Summary of Findings

A bridge is an arch of unmelted material spanning the casing above a void. Once it forms, the paste level indication reads the top of that arch rather than the surface of the melt, and the two can differ substantially. Every decision taken on the level reading from that point is taken on a measurement of the wrong surface.

The operational error that follows is predictable and common: the level reads low, so more paste is charged. That adds load and cold material to the top of a structure that formed because material was not melting, which deepens the condition rather than correcting it. Bridging is therefore not primarily a level-control problem. It is a melt-down problem that presents as a level-control problem, and the four contributors to it — unit size and packing, softening point against the thermal conditions, caked material charged as a mass, and a fines-rich layer — are all upstream of the level indicator.

Key Findings

• The level reading becomes invalid the moment a bridge forms, because the indicator reads the arch, not the melt. Divergence between the level reading and the charging record is the primary signal.
• Charging more is the classic error. It adds cold material and load above a void, and the void does not close because it was never a level deficit.
• Four contributors, all upstream. Packing voids between units; charges that did not merge because softening point was too high for the thermal conditions (Wilkinson et al., Infacon IX); caked material charged as a mass; and a fines-rich layer of reduced permeability.
• The consequence below is uneven sintering. Material that arrives at the baking zone in a discontinuous condition bakes unevenly, and the interface between charges that did not merge travels down the column as a plane of weakness.
• This article does not prescribe an intervention method. Clearing a bridge involves a hot furnace, an unstable void and evolved gas; the procedure belongs to the plant’s own safety system and to the furnace supplier.

Contents

1. What is a bridge, and how does it present?
2. What causes one to form?
3. How is it confirmed before anything is done?
4. What governs the response?
5. What does an uncleared bridge do to the column?
6. How is recurrence prevented?
7. Frequently asked questions
8. Related reading

1. What Is a Bridge, and How Does It Present?

A bridge is a self-supporting arch of unmelted or partly melted material spanning the casing, with a void beneath it and the melt somewhere below that. It forms where material stops descending — because it is not melting, because it has locked against the casing wall, or both — while material continues to be charged on top.

The presentation is a divergence rather than an event. Four signals appear, usually in this order:

• The level reading stops responding to charging as expected. Charged tonnage does not produce the level change the charging record implies, or the level appears stable when it should be falling.
• The level reads low relative to the charging record, because part of what was charged is being held above the void rather than joining the melt.
• The level reading is unusually steady. A working paste level moves with charging and consumption; an arch does not.
• Electrical behaviour drifts on that phase, since the material arriving below the baking zone is no longer what the operating cadence was calibrated for.

None of the four is conclusive alone. The combination of an unresponsive level and a divergence from the charging record is what distinguishes a bridge from an ordinary level-control difficulty.

2. What Causes One to Form?

Four contributors, none of which is at the level indicator.

Packing voids between units. Units of any shape stack with voids between them, and larger units produce larger individual voids. Until the paste melts and the voids close, they leave the material above the melt discontinuous — which is the structural precondition for an arch. The relationship between unit size, void size and melt-down time is treated in the block shape article.

Charges that did not merge. Where the binder softening point is too high for the thermal conditions at the paste level, the paste melts insufficiently: the reported consequences are low liquid paste levels, layering of paste from successive charges, and poor electrode formation (Wilkinson et al., Infacon IX). A charge that sits on top of the previous one rather than merging into it is an arch in the making. The determination of a softening point window against the furnace’s own conditions is treated in the softening point article.

Caked material charged as a mass. Material that arrived caked and was charged without being broken down to individual blocks enters as a large irregular unit with correspondingly large voids beneath it. The thermal route that produces caking in transit is treated in the receiving inspection article; the point here is that the condition travels from the container to the casing unless it is broken down at the platform.

A fines-rich layer. Fines migrate into the voids between larger units and concentrate, producing a layer of reduced permeability. Increased fines loading was identified as a contributor to mechanical stress in an electrode management investigation (Wilkinson et al., Infacon IX). A less permeable layer both impedes gas moving upward and helps an arch hold.

The four combine rather than compete. A furnace running large units, a softening point at the upper end of its window and a consignment that arrived caked has three of the four present at once, and their combined effect warrants investigation without assuming a fixed number of charges before bridging.

3. How Is It Confirmed Before Anything Is Done?

Confirmation precedes any intervention, and it is also what prevents the charging error.

1. Reconcile the charging record against the level trace over the preceding shifts. A divergence that grows charge by charge is the signature; a single anomalous reading is not.
2. Check the level with the plant’s normal measuring practice at more than one point across the casing cross-section where the method permits. An arch is rarely uniform, and a single measurement point can read either the arch or a gap beside it.
3. Review the electrical trace for that phase over the same period, against the other two phases. A per-phase divergence with common paste points to that column rather than to the material.
4. Check what was charged. Batch, form, whether the consignment showed caking on arrival, and whether the blocks were broken down before charging.
5. Check the softening point of the batch in service against the contract window and against the preceding batches, and check whether current density has changed. This is the contributor most often missed, because the batch is within specification.

Items 4 and 5 are what turn a bridging event into something that can be prevented rather than repeated. Items 1 to 3 establish that a bridge is present.

4. What Governs the Response?

This article does not prescribe a method for clearing a bridge, and no general method should be taken from any published source. The work involves a hot furnace, a void of unknown extent that can collapse without warning, material that may be evolving gas, and an energised electrode column. The procedure is properly governed by the plant’s own safety system, the furnace supplier’s instructions, and the site’s permit and isolation regime.

