pseudo-lignocellulose were reported by Carvalho et  al. (2015) on sweet sorghum in an inert atmosphere using the deconvolution method. Accordingly, Jayaraman et al. (2017) confirmed that pyrolysis and char oxidation attributed to the pyrolysis and volatile oxidation con- occurred in similar ranges by investigating the evolu- trolled by chemical reactions (Magalhaes et  al. 2017). tion of typical gases from various biomass. A significant Slow pyrolysis (< 50 °C/min) produced volatiles and chars amount of water vapor (H O), carbon monoxide (CO), majorly through dehydration and decarboxylation of hydrogen (H ), and aromatic compounds (i.e., meth- hemicellulose (127–235  °C), cellulose (227–350  °C), and ane (CH ), benzene (C H ), and so on) was detected in 4 6 6 lignin (127–527  °C) (Basu 2013; Carvalho et  al. 2015), the range of 150–400  °C, showing the typical gaseous wherein the volatiles then diffused into the ambient air Table 2 Reaction temperatures and characteristic temperatures of the combustion reaction under different heating rates Heating rate (°C/min) Reaction temperature (°C) Characteristic temperature (°C) Stage I Stage II Tp,I Tp,II Tig Tb - (material nature) 131–336 336–475 264 405 215 433 2 143–348 348–475 268 412 218 444 5 168–362 362–486 281 435 229 467 10 197–391 391–487 292 452 236 495 L uthfi et al. Bioresources and Bioprocessing (2022) 9:49 Page 7 of 12 Table 3 Mass changes including reaction durations of the combustion reaction under different heating rates Heating rate (°C/min) Remaining mass, % (standard deviation) Mass loss, % (reaction duration, s) Maximum mass- loss rate (%/min) Stage I Stage II Stage I Stage II Tp,I Tp,II - (material nature) 91.01 (1.26)–34.14 (1.42) 34.14 (1.42)–9.70 (0.22) 56.87 24.44 – – 2 91.95–32.55 32.55–9.80 59.40 (6057) 22.75 (3846) 1.42 0.66 5 91.50–35.28 35.28–9.45 56.22 (2265) 25.83 (1506) 3.38 1.86 10 89.59–34.59 34.59–9.84 55.00 (1110) 24.74 (585) 6.94 17.69 products of dehydration, decarboxylation, and scission from 3846 to 585  s during stage II. This implies that the in pyrolysis reactions. In contrast, carbon dioxide (C O ) combustion reaction proceeded faster at higher heating was released more in char oxidation between 400–750 °C rates without any obvious change in the amount of mass along with a small release of CO, representing the partial converted. oxidation in char. The maximum mass-loss rates at the corresponding Furthermore, increasing the heating rate affected the stages rose from 1.42 to 6.94%/min and from 0.66 to TG and DTG curves to shift to higher temperatures, 17.69%/min as the heating rate increased, agreeing with marked by a significant delay of 10–50  °C for both tem - the mass loss to reaction duration ratio (see Table  3). perature parameters as shown in Fig.  3 and Table  2. The values define the rate at which the non-condensed The significant linear influence is a consequence of the and condensed phases of biomass decompose to gases. decrease in the time required for heat transfer to cross As can be observed, stage II, which was dominated by the biomass interior before the surface temperature char, had lower mass-loss rates than stage I due to the increases (Elorf et  al. 2021). A major deterrent to the high energy bond of the carbon–carbon bond. Cor- higher heating rate is the low thermal conductivity of sor- respondingly, the susceptibility of sorghum charac- ghum by 0.13 W/mK (Fennell and Boldor 2014). For this terized by a high fuel ratio (> 2.0) of volatile matter to reason, even though the heating rate was high, the sor- fixed carbon (4.27) exerted influences on the results of ghum interior was heated at a considerably slower. the early stage (Lu et  al. 2013; Wilen et  al. 1996), such as low ignition temperature (200–300  °C) and higher Combustion reactivity of sweet sorghum maximum mass-loss