Experimental optimization of cynara cardunculus l. And sewage sludge pyrolysis in a fluidized bed reactor
- Autores: Andrés Morato Godino
- Directores de la Tesis: Antonio Soria Verdugo (dir. tes.), Sergio Sánchez Delgado (codir. tes.)
- Lectura: En la Universidad Carlos III de Madrid ( España ) en 2020
- Idioma: español
- Tribunal Calificador de la Tesis: Mercedes de Vega Blázquez (presid.), Isabel Ortiz González (secret.), Andrés Anca Couce (voc.)
- Programa de doctorado: Programa de Doctorado en Ingeniería Mecánica y de Organización Industrial por la Universidad Carlos III de Madrid
- Materias:
- Enlaces
- Tesis en acceso abierto en: e-Archivo
- Resumen
The continuous growth of the global population and the industrialization of both developed and developing countries are the main reasons for the current increasing energy demand. This increasing requirement of energy is mainly covered using fossil fuels. Considering the environmental problems caused by the use of fossil fuels, which contributes to enhance the greenhouse effect due to its associated pollution, alternative energy sources must be explored. In addition, predictions foresee a fossil fuel depletion within the next 100 years, what forces humanity to develop new ways of converting energy. Among the different alternatives to fossil fuels, biomass is considered a promising source of energy due to its net neutral CO2 life cycle. There are several technologies available to obtain energy from this renewable source, among which pyrolysis stands out for producing an organic liquid called bio-oil, what enables the decentralization of the energy consumption from the point of generation, together with combustible gases and a solid residue named char. The quantity and quality of the pyrolysis products are strongly influenced by the operating conditions under which the pyrolysis of the fuel is carried out. Therefore, controlling the operating conditions is essential to obtain high quality products and influence their production in terms of quantity. The most common parameters adjusted for this purpose are reactor temperature, heating rate, and residence time of the release vapors inside the reactor. A technology capable of providing good control over these operating conditions is fluidization, which consists in forcing a gas flow to pass through a bed of granular solid material from bottom to top of the bed. The solid material behaves then like a liquid, enabling fuel particles to move inside, while maintaining its inherent heat transfer capacity. In addition, the good solid mixing attained in fluidized bed reactors, together with the mentioned heat transfer capacity, results in a homogenous thermal profile inside the reactor. Hence, this technology is suitable for converting fuel particles with a good control of the operating conditions, and thus, this is the technology selected for the study of biomass pyrolysis conducted in this dissertation.
The biomass species subjected to study in this thesis, namely Cynara cardunculus L. and sewage sludge, were characterized performing proximate and ultimate analyses and heating value tests. The pyrolysis process of both biomass species was investigated separately, by means of non-isothermal thermogravimetric analysis (TGA). The simplified Distributed Activation Energy Model (DAEM) was employed to obtain the pyrolysis kinetic parameters of the samples, i.e., the activation energy and the pre-exponential factor. The temperature profile programmed to the TGA consisted of two processes occurring in series, first a drying process of the sample at 105 °C and then the pyrolysis process taking place when increasing the temperature of the sample in an inert atmosphere up to 800 °C. Nine different heating rates of 10, 13, 16, 19, 22, 25, 30, 35, and 40 K/min were used for each sample in a thermogravimetric analyzer (TGA) to obtain accurate values of the pyrolysis kinetic parameters when applying simplified DAEM. A mass of 10.0±0.5 mg of sample, sieved previously under 100 µm, was employed in the pyrolysis measurements in the TGA to reduce heat and mass transfer effects in the sample. The accurate values of the kinetic parameters that characterize the pyrolysis reaction of C. cardunculus L. and sewage sludge were reported. The activation energy and pre-exponential factor for the C. cardunculus L. samples are around 150 kJ/mol and 1012 s-1, respectively, for most of the pyrolysis conversion, increasing in both cases at the end of the process. In contrast, values ranging from 200 to 400 kJ/mol were obtained for the sewage sludge activation energy, and from 1015 to 1025 s-1 for its pre-exponential factor. From the kinetic parameters obtained, a representative value for the rate coefficient of pyrolysis of both samples was calculated, resulting in a rate coefficient of 4.73 min-1 and 2.03 min-1 for C. cardunculus L. and sewage sludge, respectively.
