Case Study: Blasting Down CO² at Mountsorrel Quarry

As part of IQ's 'Artificial Intelligence – mining data for CO2 reductions' release in September 2024's edition of QM Magazine (available to read online here), Thomas Clifford of GEARS Ltd, joint winner of the Emerald Challenge 2023 gives a detailed case study surrounding his research and project progress.

Overview: 

The Institute of Quarrying Emerald Challenge is an initiative created to invite the mineral products sector to test innovative environmentally focussed ideas that could help the industry on its path to Net-Zero. 

The Emerald Challenge grant is there to support research and development into new processes or concepts to help deliver better performance or to facilitate adoption of existing technologies within the extractive industries.

In 2023, with support from Tarmac, Ground Engineering Applied Research Services (GEARS) presented their initial findings on the CO₂ impact of poor blasting quality to the Institute of Quarrying Emerald Challenge Selection Committee. Their work was rewarded with one of the inaugural Emerald Challenge grants.
Their pivotal study was to assess CO₂ emissions in the UK hard rock quarrying sector and sought, via the use of artificial intelligence (AI), to develop more effective and efficient ways to reduce CO₂ emissions as part of the blasting process, with particular focus on the precise measurement of blast fragmentation. 

To develop the necessary AI tools GEARS worked with AI software consultancy, New Gradient, who specialise in developing bespoke AI solutions for industries that may otherwise not embrace this powerful technology.

Addressing an industry challenge using artificial intelligence: 
Blasting, despite being essential, is costly and often optimised solely for expense reduction, which can compromise blast quality, increase overall production costs and increase the CO₂ emissions. Without precise measurement of blast fragmentation, the impact on CO₂ emissions and on total operational cost and yield remains unclear. 

GEARS research revealed that UK hard rock quarries produce between 2.5 and 4 kilograms of CO₂ per tonne, primarily from diesel-powered load and haul equipment and energy-intensive processing plants. 

Recognising that businesses within the mineral products sector were unlikely to be able to economically or quickly replace their diesel-powered plant and processing operations with renewable energy alternatives, GEARS looked instead at optimising existing processing operations and specifically whether improved blast fragmentation would help reduce CO₂ emissions. 

It was also theorised that this one application of AI could serve as a catalyst for using AI to positively impact other ‘Mine to Mill’ operations, which could offer significant cost and CO₂ emissions reductions for the UK quarrying sector. 

AI rationale and data gathering
AI's ability to identify patterns and abstractions from data make it ideal for understanding complex processes like those in mineral processing. Machine learning, which is the process the AI engine teaches itself, can learn from data to make predictions, thus facilitating intelligent decision-making.

The work at Tarmac Mountsorrel Quarry
Mountsorrel Quarry, a flagship Tarmac site, was chosen due to its high production levels. Even a small improvement in efficiency could have significant CO₂ reductions and cost savings for the quarry.

The quarry had also recently invested in a new primary crusher plant, and this provided a good opportunity to develop and test the new AI powered tools (AggCAM, BlastFRAG and FragMAP), which were developed to improve blast design, implementation and monitoring and so optimise the front end of the production process. 

The research project was supported by Tarmac’s management team as part of their corporate commitments towards net zero. 

Drill and blast process
New Gradient built for GEARS; a new AI-driven tool called FragMAP that accurately measures the block sizes in blast piles. FragMAP is unlike early blast pile image analysis tools in that it has been developed using bespoke AI algorithms and Machine Learning techniques targeted at addressing this specific challenge . The AI powered software can accurately define block boundaries and so measure each block on the surface of the blast pile. 

This represents a massive transformation in technology for blasting as it is set to replace the older photograph analysis fragmentation methods that relied on shadows around edges of block to define the boundary between blocks and as such were plagued with uncertainty and therefore seldomly used by operators. 
Early tests of FragMAP at Mountsorrel Quarry have clearly demonstrated the accuracy and power of the software. To complement the AI technology GEARS, have for the last several years been developing with Tarmac better drone surveying systems for shotfires, which has enabled the creation of detailed face and blast pile 3D models. 