rate during devolatilization (see Changes in heating rate exhibit unique effects on the Table  1) (Basu 2013). Comparing the rate magnitude mass changes and reaction durations. Based on Fig. 3 and of the two, stage II rose approximately ten times higher Table  3, the mass losses fluctuated during combustion under 5–10  °C/min. In contrast, stage I increased with a tendency to decrease insignificantly (1–4%) as the merely twofold at the same level of heating rate. It may heating rate increased. Stage I experienced a decrease in be caused by the ability of oxygen to overcome mass the mass loss by 59.40–55.00% at a heating rate of 2 to transfer resistance and diffuse at high heating rates 10 °C/min. Meanwhile, the mass loss at stage II increased (Islam et  al. 2016). Unlike the devolatilization process from 2 to 5 °C/min by 22.75–25.83% and decreased from which is only sensitive to the particle temperature, the 5 to 10  °C/min by 25.83–24.74%. Fluctuating results reactivity of char oxidation can accelerate due to tem- were also obtained in the report of Fraga et  al. (2020a) perature elevation and oxidizer concentration (Li et  al. and Jayaraman et  al. (2017) using various biomasses 2016), which may affect the present findings in the sim - and broader ranges of heating rates. Fraga et  al. (2020a) ilar way. deduced this to be a form of randomness. Additionally, Supporting the given trends in the maximum mass- the insignificance of the mass changes was expressed in loss rate, the change in heating rate shows a directly pro- small standard deviations (SD) of the remaining masses, portional relationship with the ignition and combustion namely 1.24 (1.38%; relative standard deviation (RSD)), indices. Both indices signify the decomposition over time 1.42 (4.16%), and 0.22% (2.27%) for the onset of stage I, and temperature within combustion stages. The ignition transition point, and offset of stage II, respectively. This and combustion indices were obtained to increase from 4 4 3 8 implies that the mass change was probably an intrinsic 1.21 × 10 to 192.17 × 10 %/min and from 3.37 × 10 to 8 2 3 property and not affected by the heating rate (Yao et  al. 180.14 × 10 %/min °C , respectively, under 2–10  °C/min 2008). Conversely, the increase in heating rate alone heating rates (see Fig. 4). Substantially, the ignition index significantly affected the reaction durations, denoted suggests that sorghum caught fire faster with magni - by the decrease from 6057 to 1110  s during stage I and tudes of approximately 13–12 times higher as the heating Luthfi et al. Bioresources and Bioprocessing (2022) 9:49 Page 8 of 12 rate increased. This index is assigned by concerning the amount of volatile evolution within the first combustion stage or pyrolysis reaction. Meanwhile, the combustion index exhibits the activity of the substance consumption, starting when the sorghum caught fire until the chars burned out into ashes, signifying the predominance of oxidation. Following the same previous causal relation- ship, the entire combustion reaction proceeded rapidly by the increment of about 5- to 10-fold (Wnorowska et al. 2021; Fraga et al. 2020b). Kinetic behavior of sweet sorghum The Friedman method was employed to evaluate the kinetic data of the overall combustion reaction by linearly Fig. 4 Eec ff ts of different heating rates on the ignition and fitting the term ln(dx/dt) to 1/T in Eq.  7 for a series of combustion indices conversion degrees at different heating rates. The conver - sion degree in the range of 0.1–0.9 was used in this study with segmentation of 0.1. Figure 5 shows that fitted lines were nearly parallel in three conversion ranges of 0.1–0.5, 0.6–0.7, and 0.8–0.9, denoting a single reaction or uni- fication of several reactions (Yao et  al. 2008). Averaging over the corresponding parallel