Finding cheaper and more sustainable ceramic glazes using data science

Challenge

Can you imagine a world without glazes for your ceramic ornaments and household items, or for your pottery classes? In ceramics, glazes are crucial as they affect not only the aesthetics but also the physical properties of ceramics articles, such as the ability to hold liquids. This means that choosing the appropriate glaze is an important part of ceramics work. Although a myriad of glaze recipes can be found through open-sourced data online, users are unable to filter their search based on glaze ingredients. It is therefore challenging for ceramicists to find glaze recipes that match their budget and the availability of ingredients while providing the right glaze properties.

Approach

Ceramics manufacturer Box O Fun produces ceramic articles and tools. The company wanted to help ceramicists explore glaze options that fulfil cost, functionality and aesthetic requirements. Using an Innovate UK BridgeAI £15,000 Innovation Voucher, the company was able to access the Hartree Centre’s data science expertise. This allowed for data retrieval using automated methods, enabling the collation of data from approximately 25,000 glazes. Using the collated data, the team built the search function based on the properties and material compositions of these glazes. In this newly developed feature, users can input available materials and their quantities to find economical glaze recipes that deliver the desired properties.

Benefits

Using automation, our data science team was able to speed up and expand the scope for glaze data collation, such as material composition and carbon output for as many as 25,000 glazes. With the search function built upon the vast amounts of data collected, ceramicists at Box O Fun can now find appropriate glazes that they can create using available materials matching their budget. By helping ceramicists choose
cost-effective ingredients including leftover materials, Box O Fun will benefit both in terms of cost and waste reduction, with the latter also driving environmental sustainability for the company.

This project has helped us to explore the possibility of machine learning with materials. It led us to question what more could be done to transform the foundation industries through the use of technology, particularly when working towards net zero targets.

– Layla Hashim, Box O Fun