Filip Johnsson

Characterization of fluidization regimes by time-series analysis of pressure fluctuations

Abstract This work compares time, frequency and state-space analyses of pressure measurements from fluidized beds. The experiments were carried out in a circulating fluidized bed, operated under ambient conditions and under different fluidization regimes. Interpretation of results in time domain, such as standard deviation of the pressure fluctuations, may lead to erroneous conclusions about the flow regime. The results from the frequency domain (power spectra) and state-space analyses (correlation dimension, D ML , and Kolmogorov entropy, K ML , together with a non-linearity test) of the pressure fluctuations are generally in agreement and can be used complementary to each other. The power spectra can be divided into three regions, a region corresponding to the macro-structure (due to the bubble flow) and, at higher frequencies, two regions representing finer structures that are not predominantly governed by the macro structure of the flow. In all fluidization regimes, the measured pressure fluctuations exhibited an intermittent structure, which is not revealed by power spectral analysis of the original signals. Fluctuations with pronounced peaks in the power spectrum and in the auto-correlation function, corresponding to passage of single bubbles through the bed, are non-linear with a low dimension ( D ML D ML D ML >5.5 both K ML (bits/cycle) and D ML are insensitive to changes in the distribution of energy in power spectra. Thus, the state-space analysis reflects that non-linearity is mostly found in the macro-structure of the flow. Fluidized bed time series treated in this work are available at http://www.entek.chalmers.se/∼fijo

Non-intrusive determination of bubble and slug length scales in fluidized beds by decomposition of the power spectral density of pressure time series

Abstract In this paper we show that spectral analysis of non-intrusive time dependent pressure measurements in bubbling and circulating gas–solid fluidized beds permits to obtain the first estimates of bubble, gas slug, and solids cluster length scales from pressure fluctuation data. These length scales are calculated from the incoherent cross power spectra of pressure signals measured in the bubbling or circulating bed and in the plenum. Remarkable quantitative agreement with bubble diameter data is found, and equally remarkable agreement is obtained with independent estimates of gas slug lengths by others in circulating fluidized beds. These results demonstrate the possibility of greatly expanding the information that can be obtained non-intrusively from gas–solid fluidized beds.

Radial voidage profiles in fast fluidized beds of different diameters

A previously developed method for calibrating an optical-fibre probe for voidage measurement in gas-solid systems has been improved. This was applied to study the influence of operating conditions, particle properties and bed diameters on radial voidage profiles in circulating fluidized beds operating in the fast regime. The present experimental results showed that the radial voidage profile depends solely on the value of the cross-section-averaged voidage, irrespective of operating conditions, solids properties and bed diameters, and can be correlated by the expression epsilon = epsilon-BAR(0.191 + phi-2.5 + 3-phi-11, where epsilon and epsilon-BAR represent the local and cross-section-averaged voidage respectively, and phi-denotes the dimensionless radial distance from the centre of the bed.

Fluidization regimes in non-slugging fluidized beds : the influence of pressure drop across the air distributor

The purpose of the present work is to study the influence of the pressure drop across the air distributor on the bubbling conditions of the bottom bed of circulating fluidized beds (CFB). The bottom bed is the dense bubbling zone just above the distributor. The experimental work was carried out in a 12 MWm CFB boiler and in a cold CFR Three different distributions of the bubble flow in time and space, termed fluidization regimes, were identified in the cold CFB: the multiple bubble regime with many small bubbles evenly distributed in the bed; the single bubble regime, characterized by the presence of only one bubble at a time in the bed; and the exploding bubble regime with large, single, irregular voids stretching from the air distributor to the bed surface. These bubbling conditions were observed during variations in the gas velocity and the distributor pressure drop. A comparison with the 2 mz cross-section CFB boiler showed that the boiler always operates in the single or in the exploding bubble regime, which indicates a bubble flow that is not continuous and not well distributed over the crosssection of the bed. The conditions in the boiler are influenced by the relatively large area of gas passage and the low pressure drop of the boiler air distributor.

Expansion of a freely bubbling fluidized bed

Abstract The expansion of a freely bubbling fluidized bed is studied over a range of particle properties and gas velocities that applies to fluidized bed boilers. A bed expansion model is derived from a modified two-phase flow model. The results from the model are compared with measurements in both a cold two-dimensional bed and a 16 MW th fluidized bed boiler, as well as with data found in the literature. The model represents experimental results for sand particles of a diameter ranging from 0.15 mm to 4.0 mm and with gas velocities up to 3 m s −1 .

Combustion of wood particles—a particle model for eulerian calculations

A simplified model for the combustion of solid fuel particles is derived, relevant for particle sizes and shapes used in fluidized and fixed-bed combustors and gasifiers. The model operates with a small number of variables and treats the most essential features of the conversion of solid fuel particles, such as temperature gradients inside the particle, the release of volatiles, shrinkage, and swelling. Typical shapes (spheres, finite cylinders, and parallelepipeds) are also considered. The model treats the particle in one dimension, and to describe the conversion inside a fuel particle the model only needs the transfer of heat and mass to an element of its external surface. When modeling a large combustion system, it is a great advantage that the conversion is related to the external surface, because the model does not have to be limited to just a single particle. In fact, it can handle the conversion of a solid phase in a computational cell, where the conversion is related to surface area per unit volume, instead of the surface area of a single particle. The model divides the particle into four layers: moist (virgin) wood, dry wood, char residue, and ash. The development of these layers is computed as function of time. The model shows satisfactory agreement with measurements performed on more than 60 samples of particles of different sizes, wood species, and moisture contents. Comparison with the experiments shows that the simplifications made do not significantly influence the overall accuracy of the model. The model also demonstrates the great influence of shrinkage on the times of devolatilization and char combustion.