The advancements from this work have enabled a robust new drone survey system to be implemented which shotfirers and quarry operators can be trained to collect the necessary drone survey date of the face and blast pile using standard drones. The drone survey of the blast pile is then analysed by FragMAP from which key performance indicators (KPIs) such as the percentage of oversize, average size and percentage of fines in the blast pile is measured.

These blasting KPIs can then be used to assess the quality of the blast. New Gradient and GEARS have also developed a programme called BlastFRAG which takes input data from the geological constraints of the rock face and the blast design to make blast pile block size predictions that can be compared against the idealised target blast pile block size distribution.

Using a combination of BlastFRAG and FragMAP allows quick effective comparison of blast pile block size prediction with reality. This allows the operator to make informed decisions to change the blast design to bring reality closer to the optimised blast pile target. The aim is to eventually use FragMAP feedback data and AI within BlastFRAG to automate the blast design optimisation for each individual quarry, taking  account of the variations in geology, processing plant constraints and product requirements.

Crushing and screening
The limitation with FragMAP is that it can only measure the surface of the blast pile. As such it is useful to monitor the particle size distribution (grading) at the start and throughout the crushing and screening process. 

This is, however, challenging due to the environment and has been historically very expensive, yet crucial to achieve full Mine to Mill optimisation. To solve this issue GEARS has, with NewGradient, developed AggCAM, another AI tool, that continuously monitors the grading of the material on conveyor belts. 

The most challenging requirement set by GEARS for New Gradient was to develop an AI engine that was clever enough to work off a camera as basic as a GoPro, so ensuring the hardware and set up costs were minimised for operators. 

The first fully operational prototype AggCAM system was installed by Tarmac at Mountsorrel Quarry. The AggCAM data is viewed on a cloud-based system from which Oliver Kibble, Process Engineer at Tarmac was quickly able to identify several inefficiencies within the primary processing plant. This led Oliver and the management team to review the load and haul fleet and so enabled a rapid and well-informed change to optimise the load and haul operation at Mountsorrel.  

Targeted Improvements
GEARS and Tarmac initial observations have revealed that by understanding the quality of the blast fragmentation from using systems like FragMAP, BlastFRAG and AggCAM the CO₂ emissions from hard rock quarries can be potentially reduced by up to 13%. Though the expectation is that as the project continues and with more investigation further CO2 savings will be identified.

Such reductions in CO₂ can be achieved through better blast fragmentation which can not only reduce secondary breakage, but also can increase dumper loads by up to 10% whilst increasing processing plant outputs by up to 15%, subsequently requiring less energy per tonne. 

Whilst the data collection, analysis and optimisation process are still at an early-stage the background research suggests that the total production cost savings could be as high as around £1 per tonne. Whilst also reducing the amount of waste fines being produced and so increasing the high value product yield.  

Conclusion – simplifying complexity
AI's ability to identify patterns and abstractions from data makes it invaluable in optimising complex processes like those in quarrying.  

The project at Mountsorrel Quarry demonstrates the significant potential of AI technology in reducing CO₂ emissions and operational costs in quarrying. 

By integrating AI tools like BlastFRAG, FragMAP, and AggCAM, quarry management teams can now measure not only the cost of each stage of operation but the quality of each stage and so enabling the optimisation of the whole process, which means cost saving and CO2 efficiencies can now be readily achieved. 

Mountsorrel Quarry is now optimising mobile equipment utilisation and planning fleet expansions to ensure continuous production.

The AI approach also delivered some human benefits; it upskilled the team, developing their in-house drone capabilities and fostering a deeper understanding of the processing system.

The transformative power of AI lies, most notably, in its predictive analysis of data, paving the way for intelligent decisions as to how to achieve an even more sustainable future in quarrying.