Aberystwyth University (Prifysgol Aberystwyth) and Parkinson-Spencer Refractories Limited KTP 24_25 R1

A ceramics kiln will soon decide where to place its own pots, bricks, and tiles, using machine learning to pack every inch of space more efficiently. This project tackles a hidden inefficiency in manufacturing: kiln loading. Ceramic products are fired in large, energy-intensive kilns, and how they are arranged inside determines both energy use and output. Current loading methods are often manual or rule-of-thumb, leaving gaps that waste heat and increase carbon emissions. The researchers are building a computational platform that uses artificial intelligence and advanced simulations to calculate optimal product placement, treating the kiln as a puzzle to be solved mathematically. If the platform works, ceramics manufacturers could fire more product per kiln load while using less energy. That means lower production costs and a smaller carbon footprint for an industry that relies on high-temperature processes. The impact would be felt in supply chains and factory floors rather than in homes—a quieter, behind-the-scenes improvement to how everyday materials like bricks, tiles, and sanitaryware are made. This is applied engineering, not fundamental science: the goal is a practical tool for industry, not new knowledge for its own sake.