ranges, then, will give meaningful activation energy. To avoid overestimation, the relative standard deviation (RSD) was determined to be no more than 10% (see Table  4). The correlation coefficient values demonstrated the coherence of kinetic parameters generated by the Friedman method. Fraga et  al. 2020b stated that the first-order kinetic model can be considered to fit well if the coefficient is more than 0.90 (R > 0.90). Because the correlation coefficients of this study ranged from 0.9893 to 0.9999 for the whole conversion degree, these results indicate the adequacy of the kinetic parameters to represent good linear fitted plots and reaction mechanism (i.e., model and function). Fig. 5 Linear plots of ln(dx/dt) against 1/T for various conversion The distribution of activation energies and pre-expo - degree in Friedman method representing the entire combustion reaction nential factors in the function of conversion degree is Table 4 Activation energies, pre-exponential factors, and correlation coefficients (R ) for the entire combustion reaction using Friedman method Conversion (mg/ Friedman method Conversion (mg/ Friedman method mg) mg) E (kJ/mol) A (1/s) R E (kJ/mol) SD Relative SD avg 0.1 129.57 5.25 × 10 0.9998 0.1–0.5 132.91 8.60 6.47 0.2 143.38 7.30 × 10 0.9999 0.6–0.7 79.40 7.21 9.07 0.3 140.19 2.14 × 10 0.9992 0.8–0.9 169.44 2.45 1.45 0.4 132.49 2.21 × 10 0.9994 0.5 118.92 6.00 × 10 0.9961 0.6 72.20 1.26 × 10 0.9918 0.7 86.61 6.30 × 10 0.9994 0.8 166.99 7.97 × 10 0.9893 0.9 171.89 1.41 × 10 0.9993 L uthfi et al. Bioresources and Bioprocessing (2022) 9:49 Page 9 of 12 It is important to note that satisfying activation energy is required to represent the entire reaction. According to Yao et  al. (2008) and Luo et  al. (2016), fractions below 0.5–0.7, showing parallel plots at the beginning, generally provide meaningful information instead of the whole process, especially for the case of self-heating and combustion. The activation energy, which encompasses the ignition process, determines the external energy to heat the fuel reactor since the subsequent reaction will become self-sustaining above this temperature, related to exothermic chain reac- tions. Therefore, the apparent activation energy of the combustion reaction in this study was 132.91  kJ/mol regardless of the heating rate. Fig. 6 Apparent activation energy and pre-exponential factor as a function of conversion degree calculated by Friedman method Spontaneous ignition of stockpiled sweet sorghum The dimensionless parameter of Frank-Kamenetskii or critical Damkohler number (δ ), expressed by Eq.  8, was shown in Fig.  6. Converting the conversion degree of used to predict spontaneous ignition based on the energy 0.1–0.5 to temperature range from 218 to 317 °C, which equation at steady state. The thermal explosion model has the exact position in the applicable heating rates of was derived by assuming that the Biot number is equiva- 2, 5, and 10  °C/min, the unified reaction was allegedly lent to infinity (Bi = ∞). The number means that reactant suspected to be conjoint pyrolysis of hemicellulose and temperature is distributed along the body in a parabolic cellulose predominantly. The mean activation energy manner, governed by internal conduction. Thus, the was found to be 132.91  kJ/mol, while the pre-expo- critical Damkohler number depends specifically on the 9 7 −1 nential factor ranged from 5.25 × 10 to 6.00 × 10  s , geometry of reacting systems (Fisher and Goetz 1993). indicating the frequency of molecular collision for the Current work studied the relationship between criti- reaction to occur (see Table 4). Thus, the lower the acti - cal ambient temperature and silo dimension in sorghum vation energy is, the easier it is for a reaction to com- storage. Since feedstock is usually piled up in a non- mence. Considering