Disruptive substances and the burning ­behaviour of solid alternative fuels

Due to the continuing fluctuations in the price of energy and the debate on sustainable conservation of resources, the German cement industry has managed to decrease its overall energy demand continuously through technical optimization [1]. It has also increased the use of alternative fuels still further under pressure from German policy with the implementation of the ban on landfilling untreated waste (Technical Instructions) in 2005.

After the first oil crisis there was initially a changeover from oil to coal and, because of their high energy contents and relative ease of handling, also to waste oil and used tyres. These were followed later by solvents, fullers’s earth, and oil sludges. Solid alternative fuels that resembled lignite and were treated separately from industrial wastes from specific production processes were introduced via the main burner or calciner. When it was found that this minor use of separately treated wastes as solid alternative fuels did not affect either the plant’s immission balance or the process technology and product quality the waste were also allowed to be treated in mixed manner. The range of alternative fuels was completed later by pretreated high calorific fractions out of municipal waste, sewage sludge and bone meal [2]. By 2008 the  proportion of secondary fuels of the overall thermal demand in the German cement industry had increased up to an average of 54.4  % (Fig. 1).

It became obvious that particles of inhomogeneous fuel mixtures followed different trajectories, especially when introduced via the main burner, and also burnt differently. Thin plastic film with a large surface area is consumed while it is suspended in the flame.  But the residual coke of particulate matter forms the tip of the flame or even passes through it and leads to clinker burning under reducing conditions.

Through phenomenological tests and analytical classifying of different fuel mixtures it is possible to obtain information about the flight behaviour and the quality of the treating, the  starting waste composition and the suitability of the subsequent fuel mixture for the intended burning position. The results obtained provide indications of the subsequent combustion behaviour, although, of course, the geometry, density and particle shape can change dynamically and aggravate the negative factors. However, according to Zelkowski [3] such mixtures still always showed burnout times of between 1.5 and 3 s.

With increasing substitution rates the speed and completeness of the cycles of drying, pyrolysis, ignition and burnout affect the diffusion-controlled combustion process, and the former ­methods of comminution come up increasingly against their limits.

Sewage sludge, fuller’s earth and other major single-substance waste streams will not be dealt with here. The focus will be on the production of solid alternative fuels the components of which come from the disposal of wastes from industry or the municipal sector and are treated in appropriate mechanical treatment plants (e.  g. mechanical-biological treating plant, MBT) [2].

Carpets, textiles, wood, cardboard, paper, rubber, plastic film, labels, packaging and other items of daily life are obtained individually or as mixtures at the waste producer. These are checked analytically for their suitability and are jointly separated into disruptive substances and a fraction with enhanced calorific value by preliminary size reduction, screening, ferrous/nonferrous or ballistic separation and NIR spectroscopy. This is carried out with varying success because of adhering biomass, sticky moisture, and interlocking or overlapping pieces [2].

Secondary post treating specifically customized – e.  g. by air classification, screening and comminution – is therefore carried out in most cases to produce kiln-ready fuel of the required particle size. After buffer storage and homogenization the fuel is delivered equipped with the accompanying documents [4].

The entire process chain from the waste producer to usage in the rotary cement kiln is accompanied by quality checks – either voluntary or imposed by authorization requirements. About 10 years ago the SRF-producers, the permitting and monitoring authorities and the end users in Germany developed ¹ and implemented standard procedures for sampling, for chemical digestion and analysis, for data evaluation and for comparison with the permittive conditions.

First of all an unambiguous nomenclature was proposed for describing the differences in the production of solid alternative fuels (Table 1), particularly as solid alternative fuels that are intended to be introduced through the kiln inlet or the calciner are processed differently from those that are to be introduced later through the main burner. In accordance with this nomenclature the fuels were to be described unambiguously with respect to source, type and intensity of treating and quality assurance. Solid alternative fuels for introduction through the kiln inlet or calciner are usually treated more coarsely and are therefore submitted to a different quality procedure from those that have to be introduced through the main burner. These have to be comminuted more finely to achieve better burnout, although compared with pulverized coal there is still a difference of three orders of magnitude.

In most cases the bulk density is reduced to such an extent that for cost reasons it is not possible to cover long ­distances with bulk transporters. After delivery and storage the fuels are loosened up, metered gravimetrically or volumetrically and transported to the respective burning positions. They are ­pneumatically conveyed and accelerated in the same way as the pulverized coal. They leave the burner mouth and take up their trajectories while undergoing thermal conversion.

