Blast-furnace instability after a coke lot change: peripheral channelling or hanging?
A 2,500 m³ blast furnace producing 5,200 t/day runs at a coke rate of 340 kg/tHM with 160 kg/tHM pulverised coal injection. Thirty hours after a new coke lot is charged, the furnace becomes unstable. Over the next 8 h the top-gas CO2/(CO+CO2) falls from 0.48 to 0.40, the top-gas temperature climbs from 145 °C to 250 °C, and the blast pressure rises by 0.30 bar at unchanged blast volume, with fluctuations of ±0.15 bar every 20–30 min. Hot metal silicon rises from 0.60 to 1.05%, sulphur from 0.028 to 0.042%, and tap temperature from 1,485 to 1,520 °C. The stockline is normal and no slip has occurred yet. Wall-probe gas analysis shows a much lower CO utilisation than the centre probe. Burden ratio, blast humidity and blast temperature are unchanged, and the furnace has not been blown down.
1. Diagnose the instability. Which single measurement discriminates wall channelling from a hanging, and how does each of the four operating trends fit the mechanism you propose?
Read the worked answer
Discriminating measurement: the radial gas profile. The wall probe reading a lower utilisation than the centre probe proves the gas flow is radially asymmetric. A hanging has a different signature - blast pressure climbs steadily to a peak then drops 0.5–1.5 bar in a step, with a stockline fall and a dust burst - not a regular ±0.15 bar oscillation with a normal stockline.
Fitting the trends to wall channelling:
- Utilisation falls to 0.40. The gas short-circuiting up the wall never equilibrates with the ore, so it leaves rich in CO; the whole-furnace ratio is dragged down by that fraction.
- Top-gas temperature rises to 250 °C. The same gas still carries sensible heat that should have been surrendered in the thermal reserve zone. Heat released per tonne of reductant actually goes down, not up.
- Blast pressure rises at constant volume. The channel is not the reason for the pressure rise - the matrix is. The suspect coke is degrading (low CSR/high CRI), generating fines that tighten the lumpy zone and deadman; a shrinking fraction of the burden carries most of the gas, so total resistance rises.
- Si rises to 1.05% and S to 0.042%. The wall/raceway zone is coke-rich and locally hotter; SiO2 reduction is steeply temperature-dependent, so Si tracks local flame temperature. Higher superheat also lifts sulphur pickup and the metal carries more S because slag volume per tonne has not increased.
Root cause to test first: coke quality and size distribution of the new lot, not the burden.
2. List the first four actions you would take within the next cast and state what each is meant to change. Which commonly used action would you avoid, and why?
Read the worked answer
Actions:
- Cut blast volume 3–5% (keep top pressure). Lower gas velocity reduces the channel's self-reinforcing growth and gives the burden time to heat and soften uniformly; it costs 100–200 t/day for a shift, which is cheap against a slip.
- Add blast humidity (5–10 g/Nm³) or drop blast temperature 30–50 °C. This lowers raceway flame temperature, directly attacking the 1.05% Si and the local overheating driving the peripheral flow. Do it gradually and watch for cold hearths.
- Rebalance burden distribution toward a central coke charge. Opening the centre gives the gas a competing path and pulls flow away from the wall. Judge success on the wall/centre probe utilisation gap closing, not on pressure alone.
- Audit the coke: CSR, CRI, M40/M10, size analysis, moisture, and screen condition. If CSR has dropped below the specification, the fines explanation is confirmed and the burden must be compensated with extra coke until the lot is consumed.
Avoid: increasing blast temperature or cutting the coke rate to recover lost production. Both raise flame temperature, worsen the silicon and the channel, and convert channelling into a scaffold-plus-slip event. Also avoid dumping large ore batches onto the wall to "seal" it - that chills the periphery unevenly and often triggers a hanging. Cast on schedule and keep the taphole clean.
3. Which three coke properties would you write into a purchase specification, and by what mechanism does each control gas permeability?
Read the worked answer
1. CSR (coke strength after reaction with CO2), minimum 60–65%. CSR is measured after partial gasification, so it captures the coke that actually exists in the stack rather than the as-charged coke. Low CSR means the coke disintegrates in the chemical reserve zone into fines that plug the inter-particle voids of the lumpy zone and the deadman, raising the pressure drop and forcing gas to seek a wall path. CSR is the single best predictor of the event described.
2. CRI (coke reactivity index), maximum 22–25%. High CRI means the coke is gasified fast by CO2 and by any solution-loss reaction, which both consumes coke ahead of the raceway (destabilising the heat balance) and produces fines. CRI and CSR are measured in the same test and move inversely; specifying only CSR without CRI allows a supplier to meet the number with a brittle coke.
3. M40 (minimum ~85%) with M10 (maximum ~7%), plus a mean size and size-uniformity requirement (e.g. 45–60 mm mean with <5% below 25 mm). M40/M10 from the Micum drum test govern handling and charging degradation. Size uniformity controls void fraction: a narrow size range gives high voidage, a wide range lets small particles fill voids. Segregation during charging also depends on size distribution, so a coke with excess fines will create low-permeability zones regardless of its mean strength.