What can be stated generally is what not to do while the response is being decided:

• Do not charge more paste onto a suspected bridge. The level deficit is not a level deficit, and charging adds load and cold material above the void.
• Do not treat the level reading as valid for any decision — including slipping decisions — until the condition is resolved.
• Do not assume the void is small. Its extent is unknown until the condition is cleared, and a collapse transfers load suddenly onto material below.

The immediate technical decisions that do belong to the plant are: whether the electrode can continue in service while the condition persists, what the slipping cadence should be given that the level reading is unreliable, and whether the affected phase should be de-rated. Those decisions are made against the plant’s own operating standard and, where a supplier’s technical input is useful, with the batch data and the charging record to hand.

5. What Does an Uncleared Bridge Do to the Column?

Two consequences, both below the bridge and both delayed.

Uneven sintering. Material that reaches the baking zone in a discontinuous condition does not bake uniformly. Baking is the conversion of paste to conducting carbon at approximately 450–500 °C (Larsen, Platinum 2012), and a cross-section that is not uniform on arrival is not uniform after the transition either. The result is a section of column whose properties vary across its diameter.

A plane of weakness. The interface between charges that did not merge — layering — travels down the column as a discontinuity. It reaches the load-bearing region as a plane that has not bonded, and it is one of the routes by which a melt-down problem at the top of the column becomes a breakage further down. The classification of the resulting failure, and how it is distinguished from over-baking, are treated in the breakage diagnostic guide.

There is also a secondary risk worth naming. Where the column is running with unmelted material and the casing is compromised, unbaked paste can flow out into the furnace room — a hole in the electrode casing combined with unbaked electrode paste can cause unbaked paste to flow out (Larsen, 2012). Casing condition is therefore part of the assessment when a bridge is suspected, not a separate matter.

6. How Is Recurrence Prevented?

Each of the four contributors has a corresponding control, and they are worth reviewing together rather than one at a time:

• Unit size against the casing and the thermal conditions. Determined by the procedure in the block shape article, which is a per-furnace determination rather than a general rule.
• Softening point against current density and charging cadence. Determined by the procedure in the softening point article. This is the contributor that persists through a change of block size, which is why a size change alone often fails to stop recurrence.
• Breaking down caked material at the platform before charging, and specifying packing, stacking and storage conditions so that caking is less likely to arrive in the first place.
• Fines control through form selection, packing and the number of handling transfers, with a fines limit and an assessment point in the purchase specification.

Charging practice sits across all four. Smaller, more frequent charges present less cold material at once than the same tonnage charged infrequently, which raises the thermal condition at the low point of the charging cycle — the point the softening point window has to suit.

7. Frequently Asked Questions

Q1. The level reads low. How do we tell a real level deficit from a bridge before charging? Reconcile against the charging record before charging. A real deficit is consistent with the tonnage charged and the consumption since; a bridge shows a divergence that grows with each charge, and a level that responds less than expected to charging. Where the reconciliation is ambiguous, treat it as a suspected bridge — charging onto a real deficit can wait, charging onto a bridge cannot be undone.

Q2. We changed to a smaller block size and bridging continued. What next? Review softening point against current density. Where the softening point is too high for the thermal conditions at the paste level, the paste melts insufficiently regardless of unit size (Wilkinson et al., Infacon IX), and unit size addresses the margin rather than the cause. Compare the batch softening point against the contract window and check whether load or charging cadence changed at the same time.

Q3. Can bridging be detected before the level diverges? Not reliably from the level alone, which is why the charging record and the electrical trace are read alongside it. The earliest available signal is usually a level that responds less than expected to a standard charge, which precedes an outright divergence. Plants that log charge quantity and level response together see it sooner than plants that log the level alone.

Q4. Is a bridge a paste quality problem or an operating problem? Usually a combination, and the four contributors in Section 2 divide across both. Softening point and form are specification questions; caking control and charging cadence are operating and logistics questions; fines are generated across the handling route. A claim raised against the supplier on the strength of a bridging event alone is unlikely to survive the first question about charging practice, which is why the evidence in Section 3 is collected first.

Q5. Should the electrode be de-rated while a bridge is present? That decision belongs to the plant’s operating standard rather than to any general guidance, and it depends on what the column below the bridge is doing and on whether the casing is sound. What can be said is that the level reading is unreliable while the condition persists, so slipping decisions made on it are made on a measurement of the wrong surface.

8. Related Reading

• Receiving Inspection: Caking, Powdering, Binder Separation and Package Damage (companion article) — the condition that arrives at the platform and travels into the casing
• Paste Block Shape and Size: How Geometry Affects Feeding and Sintering — unit size, packing voids and melt-down time
• Matching Softening Point to Furnace Operating Rhythm — the contributor that persists through a block size change
• What Volatile Matter Actually Does Inside a Söderberg Electrode — the gas that a fines-rich layer impedes
• How the Baking Zone Forms and Migrates in a Self-Baking Electrode — where discontinuous material arrives next
• Söderberg Electrode Hard Breakage vs. Soft Breakage — how a plane of weakness presents as a failure further down

Author and Review

Written by Sterling Yang, Senior Vice President at Jiyang Carbon.

This article was reviewed by Mr. Yang, Technical Director at Jiyang Carbon, drawing on 40+ years of hands-on experience in carbon processing.

Values attributed to published sources are cited in this article. No figures are given for bridging frequency, void extent or intervention timing, because no verifiable general values exist. No method for clearing a bridge is given. The work involves a hot furnace, an unstable void and evolved gas, and the procedure is properly governed by the plant’s own safety system and the furnace supplier’s instructions.

Technical Consultation

For assistance in reviewing a bridging history against batch data, form and charging practice, please contact the technical team.

Get a Technical Consultation →