Agglomeration in fluidised bed gasification of biomass

doi:10.1016/j.powtec.2007.05.022

Powder Technology

Volume 181, Issue 3, 12 February 2008, Pages 307-320

https://doi.org/10.1016/j.powtec.2007.05.022 Get rights and content

Abstract

Three promising biomass fuels for southern Mediterranean regions were tested for their agglomeration tendency in an atmospheric lab-scale fluidised bed (FB) gasifier using quartz and olivine as bed materials. The defluidisation temperatures of the energy crops Giant Reed (Arundo donax L.) and Sweet Sorghum bagasse were respectively approx. 790°C and 810 °C, in both bed materials, while the agro industrial residue olive bagasse caused defluidisation of the quartz bed at 830°C and olivine bed at > 850 °C. Agglomerates from these tests were analysed with SEM/EDS. Coatings and necks between bed particles were formed due to ash derived potassium silicate melt. For the first two fuels cluster-type agglomerates around remains of char particles were observed. Thermodynamic equilibrium simulations of each chemical system were performed to cross examine the predicted ash melting temperatures and chemistry with experimental findings. Predictions of potassium liquid compounds, like K₂O·SiO_2(l) were verified by EDS analyses on the particle coatings. FB gasification of olive bagasse resisted defluidisation up to higher temperatures because of its lower potassium and higher calcium content, especially in the case of olivine bed. The latter experimental finding coincided with thermodynamic predictions.

Graphical abstract

Three biomass fuels were tested for their agglomeration tendency in an atmospheric lab-scale fluidised bed (FB) gasifier using quartz and olivine as bed materials. Agglomerates were analysed with SEM/EDS. Coatings and necks were formed due to ash derived potassium silicate melt. Thermodynamic equilibrium simulations were performed to cross examine the predicted ash melting temperatures and chemistry with experimental findings.

Introduction

Biomass fuels such as agricultural or agro industrial residues together with energy crops are considered promising renewable energy sources [1]. To reduce CO₂ emissions, part of the power production can be substituted by similar thermochemical technologies, such as gasification, where a solid fuel is converted into a gaseous one referred to as product gas, allowing its use more efficiently in combined power cycles. Nevertheless, biomass gasification suffers from some technical problems prohibiting the economical and trouble free operation of numerous power stations conceptually based on its advantages.

One of the most commonly studied and applied biomass gasification technology involves processing the solid fuel in a fluidised bed (FB) reactor. FB gasifiers are considered a suitable solution for biomass throughputs above 5 MW_th [2]. During the past decade several FB biomass gasifiers were reported in design phase or already operating on a demonstration or (semi-) commercial scale worldwide [3], [4]. FBs are relatively fuel-flexible concerning the feedstock's particle size and moisture content but suffer from two major problems: a generic problem associated with increased tar content in the product gas that inhibits its smooth utilisation and a fuel-specific problem, which is examined in the present study, deriving from the low melting temperatures of biomass ashes that create particle sintering, agglomeration, and eventually defluidisation of the FB.

Biomass fuels, especially those stemming from herbaceous plants, contain silicon, potassium, sodium and alkali earth metals as principal ash forming constituents, together with chlorine and sulphur to a lesser extent [5], [6], [7]. The formation of low melting ash derived compounds such as alkali silicates creates problems in FB reactors at high temperatures; the formation of sticky glassy melt causes bed particle agglomeration. The growth and accumulation of agglomerates may lead to a loss of fluidisation (defluidisation) and unscheduled shutdowns. The problem is common to both combustors and gasifiers operating at 750°C–900 °C and its prediction and tackling requires a better understanding of its mechanisms.

The objective of the present work is to study the agglomeration phenomena of three promising biomass fuels in two commonly used bed materials i) quartz sand and ii) olivine, which is a natural material active in reducing the tar content of the product gas. The fuels chosen were two energy crops, namely a) Giant Reed, and b) Sweet Sorghum bagasse, as well as an agro industrial residue: c) olive bagasse. These represent some of the most promising solid biofuels in southern Mediterranean regions. For all the fuel and bed material combinations, externally induced fluidisation loss experiments were conducted in a lab-scale fluidised bed gasifier. Coatings and necks of melt between agglomerated particles were analysed with SEM/EDS. These chemical analyses were compared with chemical equilibrium modelling predictions of the developed chemical system in the FB (ash and gasification atmosphere).

Section snippets

Bed sintering in FB

The early work of Gluckmann et al., cited by Kunii and Levenspiel [8], experimentally demonstrated that the minimum fluidisation velocity will not follow Ergun's equation but increase sharply with temperature above a certain value which they named ‘initial sintering temperature’. Comparatively they observed that addition of a liquid (viscous phase) into a gas fluidised bed led to defluidisation by making particles more cohesive.

At elevated temperatures, Seville et al., [9], described sintering

Experimental set up

A schematic of the experimental atmospheric FB gasifier used for the defluidisation experiments is shown in Fig. 1. The reactor is a stainless steel cylindrical tube of 8.9 cm ID and 1.3 m height, placed in an electrically heated oven with radiative electric resistances applied to preheat and make up for heat losses in the bed. The gasification/fluidising air was preheated and introduced into the reactor through a perforated plate type distributor. The bed consisted of 2000 g material, quartz

Equilibrium calculations for evaluation of the ash melting behaviour

An effort was made to compare the experimental results on defluidisation temperature and ash coating compositions with predictions from chemical equilibrium modelling for each fuels' ash in the presence of product gas and significant excess of bed material. For this purpose the equilibrium module of FactSAGE™ software was used, based on the Gibbs' free energy minimisation using the SGTE thermodynamic database [46].

The main solution phases used in the calculations are shown in Table 5. Two kinds

Defluidisation temperatures

The defluidisation temperatures determined at the onset of ΔP_bed drop had an accuracy of ± 3 °C. The results, the mean values of which are shown in Table 6, reveal that Giant Reed had the highest agglomeration tendency followed by Sweet Sorghum bagasse and lastly olive bagasse. The defluidisation temperature in olive bagasse gasification is even higher when olivine bed material is used. This can be attributed to the lower potassium and silicon and higher calcium and iron in its ash, in

Conclusions

The high potassium content of Giant Reed and Sweet Sorghum bagasse causes formation of melt responsible for total defluidisation in the FB agglomeration trials at relatively low temperatures. The main defluidisation mechanism is total melting of the silicate ash forming a highly viscous liquid. Melts consist mainly of alkali silicates and to a lesser extent of other oxides or alkali salts. Bed particle grains were either adhered to the sticky melt or necked together by a molten layer on their

Acknowledgements

The authors would like to thank Prof. E. Pavlidou of the Scanning Electron Laboratory of the Physics Department at the Aristotle University of Thessaloniki, and Prof. Ch. Katagas of the Geology Department at the University of Patra for their assistance in SEM/EDS imaging.