Fluidized Bed Combustion of Solid Lignin-Rich Residues from Bioethanol Production

Authors: 
Solimene, R., Consiglio Nazionale delle Ricerche
Troiano, M., Università degli Studi di Napoli Federico II
Cammarota, A., Consiglio Nazionale delle Ricerche
Chirone, R., Consiglio Nazionale delle Ricerche
Salatino, P., Università degli Studi di Napoli Federico II
The behaviour of lignin-rich residues from second-generation bioethanol production plant has been investigated during fluidized bed combustion. With the depletion of fossil fuels and the issues related to the emissions of greenhouse gases, the utilization of alternative automotive fuels has attracted worldwide attention. Biofuels are unique among available alternative energy sources in their compatibility with existing liquid transport fuel. The global production and use of biofuels have increased from 18.2 billion liters in 2000 to 60.6 billion liters in 2007, with about 85% of this being bioethanol. The production of second-generation bioethanol, i.e. ethanol produced from lignocellulosic biomass and scrap of agricultural crops, will be of 155 billion liters by 2020. The composition of lignocellulosic biomass is typically: cellulose (35-45%), hemicellulose (25-30%) and lignin (25-30%). Differently from cellulose and hemicellulose, lignin is a polymer mainly consisting of not fermentable phenylpropanoids. Therefore, the solid residues after ethanol distillation and separation in the second-generation bioethanol production, are characterized by high lignin content. These lignin-rich residues can be used to produce chemicals as well as energy. Fluidized bed combustion/co-combustion can be considered as promising and viable options to recover thermal power from lignin-rich residues of a second-generation bioethanol production plant. However, the understanding of the mechanisms of devolatilization, conversion and attrition of particles of lignin-rich residue in fluidized beds is scarce and it deserves further investigations.

This work aims at investigating the combustion behaviour of a lignin-rich residue in lab-scale fluidized beds. In particular, devolatilization, char burnout and comminution phenomena of single fuel particles were studied. Devolatilization experiments aimed at characterizing the devolatilization rate, the average volatiles molecular weight and the devolatilization time of wet fuel particles while varying particle size and operating temperature based on the analysis of time-resolved pressure signals. Fragmentation and char burnout experiments aimed at evaluating the particles fragmentation tendency during the devolatilization (primary fragmentation) and char burnout (secondary or percolative fragmentation) steps. Elutriation by attrition experiments were carried out to estimate the carbon elutriation rate under devolatilization and combustion conditions.

A stainless-steel bubbling fluidized bed reactor 17 mm ID and 1 m high was used for the devolatilization experiments. The reactor consists of a lower section, an upper section and an inter-flange of connection between the two logs. The inter-flange length contains most of the bed material. A thermocouple is inserted in the fluidized bed at about 0.01 m above the high-pressure-drop porous gas distributor plate. A stainless-steel tube is laterally welded to the column for rapid discharge of bed particles. The lower section is used as wind box. Technical nitrogen from cylinders was used as fluidizing gas. The fluidizing gas-flow rate was established by setting the upstream pressure of a critical orifice located along the gas feeding line. The orifice (0.187 mm dia.) was operated in critical flow conditions. The upstream relative pressure varied between 1 and 3.5 bar. The reactor was operated at a pressure slightly larger than atmospheric pressure by means of a calibrated flow restriction at the exhaust consisting of an orifice with an adjustable cross section. The reactor was heated by means of a cylindrical oven (1 kW) 0.4 m high equipped with a PID temperature control system. A high-precision (accuracy better than 0.06 mbar) piezo-resistive electronic pressure transducer with small-time response (less than 1 ms) and 100 mbar full-scale was used to measure the time-resolved gas pressure in the reactor. The pressure signal was acquired by means of a data acquisition unit consisting of an A/D converter and a personal computer.