It is obvious that these inhomogeneous mixtures will split up into individual particles again. The particles will then follow, and burn on, different trajectories. Thin particles with large surface areas (mainly packing residues with thicknesses <  1  mm) are consumed while they are suspended in the flame but small three-dimensional particles (e.  g. hard plastics, rubber, wood, etc.) form the tip of the flame or even pass through it and land on the kiln feed, causing clinker burning under reducing conditions (Fig. 2).

Phenomenological tests (Fig. 3), in which the solid alternative fuels were accelerated with a conveying air flow that is normal for the main burner, were carried out in order to gain a feel for the trajectories and flight characteristics of solid ­alternative ­fuels in a rotary kiln. As would be expected, particles with “comparable properties” collect at their respective distances from the mouth of the burner in “groups with the same characteristics”; this phenomenon corresponds to the principle of air classification.

In traditional classification there are forces acting on the particles in which there is a proportional relationship between the equivalent sphere diameter of the rotational solid, the density and the particle shape and the velocity of descent in air.

This means in practice that if the particles

1) have uniform density and particle shape then separation occurs on the basis of particle size:

Vs ~ 2ÎØ_particle

2) have uniform particle size and particle shape then separation occurs on the basis of density:

Vs ~ 2Îd_particle

3) have uniform particle density and particle size then the separation occurs on the basis of particle shape:

Vs ~ 2Î 1cW

where:

v_s: velocity of descent m/s

Ø_particle: particle diameter m

d_particle: density of particle g/cm³

cW: resistance coefficient of a rotationally symmetrical equivalent body

Classification can be carried out in laminar “Stokes” air flow, in the turbulent Newtonian range or in the transition range. The choice of flow range and the type of classifier (Fig. 4) depend on the objective.

In reality, solid alternative fuels consist of a mixture of many materials that can be easily classified by classification in the laminar or turbulent range. The loading of the classifier should not be greater than 1  % for the analysis to prevent mutual ­interference of the particles.

The first heavy fraction obtained (at approx. 10 m/s) is retained and weighed, while the entire light fraction that has been ­extracted is classified again at the next slowest wind velocity (9 m/s). This procedure is repeated until the possible classifi­cation range has been covered or the feed material has been used up.

This results in a typical histogram and a fractionation (Fig. 5), analogous to the familiar particle size determination, that can be used for further analysis and description of the combined fuel by the d₅₀ value and the slope of the cumulative curve between d₂₀ and d₈₀ (Fig. 6) [2].

This method provides information about the flight behaviour and about the quality of the processing, the former composition of the waste (single components are clearly identifiable) and the suitability of the respective fractions for the intended feed positions in the rotary kiln plant.

During the production and use of solid alternative fuels with components that may have widely differing particle shapes, particle densities and particle sizes it was found that separation based on the same “property” is entirely adequate for thermal conversion. This means that large two-dimensional fuel particles and small, very dense, three dimensional fuel particles collect in a fraction that ultimately exhibits similar combustion properties. Assessment by descent behaviour is therefore more suitable than the typical assessment by particle size class!

The burnout times and the ignition behaviour in a downpipe reactor at constant oxygen content were determined for such mixtures by Zelkowski [3]. The tests showed burnout times for solid, shredded, alternative fuels (25 mm diameter) of between 1.5 and 3 s.

The trajectories of the alternative fuel particles can then in fact be calculated in advance, but the conversion of the fuel particles during the combustion causes a dynamic change in properties for the entire system.

Every combustion process is controlled by diffusion, i.  e. the ­sequences are determined by the particle size of the fuel, the ease with which the resulting gases can be ignited and the burnout time of the residual coke [3, 5].

This means that a gas that ignites readily reacts immediately, while a liquid fuel must first be transformed into the vapour phase. With solid fuels the first stage is drying while the volatile, readily ignitable, constituents are driven out by the radiant heat and can then ignite.

Two-stage combustion, which is characterized by gas burnout followed by burnout of the residual coke, therefore also occurs with solid fuels (Fig. 7).

During the combustion, redox reactions take place between the hydrogen, carbon and oxygen with the release of thermal ­energy. As soon as an initial quantity of fuel had been converted the heat activates the intact (subsequent) fuel, which is dried and pyrolyzed. Its gases are ignited and burn out until one of the reaction partners has been used up. This cycle is repeated until the fuel or oxygen is consumed.

With increasing rate of substitution of pulverized coal by solid alternative fuels the speed and completeness of this sequence ultimately affect the entire burning process. Because of the high H/C ratio of liquid fuels the cycles of drying, pyrolysis, ignition and burnout are passed through more rapidly, and the time-intensive burnout of the residual coke hardly plays any part in the conversion process [6].