Fluidization regimes and transitions from fixed bed to dilute transport flow

Characterization by means of Kolmogorov entropy shows that the dynamics of the bottom bed in small size circulating fluidized bed risers are significantly different from the dynamics of the dense bottom bed in large size risers and, as a consequence, two types of circulating regimes are introduced: the exploding bubble bed for large risers and the circulating ‘slugging’ bed for small risers, the latter at high superficial gas velocities. In a pictorial fluidization diagram ten gas—solid fluidization regimes are given, seven of which are experimentally identified with the Kolmogorov entropy by varying the superficial gas velocity, riser solids holdup and diameter (or width) of the riser: bubbling bed, slugging bed, exploding bubble bed, intermediate turbulent bed, circulating ‘slugging’ bed, intermediate dilute flow, and dilute transport flow. No transition could be identified between the exploding bubble bed at captive conditions and the exploding bubble bed at circulating conditions in the dense bottom bed of the two largest facilities in this study. This suggests that the dense bottom bed in large size risers can be considered as a bubbling bed. A turbulent bed was found in none of the facilities of this study with the Geldart B solids used. As well as by the Kolmogorov entropy (chaos analysis), the hydrodynamics have been characterized by amplitude of pressure fluctuations, while a solids distribution analysis has also been carried out. The study has been made in four (circulating) fluidized beds of different size and design, all operated with 0.30 mm silica sand. The dimensions of the fluidized bed risers are 1.47 × 1.42 × 13.5 m, 0.70 × 0.12 × 8.5 m, 0.12 m i.d. × 5.8 m, and 0.083 m i.d. × 4.0 m.

Fluid-dynamic boundary layers in CFB boilers

The characteristics of the fluid-dynamic boundary layer (FBL) were studied in the 12 MW circulating fluidized bed boiler at Chalmers University of Technology. For a comparison, some additional measurements were made in a 165 MW CFB boiler. The FBL consists mainly of the falling film of particles at the furnace wall. The distance from the wall to the zero net particle flux position is used as a measure of the thickness of the layer. Horizontal profiles of particle flux, velocity and concentration were measured in the transport zone of the furnace. It was found that the FBL thickness varies only slightly or not at all with particle concentration, gas velocity and height in the furnace. An estimation of the thickness is given as a function of the horizontal bed dimension. The particle falling velocity inside the FBL is about 0.9–1.6 m/s depending on location. There is a sharp increase in particle concentration in the FBL all the way towards the wall. The disparity between the particle flux profile normal to a plain refractory-lined wall and the corresponding profile normal to a membrane-tube wall can be compensated for by taking the fin-tube configuration into account. From this point of view, the particle redistribution mechanism at the membrane-tube wall is discussed

Gas leakage measurements in a cold model of an interconnected fluidized bed for chemical-looping combustion

Abstract In chemical-looping combustion (CLC) a gaseous fuel is burnt with inherent separation of the greenhouse gas carbon dioxide. The oxygen is transported from the combustion air to the fuel by means of metal oxide particles acting as oxygen carriers. A CLC system can be designed similar to a circulating fluidized bed, but with the addition of a bubbling fluidized bed on the return side. Thus, the system consists of a riser (fast fluidized bed) acting as the air reactor. This is connected to a cyclone, where the particles and the gas from the air reactor are separated. The particles fall down into a second fluidized bed, the fuel reactor, and are via a fluidized pot-seal transported back into the riser. The gas leaving the air reactor consists of nitrogen and unreacted oxygen, while the reaction products, carbon dioxide and water, come out from the fuel reactor. The water can easily be condensed and removed, and the remaining carbon dioxide can be liquefied for subsequent sequestration. The gas leakage between the reactors must be minimized to prevent the carbon dioxide from being diluted with nitrogen, or to prevent carbon dioxide from leaking to the air reactor decreasing the efficiency of carbon dioxide capture. In this system, the possible gas leakages are: (i) from the fuel reactor to the cyclone and to the pot-seal, (ii) from the cyclone down to the fuel reactor, (iii) from the pot-seal to the fuel reactor. These gas leakages were investigated in a scaled cold model. A typical leakage from the fuel reactor was 2%, i.e. a CO 2 capture efficiency of 98%. No leakage was detected from the cyclone to the fuel reactor. Thus, all product gas from the air reactor leaves the system from the cyclone. A typical leakage from the pot-seal into the fuel reactor was 6%, which corresponds to 0.3% of the total air added to the system, and would give a dilution of the CO 2 produced by approximately 6% air. However, this gas leakage can be avoided by using steam, instead of air, to fluidize the whole, or part of, the pot-seal. The disadvantages of diluting the CO 2 are likely to motivate the use of steam.

Numerical simulation of the fluid dynamics of a freely bubbling fluidized bed: influence of the air supply system

Abstract Numerical simulations, based on an Eulerian approach, of a freely bubbling fluidized bed (BFB) are performed where emphasis is put on the importance of the inlet boundary conditions (influence of the pressure drop of the air distributor on the state of fluidization). The numerical results are compared with local instantaneous pressure measurements and time-averaged measurements (bed height, mean particle concentration). The closure of the Eulerian model is treated as follows: the drift velocity is modelled with a binary dispersion coefficient, gas-phase (continuous phase) fluctuations are modelled with a modified two-equation k1–ϵ1 model, and particle-phase (discrete phase) fluctuations are also described by a two-equation k2–k12 model derived from the kinetic theory of granular flow (modified to account for the interstitial gas) and a Langevin equation. The numerical computations (of a bubbling fluidized bed) predict qualitatively the experimental values, which shows that there is a coupling between the bed and the air supply system.