Fragmentation, elutriation and combustion experiments were carried out in a stainless steel atmospheric bubbling fluidized bed combustor 41 mm ID and 1 m operated at 850°C. Three different configurations of the reactor were used. In the first configuration, used for particle fragmentation experiments, the top section of the fluidization column was left open to the atmosphere. A stainless steel circular basket was inserted from the top to retrieve fragmented and un-fragmented particles from the bed. In the second configuration, used for fines elutriation rate experiments, a two-exit head was fitted to the top flange of the fluidization column. Flue gases were conveyed alternately to two removable filters made of sintered porous alumina. Batches of material were into the bed via a hopper connected sideways to the upper part of the freeboard. A paramagnetic analyzer and two NDIR analyzers were used for online measurement of O2, CO, and CO2 concentrations, respectively, in the exhaust gases. The third configuration, used for single particle combustion experiments, also consisted in leaving open the top section of the fluidization column. A stainless-steel probe was inserted from the top of the column to convey a fraction of the exit gases directly to the gas analyzers. Data from the analyzers were logged and further processed on a PC.

Experiments were carried out using a lignin-rich residue coming from a second-generation bioethanol production plant (LHV (as received) =5645 kJ/kg, moisture 60.7 wt%, volatile matter 25%, fixed carbon 8.1%, ash 6.1%). The lignin content was obtained by Klason method. For devolatilization experiments, fuel particles with a diameter from 3 mm to 8 mm were used, while for fragmentation/combustion/elutriation experiments fuel particles sized between 10 and 17 mm were used. Gases used in the experiments were air and nitrogen of technical grade.

Regarding devolatilization experiments, pressure was continuously monitored in the fluidized-bed reactor during the experiments. The selection of the steady-state overpressure in the reactor, dictated by the flow rate of fluidizing gas and by the calibrated flow restriction at the exhaust, determines the time evolution of the phenomena. An overpressure of about 10 mbar was chosen for the tests at 750°C and 850°C, and an overpressure of 20 mbar at 650°C. The bed of sand (400-600 mm, 13 g) was fluidized with nitrogen to rule out the influence of volatile matter and char combustion on the recorded pressure profiles. The upstream pressure at the critical orifice was kept at a constant value of about 1, 1.65 and 2.3 bar at 650, 750 and 850°C, respectively. The mass of each batch of fuel particles was fixed at 100 mg for the sizes 3-5 mm, 135, 160 and 200 mg for 6 mm, 7 mm and 8 mm-sized particles. The tests consisted of the injection of a batch of fuel particles in the fluidized bed followed by complete devolatilization. The pressure signal was continuously logged during the experiments at a sampling rate of 1,000 Hz. Pressure time series were analyzed to determine the volatile emission rate as a function of time. The time for 95% devolatilization degree and the mean molecular weight of the emitted volatile compounds were computed from the raw data. Primary fragmentation experiments were performed using the basket-equipped configuration. Single fuel particles were injected into the bed from the top of the column. After about 3 min, required to completely devolatilize the fuel, the resulting char was retrieved by means of the basket to investigate the number and size of the produced fragments. For secondary fragmentation experiments, single fuel particles were first pyrolyzed under minimum fluidization condition, then, nitrogen–oxygen mixtures (4.5% of O2 volume concentration) were used as inlet gas. Experiments were carried out at a superficial gas velocity of 0.4 and 0.8 m/s. Char particles were injected into the bed from the top of the column. The char particles were retrieved from the bed at definite time intervals and the number and size of the fragments were recorded. The experimental run was stopped when the char combustion was complete as detected by the time-resolved profiles of concentration of gaseous species measured at the exhaust. Regarding the attrition rate experiments, elutriated fines were collected by means of the two-exit head by letting the flue gas flow alternately through sequences of filters for definite time intervals. Fines collected in the filters were further analyzed to determine their fixed carbon content. This procedure enabled time-resolved measurement of carbon elutriation rates. Furthermore, combustion experiments were carried out at 850°C by injecting single fuel char particles with a gas superficial velocity of 0.4 and 0.8 m/s. Continuous analysis of the exhaust gases enabled measurement of the carbon combustion rate as a function of time until complete burnout was reached.

Results highlighted that devolatilization rate and times were relatively long and increased with decreasing temperatures. Fuel particles did not undergo primary fragmentation, whereas secondary and not percolative fragmentation occurred during the late stage of char burnout. Attrition by surface wear was limited. However, results indicated that carbon loss by attrition increased with the superficial gas velocity and it was more pronounced under inert conditions than under combustion environment.