- Title
- Laboratory evaluation of combustibility of coals for blast furnace pulverised coal injection
- Creator
- Li, Hongyu
- Relation
- University of Newcastle Research Higher Degree Thesis
- Resource Type
- thesis
- Date
- 2014
- Description
- Research Doctorate - Doctor of Philosophy (PhD)
- Description
- The concern about pulverised coal injection (PCI) into ironmaking blast furnaces has increased because of environmental issues and the shortage of coking coal resources. PCI technology incorporates the advantages of saving energy and cost, improving productivity and potentially reducing CO₂ release. However, injecting unsuitable coals affects the operation of blast furnaces negatively. One of the resulting problems is that unburnt char blocks the passage of fluid and gas transportation, therefore changing the distribution of gas and temperature in the blast furnace. Under these circumstances, it has been addressed that how to evaluate coals before being injected into a blast furnace. Direct evaluation of the combustibility of PCI coals in ironmaking blast furnace is difficult. Different test rigs have been used in the assessment, such as a pilot-scale PCI rigs firing at 25-69 kg/h, a laboratory-scale drop tube furnace (DTF) firing at 4 g/h. A PCI rig is designed to replicate the conditions in the blowpipe, tuyere and raceway of a blast furnace. Although a PCI rig can give a prediction of coal burning behaviour, it is costly and complicated and not always available. Alternatively, a DTF, with the advantages of low cost and easy operation, has been commonly used in the research of coal combustion. However, differences in temperature, heating rate and residence time exist in the two rigs. Significant influence on the devolatilisation of coal and combustion of residue char by these parameters has been suggested in the literature. This thesis therefore investigates coal combustion behaviour by burning coals of different ranks in a laboratory-scale DTF with an as-ground particle size distribution used in PCI industry, with some coals which have previously been combusted in a PCI test rig. This work is built on the hypothesis: a DTF can substitute for a PCI rig in the evaluation of PCI coals by ranking coal combustion performance against coal rank as quantified by the volatile matter (VM) content, and can identify coals that perform better or worse than expected. The objectives are specifically to: Compare the combustion performance of a range of coals combusted in a pilot-scale PCI rig and a laboratory-scale DTF. Establish the methodology to find if the ranking of coal combustion performance in PCI can be predicted by DTF tests on the same coals. Identify coals not fitting the ranking of combustion performance in the PCI rig and suggest reasons, including the clarification of combustion process in the PCI rig by a single particle combustion model. Twelve coals have been tested in the DTF (nine of them have corresponding combustion data in the PCI rig). The gas temperature of the DTF was set at 1450 °C, and coal feed rate was about 4 g/h. The yields of VM in devolatilisation and coal burnouts (at 21%, 22.6% and 26% O₂) were measured, respectively. The measured burnouts from the DTF were compared with the data previously obtained from the PCI rig. It was found that a linear trend of burnouts was returned from each rig but their slopes appear to be different. The burnouts of coals in the DTF cover a wider range against the range of VM contents and therefore a steeper slope. Char specific surface area is significantly impacted by combustion conditions, and PCI chars present higher specific surface area than DTF chars across the range of coals. A small amount of char sampled from coal burning at 21% O₂ in the two rigs was re-burned in a thermalgravimetric analyser (TGA) to compare the TGA reactivity at 650 °C. PCI chars present higher apparent reactivity, which was directly measured in the TGA, than DTF chars. However, after being normalised by char specific surface area, their intrinsic reactivities are in the same order, indicating that the different conditions of the two rigs exert little effect on it. Coal and char particle size distribution was measured by a Malvern 2600 particle-size analyser. There is no evidence to suggest that one rig produces consistently finer or coarser chars. The char particle size is a function of the coal particle size but different behaviours were observed from different coals within the two rigs. The char morphology of both surface and cross-section was investigated through a scanning electron microscope (SEM). Small fragments observed in the two low-VM coals in the PCI chars certify that fragmentation occurred during combustion. A single-particle combustion model was developed to predict VM yield in the PCI rig, with allowance of the particle size distribution of fired coals in the model. This model incorporated the char combustion kinetics parameters derived from the TGA measurements. The predictions of the PCI rig from the model were compared with the measured VM yields in devolatilisation in the DTF and the char weight losses as well. The results show that VM yield in the PCI rig is generally higher than that in the DTF. Nevertheless, the Q factors of the coals present different trends in the two rigs. They demonstrate a descending trend in the PCI rig against coal VM contents while increase with increasing VM content in the DTF. Char weight loss shows little difference across the range of coals but is related to particle size in the PCI rig while that returns an increasing trend with a few scattered points in the DTF. The results suggest that the devolatilisation is a significant process during combustion in the PCI rig. The burnout of coals during combustion in the DTF is more sensitive to coal VM content. For the results from medium (18.4% VM db) to high (41.7% VM db) volatile coals, the tests in the DTF can provide a reasonable indication of coal combustion performance in PCI. However, the burnouts of the outliners - two low volatile coals (Coal 2 and Coal 3) - have significantly higher burnouts than expected from their VM contents according to the trend suggested by other coals, yet fit well with the burnout trend in the DTF. The sensitivity analysis of fragmentation suggests that the contribution of enhanced char burnout caused by particle fragmentation to produce finer char is limited. Therefore, higher VM yield as well as greater char burnout contributed to the unusual high burnouts of the two low-VM coals in the PCI rig according to the prediction of the model. Based on the results of this study, it is safely concluded that a low-cost laboratory-scale DTF can substitute for a pilot-scale PCI rig in the evaluation of PCI coals, but testing of low-volatile coals must await an understanding of the reason for evaluation of such coals from the current data.
- Subject
- coal combustion; pulverised coal injection; reactivity
- Identifier
- http://hdl.handle.net/1959.13/1051120
- Identifier
- uon:15249
- Rights
- Copyright 2014 Hongyu Li
- Language
- eng
- Full Text
- Hits: 1531
- Visitors: 1999
- Downloads: 651
| Thumbnail | File | Description | Size | Format | |||
|---|---|---|---|---|---|---|---|
| View Details Download | ATTACHMENT01 | Abstract | 275 KB | Adobe Acrobat PDF | View Details Download | ||
| View Details Download | ATTACHMENT02 | Thesis | 8 MB | Adobe Acrobat PDF | View Details Download |