By Wu Peng, Senior Process Instrumentation Engineer · Last reviewed August 24, 2026
Blast furnace level measurement tracks the burden surface, called the stockline, at the top of an ironmaking furnace. The working standard today is non-contact FMCW radar reading 2 to 15 m below the zero line with an accuracy of ±5 mm or ±0.2 percent of span. Readings are cross-checked against mechanical sounding rods.
The measurement point is one of the harshest in heavy industry. The throat runs 200 to 400 °C, dust rises on every charge, and reducing gas moves upward at 5 to 15 m/s. This guide compares the four measuring technologies and works through radar frequency selection with beam footprint math. It then sets out the purge air, mounting geometry, and commissioning checks that decide whether the instrument lasts.
Contents
- Why measurement is hard
- Four technologies compared
- Choosing the frequency
- Purge and cooling
- Mounting geometry
- Range and outputs
- Commissioning cross checks
- Reading the trend
- FAQ
Why measurement is hard
Three furnace conditions defeat ordinary level instruments. The first is heat. The gas space above the burden runs 200 to 400 °C in normal operation, and the charge material a few meters down glows at 800 to 1200 °C. Any instrument at the top sees both conducted heat through the nozzle and radiant heat from below.
The second is the atmosphere. Top gas is a reducing mixture of CO, CO2, H2, and nitrogen moving upward at 5 to 15 m/s, loaded with fine coke and ore dust. On a modern high-top-pressure furnace the gas space also sits at up to roughly 0.25 MPa gauge, so every penetration must hold furnace pressure.
The third is the charging cycle itself. Each charge drops tens of tonnes of coke or ore onto the surface in a few seconds. The impact throws up a dust cloud of 10 to 100 µm particles that scatters light completely; laser instruments have no place on a furnace top for that reason. The surface being tracked also moves the moment the material lands.
Four technologies compared
Ironmakers have handled stockline measurement with weighted rods for more than a century, and with radar as the mainstream choice since roughly the late 2010s. Radioactive sources filled the gap in between. The table compares all four on the points that matter to an operator.
| Technology | Reading | Behavior during charging | Main limitation |
|---|---|---|---|
| FMCW radar | Continuous, updated in under a second | Tracks through dust when purged; brief echo loss is averaged out | Needs purge air and cooling to survive |
| Mechanical stock rod | One point per sounding cycle | Must be winched clear before material falls; blind while raised | Contact wear, chain and drum maintenance |
| Radiometric (isotope) | Point detection at fixed elevations | Unaffected; measures through the shell | Source licensing, handling rules, no continuous profile |
| Laser | Continuous in clean air | Blinded by the dust cloud | Not viable in furnace dust |
Published sounding systems from a major Japanese maker winch the rod down at 42 m/min and back up at 60 m/min, so even a short 3 m sounding takes several seconds each way plus settling time, once per cycle.
In practice the technologies are not rivals. A typical modern furnace top carries one to three radars for continuous reading and keeps two mechanical rods as the calibration reference and backup. Radiometric point gauges persist where a plant already holds source licenses. The pairing also gives a new radar its reference at commissioning.
Choosing the frequency
Radar frequency sets two things at once: how tightly the beam focuses, and how the wavelength interacts with dust. The beam footprint on the burden follows a simple cone formula.
Footprint diameter = 2 × R × tan(θ/2)
At R = 15 m: a 3° beam covers 2 × 15 × tan(1.5°) = 0.79 m, while a 12° beam covers 3.15 m.
| Frequency | Wavelength | Typical beam | Footprint at 15 m | Furnace behavior |
|---|---|---|---|---|
| 6 GHz | 50 mm | 10−15° | 2.6−4.0 m | Best dust penetration; wide beam clips wall and chute |
| 26 GHz | 11.5 mm | 6−8° | 1.6−2.1 m | Workable compromise; echo fades when dust load peaks |
| 80 GHz | 3.75 mm | 3−4° | 0.8−1.1 m | Footprint stays on the burden between wall and chute |
The physics pulls in both directions. A longer wavelength passes through suspended dust more easily, which is why furnace radars of the 1990s and 2000s ran at 6 to 10 GHz. But a wide beam on a 6 to 10 m throat also returns echoes from the armor plates and the charging chute. Separating those from the burden echo is where legacy systems spent their engineering budget.