the same way to elucidate the data, geometric shape, the stockpile was assumed to comply lignin-dominated pyrolysis underwent in the second with two common geometric shapes in the market: (1) a half of the conversion degree (0.6–0.7), settling in the cylinder with dimensions of diameter (d) and height (h); temperature from 304 to 395  °C. The required activa - and (2) a rectangular box with dimensions of height (h), tion energy decreased to 79.40 kJ/mol, presumably due length (l), and width (w). If the critical ambient tempera- to the least presence of hemicellulose and cellulose ture is set within a specific range, and the Damkohler remaining at the high temperature. At a conversion number for a given geometry is specified, the response degree of 0.8–0.9, equivalent to a temperature range of the selected silo dimension to the critical temperature from 386 to 453 °C, char oxidation was presumed to be can be calculated. According to Fisher and Goetz (1993), the reaction to occur. The minimum energy for oxida - Damkohler numbers for finite cylinder (heat loss one tion to react increased drastically to 169.44 kJ/mol as a end) and rectangular box are shown in Eq.  9 and Eq.  10, consequence of the gradual deposition of carbon con- respectively. stituents. Additionally, both possible reactions were disclosed to have pre-exponential factors spreading δ = 2.0 + 0.195(d/h) , (9) 3 3 9 from 1.26 × 10 to 6.30 × 10 and from 7.97 × 10 to 10 −1 1.41 × 10  s , respectively. Table 5 Calculation parameters used in Frank-Kamenetskii model Parameter Value Parameter Value Parameter Value 5 3 T (K) 273.15–473.15 ΔH (J/kg) 17.79 × 10 E (J/mol) 129.57 × 10 3 5 9 C (kg/m ) 38.86 Q (J/kg) 173.91 × 10 A (1/s) 5.25 × 10 λ ( W/mK) 0.13 xig 0.102 R (J/molK) 8.314 Luthfi et al. Bioresources and Bioprocessing (2022) 9:49 Page 10 of 12 Table 6 Geometric shapes and dimension ratios used in Frank-Kamenetskii model Geometric shape Dimension ratio Geometric shape Dimension ratio d/h Note h/l h/w Note Cylinder 1/10 (0.1) (h > d) Box 1/2 (0.5) 1/2 (0.5) (h < l,w) (l > w) 1/3.3 (0.3) 1/4 (0.25) 1/2 (0.5) 1/2 (0.5) 1/8 (0.125) 1/2 (0.5) 1/1.4 (0.7) 1/4 (0.25) 1/4 (0.25) 1/1.1 (0.9) 1/8 (0.125) 1/4 (0.25) 1/16 (0.0625) 1/4 (0.25) 2 2 δ = 0.825(1.067 + (h/l) + (h/w) ). (10) Regarding the calculations, the critical ambient temperature (T ) was set between 0–200  °C (273.15– 473.15  K) to represent the relative reference that changed in the input (see Table  5). A series of diam- eter and height were generated as the selected dimen- sions of cylindrical and box silos, which are equivalent to twofold the characteristic dimension (r). Meanwhile, other storage dimensions were converted into ratios such as d/h, h/l, and h/w, so that Eq. 9 and Eq. 10 could be calculated mathematically and substituted into Eq. 8 (see Table 6). Concerning the rest parameter in Table 5, the thermal conductivity (λ) was cited from the work of Fennel and Boldor (2014), while the bulk density (C ) and calorific value (Q ) were, respectively, experimented based on the method of Bhagwanrao and Singaravelu (2014) and the standard of adiabatic calorimeter (O.S.K 150, Ogawa Sampling Co., Japan). Upon extrapolating at 0 °C/min, the conversion degree of ignition (x ) was ig multiplied by calorific value to determine the heat reac - tion (ΔH). Meanwhile, the activation energy (E) and pre-exponential factor (A) were fitted to the linear plots in Fig.  5, which satisfied the conversion degree of igni - tion of 0.102. Observing Fig.  7, the resulting curves suggest that at certain levels of temperature and silo size, the areas under the curves are appropriate for storing sorghum piles safely. The opposite meaning applies to the area above the curve. For both geometric shapes of cylinder and box, the curves exhibit vertical (va) and horizon- Fig. 7 Relationship between