Case Study - AI industrial process optimization
AggCam: Quarry Rock Processing Optimization
Summary
New Gradient collaborated with Tarmac on their "Blast Down CO2" project, which won the Emerald Challenge. We developed AggCam, an AI-powered solution that significantly reduced energy use, costs, and CO2 emissions in the rock processing pipeline for aggregate production.
AI quarry process optimization
Hard rock quarries face significant challenges in optimizing their rock crushing process. Multiple crushing stages require vast amounts of energy, with crusher settings needing constant adjustment based on input rock sizes. Continuous determination of rock size in a high-speed production line is difficult, as is identifying and responding to downtime in specific stages of the processing pipeline. These inefficiencies can lead to substantial energy waste, potentially costing sites up to £200,000 per year and producing significant extra CO2 emissions. This complex industrial challenge required an innovative machine learning solution to optimize processes and reduce inefficiencies.
Our AI developers created AggCam, a lightweight deep neural network that analyzes high-speed imagery from cameras positioned above crusher input and output feeds. This innovative system identifies rock size distribution in real-time and detects processing downtimes.
Initially implemented for cloud computing as part of Tarmac's "Blast Down CO2" project, AggCam helped identify and eliminate 5-10% of downtime where crushers were running empty. We are further developing this technology for edge devices to enable fully real-time monitoring and adjustment of crusher settings based on rock size disparities.
The implementation of AggCam has already yielded significant benefits, including increased production throughput and reduced wasted energy costs and CO2 emissions. This project not only contributed to Tarmac's victory in the Emerald Challenge but also demonstrates New Gradient's AI consulting expertise in applying cutting-edge artificial intelligence to complex industrial processes, delivering tangible environmental and economic benefits. The ongoing development of edge-based solutions promises even greater real-time optimization capabilities in the future.
