· It is often necessary to provide coals to a specification, and
· The coal analysis may be all that is available to estimate the performance of a coal product.
Ultimately it is important to know:
· How well the specification of the blend will predict the performance, and
· If the performance of the component coals is already known, whether the performance measurements are additive in a blend.
If the component coals behave independently during combustion processes, it may be
expected that the performance of the blend can be predicted by summation. However in some cases there are interactions between the coals that may produce unexpected results, favourable or otherwise. It is generally believed that the likelihood of these interactions is greater when the properties of the component coals are more dissimilar.
Milling
Moisture content, CV, HGI and Abrasion Index have been identified as coal quality
parameters associated with mill performance. The effect of moisture on mill heating requirements, and of CV on mill throughput, are straightforward and additivity therefore applies.
HGI is known to be an imprecise predictor of mill capacity, power consumption and product fineness for single coals and the same applies for blends. If predictions are to be made it is better to base them on a weighted HGI rather than on a measurement.
If the performance of the individual coals is known, experience suggests that most blends will perform approximately as the weighted average of the components. This is illustrated in Figure 1 for mill power consumption and seems to work for pairs of very different coals.
Figure 1: Mill Power Consumption for Blends of Coal Pairs
Mill wear rates for blends generally lie somewhere between those of the component coals, but do not appear to be strictly additive5 (Figure 2).
Figure 2: Mill Wear Rates for Blends of Coal Pairs
5) Most of the coals shown had low abrasivity, where the precision of the measurement was poor. One coal was very abrasive and appeared to have a strong influence on the blends.
Flame Stability
This performance characteristic is difficult to measure quantitatively. Flame stability is mainly aided by the rapid release of heat with the volatiles, and is hindered by high moisture levels. The calorific value of the volatiles is difficult to predict because the quantity of volatiles released in a boiler flame is different (normally higher) than is measured in the proximate VM analysis. It is not known whether this process is additive.
Burnout Efficiency
High burnout efficiency is normally expected of high VM coals, that is low Fuel Ratio coals. However it is well known that FR or VM provide only very rough estimates of burnout. For blends the burnout efficiency is mostly not additive; it is not uncommon for a blend to perform worse or better than either of the component coals (Figure 4). It is reasonable to suggest that two coals in a blend do not burn independently because each one influences the environment of temperature and oxygen availability which affects the other one.
Another factor contributing to the erratic burnout behaviour of blends is the fact that they are milled together. Though the gross behaviour of mill power consumption and PF fineness appears to be approximately additive, the particles from the softer coal in a blend will be finer than those of the harder coal; the single coals in Figure 4, and the blended coals overall, were all at the same fineness (70% passing 75 μm).
Figure 3: Burnout Efficiency for Blends of Coal Pairs
Deposit Formation
Slagging and fouling are very difficult to predict for single coals based on coal properties because so many mechanisms are involved. For an ash deposit made up of finely mixed elements of known proportions, the fusibility behaviour is complex because of eutectic behaviour6, but can be predicted using complex mathematical models. However in a real case this is complicated by factors such as:
· The deposit is not homogeneous and includes large particles of different compositions, so that eutectic equilibrium may never be reached,
· The bulk composition of the deposit is not the same as the coal ash composition
because of differences in the stickiness and size of different particles that impinge on the boiler surfaces, giving preferential deposition. It may therefore not be possible to predict this composition.
6) A eutectic is a critical mixture of substances that melt at a lower temperature than other mixtures of slightly different composition. Given the great number of elements in coal ash there are many eutectic combinations.
For a blend, the complexity of the ash composition and distribution is magnified and the possibilities for eutectic interactions are great. Even if the above interactions were not present, the temperature environment is normally modified by adding a second coal. Therefore it is possible to combine two relatively harmless coals and to make a blend that fouls or slags.
However it would be wrong to suggest that all blends cause unexpected deposition problems. In an average case it makes sense to use a trouble-free coal in a blend to upgrade a troublesome one. In spite of the risk of trusting the Ash Fusion Test to predict deposition problems, it should be used to test laboratory samples of blends to try to avoid poor coal combinations.
Electrostatic Precipitation
The most important ash property for ESP collection efficiency is its electrical resistivity. Over a relatively narrow range of resistivity there tends to a marked change in collection efficiency; above this range the efficiency is uniformly poor, while below the range it is generally favourable.
When two coals of different resistivity are blended the resistivity of the blend ash tends to lie between those of the component coals; when the results are plotted on a log scale (Figure 4) the appearance is of approximate additivity. In the Figure the resistivity of the blend is in the middle of the range between those of the parent coals (based on the log scale), but the slippage (ie, emission for a constant ash loading) is closer to that of coal A.
The example given demonstrates a bonus to be obtained by blending these two coals. In another case (Figure 5) where the two blended coals have higher resistivities the blend may turn out worse than anticipated.
These two examples seem to explain most results obtained for blends in ACIRL’s pilot-scale Boiler Simulation Furnace. The collection efficiency is often not additive but the result for the blend lies somewhere in the range between those of the parent coals (not better or worse than both parent coals).
The above does not take into account the effects of coals moisture and ash content. For high resistivity ash coals, moisture in the flue gas tends to lower the ash resistivity and therefore improve collection efficiency, and moisture in coal is additive. Ash content does not impact appreciably on collection efficiency but impacts on emissions. Ash is additive and needs to be considered when predicting emissions.
Figure 4: Impact on ESP Efficiency of Blending two Coals with Low to Moderate
Ash Resistivity
Figure 5: Impact on ESP Efficiency of Blending two Coals with Moderate to High
Ash Resistivity
Sulphur Dioxide Emissions
Sulphur Dioxide emissions are almost proportional to the sulphur content of the coal (% daf) because only a small proportion of the sulphur is absorbed by the ash. Therefore a good estimate of SO2 emissions can be calculated for a blend by calculating the daf sulphur content (Figure 6). It is not clear whether the small proportion absorbed by the ash is additive.
Figure 6: SO2 Emission from Two Blends from Coal Nos. 299 and 240
Emission of Oxides of Nitrogen
The emission of NOx does not correlate with the coal nitrogen content, and other reliable methods have not been developed.
A number of sets of blends have been tested in ACIRL’s pilot-scale Boiler Simulation
Furnace with the results shown in Figure 7. The NOx level of blends certainly cannot be predicted based on that of two parent coals that have very different levels. In these cases the blend seems to behave approximately like one or other of the parent coals. However when the two parent coals have fairly similar emissions, blends made from them do not appear to give any great surprises.
Figure 7: NOx Emissions for Blends of Coal Pairs
CONCLUSIONS
Enormous possibilities for satisfying coal specifications by blending, including the use of optimisation.
There are pitfalls and risks because satisfying a specification does not guarantee satisfactory utilisation performance. Nevertheless blending to a specification is a necessary precursor to combustion trials of blends that look promising. When blends are planned between two coals with very different properties, it is advisable to allow for trials at the pilot-scale or full-scale.