the critical ambient temperature and tal asymptotes (ha) around the coordinates y = 8  km; va selected silo dimension for a cylindrical and b rectangular-boxed x = 60 °C and y = 6 km; x = 60 °C when the selected va ha ha shapes, including the enlarged curves to show the effect of different dimensions were increased or decreased to near infinity dimension ratios while keeping the respective ratios. The findings con - firm the report of Murasawa et al. (2013) on storage of soy sauce residue and fishmeal, in which similar asymp - 10 m for both geometric shapes to study the spontane- totic responses were detected. ous ignition of sorghum on a relevant scale (see Fig. 7a The market designs were employed to present and Fig. 7b). These values were chosen because they are enlarged curves around diameter and height of 15 and L uthfi et al. Bioresources and Bioprocessing (2022) 9:49 Page 11 of 12 Abbreviations commonly used in large-scale industries and are decent TGA : Thermogravimetric analysis; TG: Thermogravimetry; DTG: Derivative to apply to several dimension ratios in the current thermogravimetry; DTA: Differential thermal analysis; TBM: Tangent-bisection study. Based on the cylindrical silo design (d = 15  m), method; IM: Intersection method. the spontaneous ignition of sorghum will not occur at Acknowledgements ambient temperature below 83–84 °C, within the incre- Not applicable. ment of d/h ratio from 0.1 to 0.9. It is possible to notify Author contributions that the shift due to various d/h ratios was only 1  °C. The contributions of each author are as follows: NL performed the experi- Similar findings were found in the box-shaped silo ments, analyzed the data obtained, and wrote the manuscript. TO performed design (h = 10  m). The critical temperature showing the experiments and analyzed the data obtained. Y T administered the project and acquired funding. TF supervised the study. KT conceptualized and super- a safe limit for spontaneous ignition ranged between vised the study. All authors read and approved the final manuscript. 84–87  °C. The temperature shifted by 3  °C as the h/l and h/w ratios rose from 0.0625 to 0.5 and from 0.25 Funding This study was funded by Chubu Electric Power Co., Inc., Aichi, Japan. to 0.5, respectively. The combination of the two ratios is shown in Table  6. In addition to finding the critical Availability of data and materials temperatures for each design, although not significant, All data analyzed during this study are included in this article. it can be seen that the smaller the resized dimension to the ratio is, the greater the ambient temperature for Declarations a fire to ignite is required, owing to the immense heat Ethics approval and consent to participate dissipation. Not applicable. Consent for publication Conclusions Not applicable. The combustion kinetics and spontaneous ignition of sweet sorghum have been investigated successfully using Competing interests The authors declare that they have no competing interests. TGA and the Frank-Kamenetskii theory. The findings are highly related to sorghum utilization as a fuel in a com- Author details bustion reactor and to safety storage as a feedstock from Department of Environmental Science and Technology, Graduate School of Bioresources, Mie University, 1577 Kurimamachiyacho, Tsu, Mie 514-8507, fire hazards. Investigation shows that activation energy of Japan. Staff Service, 85 Kandaneribeicho, Chiyoda-ku, Tokyo 101-0022, Japan. 