The 80 GHz generation resolves the conflict differently: dust particles of 10 to 100 µm are still far smaller than the 3.75 mm wavelength, so scattering stays modest. The 3° beam then simply avoids illuminating anything except burden. That combination is why 80 GHz FMCW has become the default blast furnace radar level specification for new installations.
FMCW itself beats pulse radar here for a signal reason. The transmitter sweeps a continuous chirp and converts echo frequency shift to distance, averaging over a window of a few hundred milliseconds. Continuous averaging at low peak power holds signal-to-noise through the dust bursts that punch holes in a short-pulse return.
Purge and cooling
A furnace radar fails in two ways: the antenna cokes over with dust, or the electronics overheat. The purge and cooling specification prevents both, and it belongs on the instrument data sheet with numbers, not as a note that says purge required.
Purge air keeps the antenna face clean. Specify regulated, dry, oil-free air with the supply held above furnace top pressure with margin. A continuous bleed of 5 to 15 Nm³/h is the common band.
Add a solenoid-timed burst of 30 to 50 Nm³/h fired in rhythm with the charging cycle, for example a few seconds out of every half minute. Air dewpoint should stay at or below −20 °C so nothing condenses where the air meets the cooler antenna face.
Cooling is a separate circuit. The electronics head on any transmitter carries an ambient rating near 80 °C. Holding it there next to a 200 to 400 °C flange takes either a cooled standoff or a low-pressure cooling air line sized against the housing rating. A thermal-bridge extension between flange and head, one to two meters of waveguide or neck, drops the conducted heat so the head end runs at ordinary electrical-room temperatures.
Two engineering rules follow. First, feed the purge from a dedicated dryer rather than the general plant air header, because one wet-air event fouls the lens and costs a cleaning shutdown. Second, alarm on loss of purge: a radar that keeps reporting while its purge has failed produces plausible numbers right up until the antenna blocks.
Mounting geometry
Modern furnaces charge through a bell-less top, a rotating chute distributing material across a throat 6 to 10 m wide. The radar cannot mount on the centerline because the chute sweeps through it. The workable compromise is off-axis.
- Offset the nozzle 1.0 to 2.0 m from the furnace centerline, clear of the chute swing envelope.
- Keep the antenna 1.5 to 3.0 m above the highest stockline the operator will run.
- Tilt 2 to 3° toward the burden centroid and confirm the aim against the maker’s beam plot.
- Mount through a full-bore ball or knife valve so the instrument comes out for service without breaking furnace pressure.
The nozzle standoff doubles as thermal protection. A waveguide extension of about two meters between furnace flange and antenna keeps the mounting flange in the workable range. It also gives the purge air a defined duct to sweep.
On coke dry quenching chambers and raw material silos the geometry is simpler: a straight nozzle clear of the fill stream. The same valve isolation rule still applies.
Range and outputs
Stockline service on a mid-size furnace needs 2 to 15 m of range measured from the antenna. Sister applications at the same plant stretch further, and the same high-temperature radar line covers them with different antenna and range settings.
| Application | Typical range | Accuracy target |
|---|---|---|
| Blast furnace stockline | 2−15 m | ±5 mm or ±0.2% of span |
| Coke dry quenching chamber | 2−25 m | ±0.2% of span |
| Raw material silos | 2−30 m | ±0.2% of span |
Output is 4−20 mA with HART as the universal interface, and most furnace projects also wire Modbus RTU over RS-485 so the level model in the process computer reads the same instrument digitally. Where top gas hydrogen puts the area under an explosive-atmosphere classification, specify the certification on the data sheet rather than assuming the standard build carries it.
