Foreword

Seeing alternative fuel handled in a cement plant is to watch a constant battle to make order out of chaos. The sour-sweet silage smell lingering in the hangar-like reception yards reminds you of the biogenic component of the piles of refuse that, while unpleasant to handle, gives this fuel a lower carbon footprint than the coal or petcoke it replaces.

While large new facilities are built to sort, grade and store plastic and other waste, most plants have to pay close attention to the weather, as humidity or rainfall can play havoc with how the fuel burns. It makes you wonder if, in fifty years, we’ll look back on burning waste as a lasting success, as it has been in Germany over the last decade, or if advances in plastic reuse and reduction will keep us shifting towards new alternative fuel sources.

Amid all the media hype about how artificial intelligence is going to change the world, efficient waste disposal is not the first topic that springs to mind. But creating order from chaos is what AI does best. Just as Large Language Models can turn vast amounts of text into a coherent response to your question, machine learning can take the enormous datasets painstaking gathered by the cement industry into new insights about the process. We can apply these insights to answer questions like: How can we optimise cement production despite a constantly changing fuel supply? And, how can we practically leverage AI to reduce carbon emissions? These are critical questions, especially as the European Union begins phasing out free carbon allowances over the next decade.

In this whitepaper, we outline how AI is being used to optimise alternative fuel usage today, illustrated by the publication of our recent case study results. We also highlight the scope for future development of AI applications in this space.

We welcome your thoughts and experiences. As people working in the cement industry, you have the best insight into our most pressing question: What problem should AI tackle next?