132.91  kJ/mol was required to undergo the combustion Department of Life Sciences, Graduate School of Bioresources, Mie University, in a reactor, fitted well by the first-order model. The com - 1577 Kurimamachiyacho, Tsu, Mie 514-8507, Japan. bustion of sweet sorghum occurred in the temperature Received: 5 December 2021 Accepted: 7 April 2022 range of 131–475  °C and comprised two different ther - mal stages, corresponding to pyrolysis and char oxida- tion. The flame was predicted to give off at 215  °C, and the sorghum was almost consumed completely at 433 °C. References The effect of heating rate on reactivity suggests that reac - Basu P (2013) Biomass gasification, pyrolysis and torrefaction, 2nd edn. Else - vier, London tor operation was preferably at 10  °C/min due to time Bhagwanrao SV, Singaravelu M (2014) Bulk density of biomass and particle effectiveness and no distinct mass change. The ignition density of their briquettes. Int J Agric Eng 7(1):221–224 and combustion indices were evidenced to rise to 12 and Boonmee N, Pongsamana P (2017) Spontaneous ignition of bagasse stockpiles in Thailand: a fire safety concern. Eng J 21(3):37–50. https:// doi. org/ 10. 10 times higher. 4186/ ej. 2017. 21.3. 37 Regarding sorghum storage, the stockpile was assumed Cao W, Li J, Lue L (2017) Study on the ignition behavior and kinetics of com- to comply with typical cylindrical and box silos in the bustion of biomass. Energy Procedia 142:136–141. https:// doi. org/ 10. 1016/j. egypro. 2017. 12. 022 large-scale industries, having diameter and height of 15 Carvalho WS, Oliveira TJ, Cardoso CR, Ataide CH (2015) Thermogravimetric and 10  m, respectively. Based on d/h ratios of 0.1–0.9, analysis and analytical pyrolysis of a variety of lignocellulosic sorghum. sorghum in the designated cylindrical silo (d = 15  m) Chem Eng Res Des 95:337–345. https:// doi. org/ 10. 1016/j. cherd. 2014. 11. had critical temperature showing a safe limit between Elorf A, Kandasamy J, Belandria V, Bostyn S, Sarh B, Gokalp I (2021) Heating rate 83–84  °C. Meanwhile, the spontaneous ignition is not effects on pyrolysis, gasification and combustion of olive waste. Biofuels possible to occur at ambient temperature below 84–87 °C 12(9):1157–1164. https:// doi. org/ 10. 1080/ 17597 269. 2019. 15945 98 Fennell LP, Boldor D (2014) Dielectric and thermal properties of sweet for the designated box silo (d = 10  m), within the incre- sorghum biomass. J Microw Power Electromagn Energy 48(4):244–260. ment of h/l and h/w combination ratios between 0.0625– https:// doi. org/ 10. 1080/ 08327 823. 2014. 11689 888 0.5 and 0.25–0.5. Fisher HG, Goetz DD (1993) Determination of self-accelerating decomposi- tion temperatures for self-reactive substances. J Loss Prev Process Ind 6(3):183–194. https:// doi. org/ 10. 1016/ 0950- 4230(93) 85008-9 Fraga LG, Silva J, Teixeira S, Soares D, Ferreira M, Teixeira J (2020) Influence of operating conditions on the thermal behavior and kinetics of pine wood Luthfi et al. Bioresources and Bioprocessing (2022) 9:49 Page 12 of 12 particles using thermogravimetric analysis. Energies 13(11):2756. https:// Polym Degrad Stab 93(1):90–98. https:// doi. org/ 10. 1016/j. polym degra doi. org/ 10. 3390/ en131 12756dstab. 2007. 10. 012 Fraga LG, Silva J, Teixeira S, Soares D, Ferreira M, Teixeira J (2020b) Thermal conversion of pine wood and kinetic analysis under oxidative and non- Publisher’s Note oxidative environments at low heating rate. Proceedings 58(1):23. https:// Springer Nature remains neutral with regard to jurisdictional claims in pub- doi. org/ 10. 3390/ WEF- 06921 lished maps and institutional affiliations. Huang S, Jia H, Sheng JJ (2016) Research on oxidation kinetics of tight oil from wolfcamp field. Pet Sci Technol 34(10):903–910. https:// doi. org/ 10. 1080/ 10916 466. 2016. 11747 14 IRENA (2021) World energy transitions outlook: 1.5 °C pathway. International Renewable Energy Agency, Abu Dhabi. https:// irena. org/ publi catio ns/ 2021/ Jun/ World- Energy- Trans itions- Outlo ok Islam MA, Auta M, Kabir G, Hameed BH (2016) A thermogravimetric analysis of the combustion kinetics of karanja (pongamia pinnata) fruit hulls char. Bioresour Technol 200:335–341. https:// doi. org/ 10. 