On the product side, a high-temperature radar level sensor covers the furnace top itself, while an 80 GHz radar level transmitter brings the narrow beam to stockline and deep silo work. A long-range radar level sensor reaches the tallest raw material bins. The wider technology landscape is mapped in our radar level sensor range and the level transmitter types guide.
Commissioning cross checks
A stockline radar is commissioned against the sounding rods, not against a calibration bath. The furnace is never empty, so there is no empty-tank echo to fingerprint. Zero is established from the drawing datum instead: survey the antenna reference plane against the stockline zero elevation and enter the offset. Then prove the reading against reality.
Run the radar and the rods in parallel for the first month. Every sounding cycle yields one direct contact reading; log it beside the radar value at the same timestamp and watch the difference. A stable offset points to a datum error, which is correctable in configuration. A scattered difference points to echo trouble, usually the beam clipping the chute or a wall echo winning during charging.
One timing rule keeps the comparison clean. Judge descent rate only in the idle interval when the charging gates are closed. During a charge the surface is a moving pile of falling material, and neither instrument reads anything comparable. Gating the comparison this way removes the largest source of scatter from the log.
Reading the trend
The stockline number feeds the charging control loop: when the surface descends past the charging threshold, the next batch goes in. But the trend carries more diagnosis than the absolute value, and the radar makes the trend continuous for the first time.
Normal operation shows a sawtooth: steady descent at centimeters per minute, then a step up as a charge lands. Two abnormal signatures matter.
A slip appears as a sudden drop of a meter or more within seconds, usually after a period where descent had stalled. It warrants an alarm because the collapse drives a pressure surge. A hang shows the opposite signature: the level stops descending while the process continues underneath, and the gap between charges stretches.
Configure the control system to flag both patterns from the level trend. The measurement pairs naturally with the same radar architecture used on other extreme services. The closest relative is molten salt level measurement, which shares the cooled-flange design problem with a different chemistry.
Application note
Specifying a stockline radar. A complete inquiry for furnace-top service pins down six items. They are throat geometry and charging type, the stockline range, top pressure and gas temperature, purge air quality at the top platform, the isolation valve standard, and the output protocols the process computer expects. With those six fixed, the frequency, antenna, and cooling configuration follow almost mechanically.
The flow-side companion for the same plant is our guide to coke oven gas flow measurement.
FAQ
What is the burden in a blast furnace?
The burden is the charged raw material column: alternating layers of coke and iron-bearing material such as sinter, pellets, and lump ore, plus flux. The burden surface at the top is the stockline, and its elevation is what blast furnace level measurement tracks.
How is the burden level measured in a blast furnace?
Modern furnaces use non-contact FMCW radar, usually at 80 GHz, mounted off-axis on the furnace top and purged with dry air. Mechanical sounding rods remain installed as the contact reference, and radiometric point gauges appear on some older installations. Radar gives the only continuous reading through charging dust.
How do you calculate the blast furnace capacity?
Multiply working volume by productivity. Working volume is measured from the tuyere level up to the stockline zero. Modern furnaces produce roughly 2 to 3 tonnes of hot metal per cubic meter of working volume per day. A 3,000 m³ working volume at 2.5 t/m³ gives about 7,500 tonnes per day.
Can you provide an analysis of blast furnace gas?
Typical top gas runs about 20 to 24 percent CO, 18 to 23 percent CO2, 2 to 5 percent H2, and the balance nitrogen. The heating value sits near 3 MJ/Nm³. The CO and H2 content is why furnace-top instrumentation must respect explosive-atmosphere rules and why every penetration seals against gas leakage.
Request a quote
Tell us the throat geometry, stockline range, top pressure, and purge air available. We will return a furnace-rated radar configuration with the cooling and valve interface defined. Reach our application engineers or use the form below.
Written and technically reviewed by Wu Peng and the Instranova engineering team.