The first experimental campaign studied the pyrolysis of sewage sludge in a stainless-steel reactor operated as a fixed or fluidized bed. Silica sand was employed as bed material since it is known to be inert, not affecting the reaction rate during the thermochemical decomposition of biomass. The silica sand particle diameter, dbm, was in the range of 425-600 µm and the particle density, ρbm, was 2600 kg/m3. A mass of 240 g of sand was used in each test to reach a fixed bed height, hb, of 9.4 cm (bed aspect ratio hb/di = 2), corresponding to a void fraction, ε, of 0.44. The variation of the density, ρg, and the dynamic viscosity, μg, of the gas with temperature was considered to determine the minimum fluidization velocity, Umf, using the correlation of Carman-Kozeny. The sewage sludge pyrolysis experiments were conducted employing a novel technique to obtain the time evolution of the mass of sewage sludge supplied to the bed in batch during its pyrolysis while moving freely in the bed. This technique consisted in recording the mass signal measured by the scale while the pyrolysis process of the sewage sludge sample was taking place inside the reactor. Therefore, the mass released during the pyrolysis of the sampled could be determined. First, the reactor, filled with the sand particles that conformed the bed, was heated up by the electric resistors surrounding the reactor to the desired reactor temperature, T, while an air flow rate was used as fluidizing agent. Once the reactor temperature of the test was reached, the fluidizing gas was switched to nitrogen, and the flow rate was adjusted. When the operating conditions of the bed, i.e., reactor temperature and nitrogen flowrate, were selected, the scale, on which the reactor rested, was tared and a batch of around 10 g of dry sewage sludge particles was introduced through the top of the reactor. The particle size of sewage sludge was dp < 3 mm and they were dried at 105 °C in a universal oven UFE 500 from Memmett for 5 h, after which no mass variation of the sample was detected. Each experimental measurement was replicated to test the reproducibility of the experimental procedure, obtaining deviations lower than 5%. From the measurement of the mass of the solid residue remaining in the reactor, the pyrolysis time of the sewage sludge sample can be obtained accurately for each operating condition. Different operating conditions were selected to analyze the time evolution of the sample mass during the pyrolysis process, including variations of the bed temperature (T = 500 and 600°C) and the velocity of the nitrogen used as inert gas (U/Umf = 0.8, 1, 1.5, 2, 2.5, and 3), from a velocity lower than Umf, corresponding to a fixed bed reactor, to 3 times Umf, which induces a bubbling fluidized bed regime in the reactor. An increase of the velocity of nitrogen from that of a fixed bed, 0.8 Umf, to that of a low velocity bubbling fluidized bed, 2.5 Umf, accelerated remarkably the pyrolysis process, i.e., reduced the pyrolysis time. However, increasing the nitrogen velocity further had a slight effect on the characteristic velocity of the pyrolysis process. The pyrolysis process of sewage sludge could also be accelerated by increasing the bed temperature, even though the effect of temperature was lower than that of the nitrogen velocity. The percentage of volatile matter released by the samples during the pyrolysis process in the reactor was very similar to that obtained in the TGA, provided that the nitrogen velocity is sufficient to induce a proper fluidization of the bed (U/Umf ≥ 1.5). However, when the pyrolysis occurred in a fixed bed (U/Umf = 0.8) or in a bed at minimum fluidization velocity (U/Umf = 1), the mass released obtained from the reactor was lower than that of the TGA. This can be attributed to heat transfer effects inside the sample when no bubbles are present in the bed (U/Umf ≤ 1) and the fuel particles accumulate on the bed surface after being supplied as a batch through the bed top and, thus, the low conduction of heat inside this accumulation of fuel particles in the surface is relevant. When the gas velocity is increased above the minimum fluidization velocity (U/Umf > 1), bubbles appear in the bed and induce the motion of fuel particles, breaking the typical accumulation of fuel found in fixed beds, and enhancing the axial dispersion of fuel inside the bed. Therefore, in the case of fluidized beds, the fuel particles are separated from each other due to the higher dispersion of fuel induced by the presence of bubbles, hence the effect of heat transfer inside the sample is reduced, and the heating rate is increased. Furthermore, a mathematical model based on a first order apparent kinetics for the pyrolysis of sewage sludge was proposed. The model was employed to estimate the pyrolysis time for each operating condition, obtaining a proper agreement with the experimental measurements.