1016/j. biort ech. 2015. 09. 057 Jayaraman K, Kok MV, Gokalp I (2017) Combustion properties and kinetics of different biomass samples using TG-MS technique. J Therm Anal Calorim 127:1361–1370. https:// doi. org/ 10. 1007/ s10973- 016- 6042-1 Khare P, Baruah BP (2014) Thermogravimetric analysis of perhydrous Indian coals. Energy Sour 36(7):774–782. https:// doi. org/ 10. 1080/ 15567 036. 2010. 547919 Li J, Paul MC, Younger PL, Watson I, Hossain M, Welch S (2016) Prediction of high-temperature rapid combustion behavior of woody biomass parti- cles. Fuel 165:205–214. https:// doi. org/ 10. 1016/j. fuel. 2015. 10. 061 Lu JJ, Chen WH (2015) Investigation on the ignition and burnout temperatures of bamboo and sugarcane bagasse by thermogravimetric analysis. Appl Energy 160:49–57. https:// doi. org/ 10. 1016/j. apene rgy. 2015. 09. 026 Lu L, Devasahayam S, Sahajwalla V (2013) Evaluation of coal for metallurgical applications. In: Osborne D (ed) The coal handbook: towards cleaner production. Woodhead Publishing, Cambridge, p 352 Luo Q, Liang D, Shen H (2016) Evaluation of self-heating and spontaneous combustion risk of biomass and fishmeal with thermal analysis (dsc-tg) and self-heating substances test experiments. Thermochim Acta 635:1–7. https:// doi. org/ 10. 1016/j. tca. 2016. 04. 017 Magalhaes D, Kazanc F, Riaza J, Erensoy S, Kabakli O, Chalmers H (2017) Combustion of Turkish lignites and olive residue: experiments and kinetic modelling. Fuel 203:868–876. https:// doi. org/ 10. 1016/j. fuel. 2017. 05. 050 Murasawa N, Koseki H, Iwata Y, Gao L (2013) Study on spontaneous ignition of stored food waste to be used for recycling. Fire Mater 37:520–529. https:// doi. org/ 10. 1002/ fam. 2144 Penvern H, Zhou M, Maillet B, Murias DC, Scheel M, Perrin J, Weitkamp T, Bardet S, Care S, Coussot P (2020) How bound water regulates wood drying. Phys Rev Appl 14:054051. https:// doi. org/ 10. 1103/ PhysR evApp lied. 14. Pickard S, Daood SS, Pourkashanian M, Nimmo W (2013) Robust extension of the Coats–Redfern technique: reviewing rapid and reliable reactivity analysis of complex fuels decomposing in inert and oxidizing thermo- gravimetric analysis atmosphere. Energy Fuels 27(5):2818–2826. https:// doi. org/ 10. 1021/ ef400 477u Rahib Y, Elorf A, Sarh B, Bonnamy S, Chaoufi J, Ezahri M (2019) Experimental analysis on thermal characteristics of argan nut shell (ans) biomass as a green energy resource. Int J Renew Energy Res 9(4):1606–1615. https:// doi. org/ 10. 20508/ ijrer. v9i4. 9724. g7763 Sami M, Annamalai K, Wooldridge M (2001) Co-firing of coal and biomass fuel blends. Prog Energy Combust Sci 27(2):171–214. https:// doi. org/ 10. 1016/ S0360- 1285(00) 00020-4 Sheng C, Yao C (2022) Review on self-heating of biomass materials: under- standing and description. Energy Fuels 36:731–761. https:// doi. org/ 10. 1021/ acs. energ yfuels. 1c033 69 Wilen C, Moilanen A, Kurkela E (1996) Biomass feedstock analyses. VTT Techni- cal Research Centre of Finland, Espoo. https:// cris. vtt. fi/ en/ publi catio ns/ bioma ss- feeds tock- analy ses Wnorowska J, Ciukaj S, Kalisz S (2021) Thermogravimetric analysis of solid biofuels with additive under air atmosphere. Energies 14(8):2257. https:// doi. org/ 10. 3390/ en140 82257 Xie Q, Tang S, Wang Z, Chen C, Xie P (2018) The prospect of sweet sorghum as the source for high biomass crop. J Agric Sci Bot 2(3):5–11. https:// doi. org/ 10. 35841/ 2591- 7897.2. 3.5- 11 Yao F, Wu Q, Lei Y, Guo W, Xu Y (2008) Thermal decomposition kinetics of natu- ral fibers: activation energy with dynamic thermogravimetric analysis.

Journal

Bioresources and Bioprocessing – Springer Journals

Published: Apr 28, 2022

Keywords: Sweet sorghum; Thermogravimetric analysis; Frank-Kamenetskii theory; Combustion kinetics; Spontaneous ignition

There are no references for this article.