BRENNEN

Industries / Mining and materials / Glass and ceramics / Refractories

Poured linings

Poured linings, called monolithics, arrive as dry mixes and set in place: water and a binder turn them into a solid lining after they are poured, vibrated, pumped, rammed or sprayed into the furnace, or cast into blocks beforehand.

5

Milestones achieved, from 1908 to 1992.

1

To watch for: announced, not yet confirmed outside.

2

Open deliverables. Nobody has shipped them.

The road

AchievedWatch forOpen

Achieved

  1. 1908

    A cement made with bauxite

    Calcium aluminate cement is made by burning or melting lime with alumina instead of the clay and lime of ordinary Portland cement. J. & A. Pavin de Lafarge filed a French patent on 13 June 1908 (FR 391,454); its British counterpart describes adding bauxite and lime to molten blast-furnace slag in a furnace, then granulating and grinding the product. Wikipedia credits the 1908 patent to Jules Bied of the Lafarge company. The cement sets fast, gains its strength within 6 to 24 hours of curing and resists chemical attack; Ana Paula Luz and colleagues, in a 2022 review, call it the most representative binder of refractory castables.

    patents.google.compatents.google.comarxiv.orgen.wikipedia.org

    Credit

    Jules Bied; J. & A. Pavin de Lafarge

  2. 1969

    Less cement, less water

    Dense refractory concretes of the 1960s held about 12 to 20% calcium aluminate cement, and they lost much of their strength between the temperature at which the cement's hydrates break down and the temperature at which a ceramic bond forms. Lucien Prost and André Pauillac of Prost-Produits Réfractaires in France filed a patent in October 1969 for mixes with 5 to 8 parts of cement, a few parts of powder finer than 50 µm (preferably finer than 1 µm) and a dispersing agent, which lets the mix flow with less water: 6 parts of water instead of 10 for a conventional concrete in one of its examples. Krosaki Harima says the high-strength low-cement castable it developed in 1978 ended up sold under a non-exclusive license from Prost after a patent dispute.

    patents.google.compatents.google.comkrosaki.co.jp

    Credit

    Lucien Prost and André Pauillac, Ets Prost-Produits Réfractaires

  3. 1985

    Fibers let the steam out

    A newly cast lining still holds water. On its first heating the water turns to steam, and in a dense, tight castable the steam can build pressure faster than it escapes until the lining bursts, which is called explosive spalling; it can begin at about 150 °C. W.H. Gitzen and L.D. Hart described explosive spalling of cement-bonded castables in 1961, and by 1983 to 1985 papers by P.H. Havranek and by T.R. Kleeb and J.A. Caprio were reporting castables made to resist it and service experience with monolithics containing organic fibers, as cited in a 2022 review by Luz and colleagues. Plastic fibers such as polypropylene soften and melt at 120 to 170 °C and leave channels through which steam can leave. Calderys still calls the dry-out of dense low-cement castables "always" a challenge.

    LessonMaking castables denser and stronger with less water also made them hold their steam, so a lining's ability to vent on its first heating matters as much as its final strength.

    arxiv.orgprozesswaerme.net

    Credit

    W.H. Gitzen and L.D. Hart; P.H. Havranek; T.R. Kleeb and J.A. Caprio

  4. 1980s

    Castables replace ladle bricks

    Nippon Steel's engineers write that steel ladles were lined with high-silica bricks, then zircon bricks, and that castables began to spread around 1970 to mechanize lining work and save labor. Alumina-spinel castables, developed in the late 1980s, became the main ladle lining; spinel is a mineral made of magnesium and aluminum oxides. They wore where slag soaked into the wall and flaked it off, and the bottoms arched and cracked from thermal expansion. Alumina-magnesia castables followed. These form spinel in use, expanding quickly above about 1,400 °C; held in by the ladle, the working face grows denser, with smaller pores, and keeps slag out, so they are now widely used in ladles.

    LessonA mix that keeps reacting in service can be designed so that the reaction closes its pores against slag.

    nipponsteel.com

    Credit

    Japanese steel and refractory makers; Nippon Steel's 2013 review names no single developer

  5. 1992

    Poured linings overtake bricks

    Krosaki Harima writes that production of shaped refractories peaked in 1970 and then declined while monolithics kept growing, and that by 1992 more monolithic refractories than shaped ones were being produced. It ties their fast growth in the 1970s to spray-repair systems, and notes that within monolithics, castables grew while slinger mixes (thrown into place by machine) and ramming mixes (packed in by hand or tool) declined. Nippon Steel likewise dates the wider use of castables in steelworks to around 1970, for mechanized lining work and labor saving. A 2005 report by James Hemrick and colleagues at Oak Ridge National Laboratory, for the US Department of Energy, found that monolithics made up at least 60% of the refractories used in US ironmaking and steelmaking and over 70% in Japan.

    krosaki.co.jpnipponsteel.comosti.gov

    Credit

    Japanese refractory makers and steelmakers

Watch for

  1. Fast-drying castables for whole furnaces

    Calderys says its CALDE FD castables, cement-bonded low-cement mixes formulated to release their water faster and more evenly, can cut the dry-out needed to bring a new lining to 1,200 °C by up to 72 hours, against a usual controlled dry-out window of four to seven days. It says they can line an entire furnace, where fast-drying products were earlier limited to local repairs or certain zones.

