Coproduction of biofuels and biochar by slow pyrolysis in a rotary kiln
Item statusRestricted Access
Embargo end date31/12/2100
Biochar has been heralded as a promising technology for climate change mitigation that can also benefit soils. Biochar is a carbonaceous solid produced by pyrolysis of biomass – the thermal decomposition of plant and plant-derived matter in the absence of oxygen. When added to soils, biochar has the potential to increase crop yields and suppress soil emissions of greenhouse gases, whilst sequestering carbon in a stable form. In addition to biochar, biomass pyrolysis produces liquids and gases that can serve as biofuels. Biochar production systems that generate excess heat or power are particularly environmentally and economically attractive. Rotary kilns are the favoured process reactor in many industries, given their potential to handle a wide range of feedstocks and provide good process control. This thesis investigates the potential to coproduce biochar and excess biofuels by slow pyrolysis in a pilot-scale rotary kiln. The work attempts to progress towards the ultimate aim of scaling up the rotary kiln and optimising its operating conditions to produce biochar of good quality along with an excess of useful biofuels. Experimental work, involving the development and application of new methodologies, was used to gain a better understanding of the process. The data gathered were then used to support preliminary numerical simulation efforts towards the development of a comprehensive process model. Five biomass feedstocks were considered: softwood pellets, miscanthus straw pellets, wheat straw pellets, oilseed rape straw pellets and raw rice husks. The granular flow of biomass feedstocks was observed in a short closed drum faced with acrylic and resting on rollers. All pelletized feedstocks displayed similar angles of repose, validating the use of softwood pellets as a model biomass for these feedstocks. Bed mixing, which can improve product uniformity, was slow under typical operating conditions, requiring 5 min to complete at 4 rpm for softwood pellets. Mixing quickened considerably at higher rotation rates. A digital image analysis method was developed to measure the distribution of solid residence times inside the rotary kiln. The mean residence time of softwood pellets ranged from 19 to 37 min under typical operating conditions, decreasing with increases in kiln rotation rate, but mostly unaffected by feeding rates. These findings show that kiln rotation rates must be selected to balance the residence time of solids inside the kiln with bed mixing levels. Thermogravimetry and differential scanning calorimetry were performed on samples of ground softwood pellets under five different heating profiles to study the kinetics and heat flows of the pyrolysis process. Both exothermic and endothermic regions were identified, with most reactions taking place between 250°C and 500°C. Results suggest that exothermic pyrolysis reactions can be promoted by altering the process heating rate, thereby improving net biofuel yield from the process. The thermogravimetric data collected was used to develop a distributed activation energy model (DAEM) of the kinetics of softwood pellet pyrolysis for integration into a comprehensive model of the process. The applicability of the kinetic model to large-scale processes was confirmed using a simplified process model developed to simulate biomass pyrolysis inside the pilot-scale rotary kiln. Although crude, the simplified process model produced sufficiently accurate estimates of char yield for preliminary design purposes. The simplified model also allowed important process parameters, such as kiln filling degree, solid residence time and heating rate, to be evaluated. A series of pyrolysis experiments was performed on the pilot-scale rotary kiln to evaluate the yields of biochar and biofuels and determine the temperature profile inside the kiln. This work required the design of a suspended thermocouple system that measures temperatures along the kiln, both in the gas phase and inside the solid bed. For most experiments at 550°C, a region of high temperature gas and solids was observed, possibly indicative of exothermic reactions. Biochar yield varied from 18% to 73% over the range of feedstocks and operating conditions tested. A vapour sampling methodology that relies on the use of a tracer gas was developed to determine the yield of pyrolysis liquids and gases. Due to analytical difficulties, it was not possible to obtain accurate mass closure with this method. However, the methodology revealed significant air ingress into the pilot-scale rotary kiln that is responsible for partially combusting biofuels produced by the process, thereby reducing their calorific value. Energy balances on the kiln confirmed that the calorific content of pyrolysis liquids and gases exceeds the energetic demand of the process, yielding between 0.3 and 11 MJ in excess biofuels per kg of biomass feedstock. An attempt was made to develop a multiphase model of the flow of vapours and solids inside the rotary kiln using computational fluid dynamics (CFD), but the continuous modelling approach was found inadequate to simulate the dense bed of biomass inside the kiln. The discrete element method (DEM) was sought as an alternative to model the granular flow of biomass inside the kiln. Extensive parameter calibration was required to reproduce the experimental behaviour of softwood pellets observed in the short closed drum. A model of the pilot-scale rotary kiln was constructed to simulate particle residence times. Further parameter calibration was required to replicate softwood pellet holdup inside the kiln. The calibrated model was able to reproduce the mean residence time of softwood pellets within 10% under different kiln operating conditions. However, simulated residence time distributions could not be established as a result of the long execution times required for this modelling work. Few data are currently available on large-scale continuous biomass pyrolysis processes; the experimental data gathered in this thesis help to fill this gap. Along with the numerical simulation work presented herein, they provide the foundation for the development of a comprehensive model of biomass pyrolysis in rotary kilns. Such a numerical model would prove invaluable in scaling up the process and maximizing its efficiency. Future work should consider the agronomic value and carbon sequestration potential of biochar produced under different operating conditions. In addition, the performance and efficiency of different conversion technologies for generating heat and power from biofuels need to be investigated.