The starting components for solid alternative fuels are, as a rule, former objects of everyday life, i.e. there is a wide spectrum from plastics that contain hydrocarbons to carbon-rich paper or wood and a correspondingly wide spectrum of different burning characteristics [7, 8]. Solid alternative fuels are therefore located in the van-Krevelen diagram at a molar H/C ratio between 0 and 1 and a molar O/C ratio between 0.5 and 0.6 (Fig. 8).

The insulating effect, the geometry and other surface effects mean that the diffusion in the fuel particles and the rates of combustion take place at differing speeds. Mechanically, the diffusion path can therefore be shortened, meaning the surface of the fuel particles can be enlarged, by producing even finer particle sizes in high output comminution equipment. “Edge running or pelletizing” (comminution by pressing through special ring or flat matrices) and “shredding” (cutting, chopping and passing through a sieve tray) have now become established. However, the desire for increasing substitution means that the previous methods of comminution are increasingly coming up against their limits.

For this reason further development is on the way for utilisation of solid alternative fuels. For many years efforts were focused on the air-whirl-mill. Various producers have been ­working with this special impact mill technology for over 50 years, but so far it has not been adapted for use with secondary fuels.

The comminution principle of this mill is based on a rotor with a high circumferential speed and a high air throughput, in which a high degree of turbulence is imparted to the air flow in the grinding zone (Fig. 9). The trajectories and velocities of the particles change very rapidly within this vortex and the collisions between the grinding elements, the wall and the mill feed lead to comminution of the particles (Figs. 10 –12).

The particles remain in suspension during the entire grinding process until they are discharged and then collected. The enormous increase in surface area and the high air throughputs mean that this principle is also suitable for simultaneous drying. This effect can be significantly increased by the application of waste heat.

Initial investigations in the IEVB (Institute for Energy Technology and Fuel Technology) at Clausthal University of Technology various pulverized alternative fuels showed astonishing rates of ignition and conversion similar to those achieved with pulverized coal and lignite (Fig. 13). The ignition temperatures (IT) of the pulverized alternative fuels (AF) tested, which came from widely differing treating plants (MBT) = EBS 03/08 and industrial waste = EBS 03/07), clustered between 680 °C and 711 °C in the transition region between pulverized lignite with an IT of 620 °C and pulverized coal with an IT of 760 °C. This means that the pulverized alternative fuels exhibited an ignition behaviour that is slightly better than pulverized coal and somewhat more sluggish than pulverized lignite.

The test results achieved so far have shown that the principle of the air-whirl-mill, which was originally designed for grinding foodstuffs and other substances, can also be used successfully for grinding solid alternative fuels (Fig. 14). Depending on the grinding resistance, size reduction rates of up to 100:1 are possible. It is not essential that hard pellets are used as only the particles that are three-dimensional, hard or brittle and in the past had led to problems in the kiln feed are rapidly and effectively comminuted to particle sizes significantly <  1  mm, while thinner plastic film and other two-dimensional particles hardly experience any comminution.In fact there is no need regarding the burnout behaviour of these particles as what matters in thermal conversion is the speed of the sequence of drying, pyrolysis, ignition and burnout.

Pulverized alternative fuels have a tendency to bridging and agglomeration, so long static phases (intermediate storage, etc.) should be avoided. Worth to mention the additional drying effect results by the increase in surface and the excess air, as well as the halving of the initial bulk densities of some pulverized alternative fuels, although doubling was also observed with some biomasses.

The chemical properties of the alternative fuel components are not altered, which means that with residue-derived alternative fuels the known energy exchange ratio, often caused by fairly high ash and water inputs, continues to apply, i.  e. a slightly higher mass input of solid alternative fuels is needed to cover the thermal energy requirement.

As far as the flight and ignition behaviour is concerned the physical properties of the pulverized alternative fuels are now comparable with those of pulverized lignite. On the basis of many years of experience with cement plant construction Polysius AG can now offer its customers a complete handling and burner system for solid alternative fuels throughout the world.

I would particularly like to thank Professors R. Scholz and R. Weber from the IEVB at Clausthal University of Technology and my colleagues at Polysius AG, Beckum, for their year-long monitoring of the work and their friendly support. I would also like to thank Bückmann GmbH & Co. KG from Mönchen­gladbach, Altenburger Maschinen Jäckering GmbH from Hamm and Mahltechnik Görgens from Dormagen for carrying out the pilot plant trials and the tests.