    Achieved whena furnace operator or outside party publishes measured dry-out times and lining condition for a furnace lined entirely with these castables, compared with a conventional low-cement lining.

    prozesswaerme.net

    Claimed by

    Calderys, in the trade journal heat processing, September 2025

Open

  1. 01

    A table of castable mixes

    An open table of refractory castable recipes and test results gathered from published studies and open datasets: aggregates and their sizes, fine powders, cement content, microsilica, dispersant, water, curing and firing conditions, and the measured flow, density, porosity, cold crushing strength, permanent change in length after firing and hot bending strength, with a source on every row. Add a baseline model that predicts strength from the recipe.

    The brief
    Credit · unclaimed
  2. 02

    A benchmark for steam spalling

    A public set of published heating tests on castable samples, each recording the mix, sample size, curing, permeability where measured, heating schedule and outcome (whether the sample burst, and at what temperature), plus a scoreboard for models that predict the outcome. This is desk work: the data come from papers and theses, not from new furnace tests.

    The brief
    Credit · unclaimed

Open 01 · the brief

A table of castable mixes

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

An open table of refractory castable recipes and test results gathered from published studies and open datasets: aggregates and their sizes, fine powders, cement content, microsilica, dispersant, water, curing and firing conditions, and the measured flow, density, porosity, cold crushing strength, permanent change in length after firing and hot bending strength, with a source on every row. Add a baseline model that predicts strength from the recipe.

Why it is still open

Raw data from castable studies appear one study at a time, for example a March 2026 Zenodo set from Lapenko and colleagues comparing cement-free and cement-bonded cordierite castables. Machine learning work on refractories has centered on MgO-C bricks, and its reviewers name the lack of high-quality data as the main obstacle. Searches of Europe PMC, GitHub, Hugging Face and Zenodo in November 2026 found no table that pools castable mixes across studies.

zenodo.orgebi.ac.uk

Built with what exists

Python with pandas and scikit-learn; WebPlotDigitizer for values printed only in charts; an AI model to find papers and pull their tables, every value checked by hand against the source (see the Note "Seven more flagged"); Zenodo for a versioned release with a DOI.

  1. First step

    Tabulate the mixes and results in ten open-access castable studies, starting with the Zenodo datasets and the arXiv preprints of the Federal University of São Carlos group.

  2. Done looks like

    At least 300 mixes from at least 30 sources in one openly licensed file, each traceable to a table or figure, and a model predicting cold crushing strength after drying, with its error on sources left out of training published.

Open 02 · the brief

A benchmark for steam spalling

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A public set of published heating tests on castable samples, each recording the mix, sample size, curing, permeability where measured, heating schedule and outcome (whether the sample burst, and at what temperature), plus a scoreboard for models that predict the outcome. This is desk work: the data come from papers and theses, not from new furnace tests.

Why it is still open

Murilo Moreira, Roberto Ausas and Stefano Dal Pont released an open-source model of drying damage and explosive spalling in castables in October 2026; they validated it against literature results and compared it with experiments only qualitatively. Moreira's thesis data are on Zenodo, but each experimental study reports a few castables in its own format, and no table brings together the outcomes of heating tests from many studies so that models can be scored on them. The 2022 review by Luz and colleagues says heating schedules are still "usually based on empirical knowledge", and in 2025 a Lithuanian team had to design a new test after a boiler lining spalled on its first heating. Searches of Zenodo, Hugging Face and GitHub in November 2026 found papers and posters on castable drying, and no pooled table of test results.

arxiv.orggithub.comzenodo.orgarxiv.orgdoi.org

Built with what exists

The THM-PFM code (Python with FEniCS); WebPlotDigitizer; Python with pandas; a simple statistical baseline such as logistic regression on permeability, sample thickness and heating rate.

  1. First step

    List every heating test with a reported outcome in Moreira's thesis data and in the two drying reviews by Luz and colleagues.

  2. Done looks like

    At least 60 heating tests with reported outcomes from at least 10 studies; THM-PFM and the baseline each run on all of them, with the share of outcomes each predicts correctly published.