The second experimental campaign investigated the effect of different parameters on the pyrolysis of C. cardunculus L. by conducting experiments using the aforementioned innovative measuring technique based on a precision scale, capable of measuring the time evolution of the biomass samples mass during their thermochemical conversion process, while they are moving freely inside a fluidized bed. A silica sand bed reactor, in this case with a particle size, dbm, selected in the range of 180 – 600 μm, operated under different values of excess gas velocity (U/Umf = -2, 0, 3.5, 7, 10.5 cm/s) and reactor temperature (T = 450, 550, and 650°C), was employed to hold the pyrolysis reaction of cardoon particles of three different size ranges. The pyrolysis was accelerated for higher excess gas velocities, obtaining pyrolysis times as short as 17 s for experiments conducted under bubbling fluidized bed regimes, compared to 186 s required to complete the pyrolysis of the same sample in a fixed bed configuration. Similarly, the effect of increasing the reactor temperature promoted faster heating rates across the fuel samples, especially under fixed bed configurations, for which the pyrolysis time is reduced from 322 s to 132 s when increasing the bed temperature from 450 to 650 °C. Regarding the biomass particle size, small sizes are recommended to minimize the conduction thermal resistance inside the fuel particles and, thus, reduce pyrolysis times and increase volatile yields for the pyrolysis in a bubbling fluidized bed reactor. The opposite effect was found for non-bubbling bed operations, where reducing particle size resulted in a decrease of the reaction rate. Under these configurations of the bed, U-Umf ≤ 0 cm/s, for which the biomass rests stationary over the inert bed material forming a package of biomass particles, the samples compactness depends on their particle size. The compactness of smaller particles tends to be higher than for the large pellets, whose shape enhances the appearance of gaps between the pellets conforming the package of particles, through which the hot inert gas employed can easily percolate and heat up the package of particles. Therefore, the effects of interparticle heat transfer inside the package of small particles is significantly higher, delaying the conversion process and enhancing charring reactions.
The last chapter of the thesis presents an experimental study on pyrolysis of C. cardunculus L. samples in a reactor operated as a fixed and a bubbling fluidized bed. The aim of this study was to evaluate the effect of reactor temperature and bed stage on the pyrolysis products, i.e., solid residue, permanent gases, and liquid product. Further study on the permanent and the liquid fractions was conducted for a deeper analysis of the effect of the mentioned parameters. As in the previous chapter, silica sand with a particle diameter, dbm, of 180-600 μm was employed as bed material due to its inert behaviour. The pyrolysis temperatures tested were 450, 550, and 650 °C; whereas fixed bed and fluidized bed regimes were imposed to the bed of silica sand particles to conduct the experiments by adjusting the gas excess over minimum fluidization velocity, U-Umf, to -2 and 7 cm/s, respectively. The results showed that bed configuration has a strong effect on product yields. A fluidized bed operation induces a decrease in solid residues and gas yields due to the better axial mixing caused by bubbles motion, what results in an enhancement of the heat transfer and the prevention of secondary cracking reactions for the pyrolysis vapors released. Therefore, the liquid production was enlarged when pyrolysis occurred under the bubbling fluidized bed regime for all the bed temperatures tested. An increase of bed temperature for a specific bed stage causes larger conversions, i.e., decreases the solid residues and gas yields. Concerning the liquid yield, the operation of the bed under a bubbling fluidized bed promotes the generation of liquid from condensation of the pyrolysis vapors, as a consequence of the faster heating rate of the biomass particles for this bed regime. Further analysis of the liquid phase revealed that the bio-oil fraction reached a maximum of 33.5 wt.% for a reactor temperature of 450 °C and the bubbling fluidized bed configuration. The permanent gases fraction was also subjected to deeper analysis, determining the variation of its components with temperature and bed stage. CO2 was the main component of the permanent gases for all the temperatures and bed configurations analyzed. The maximum yields obtained for this compound correspond to a temperature of 650 °C for both the fixed bed and the fluidized bed configuration. A lower increase in this compound was observed as temperature was raised from 550 to 650 °C under a fixed bed regime, compared to the increase between 450 and 550 °C. This could be attributed to the cracking reactions of the pyrolysis vapors due to the long residence time and temperature, resulting in a diminution of CO2 in favor of the CO formation. CO was the gas with the second largest yield, attaining maximum values of 9.6 and 3.0 wt.% for a temperature of 650 °C under fixed bed and bubbling fluidized bed configuration, respectively. SO2 increased slightly with temperature at the two bed regimes, obtaining a maximum concentration of 0.36 wt.% for the fixed bed and 0.16 wt.% for the bubbling fluidized bed, operated in both cases at 650 °C. The study also presented the effect of the different operating conditions on the total retention of C, O, and S in the permanent gases. An increase in reactor temperature resulted in larger quantities of C, O, and S in the produced gas for the two bed configurations, being this fact consistent with the aforementioned results, where larger gas yields were attained as temperature increased. The operation of the bed under a bubbling fluidized regime was found to promote the retention of sulfur in the solid and/or liquid yield instead of generating SO2 in the permanent gases, preventing eventual environmental problems derived from the emission of this pollutant. Similarly, lower retention of carbon and oxygen were found when the pyrolysis of biomass occurred under a bubbling fluidized bed regime, since the total gas yield was decreased for this bed configuration.
