AI-Optimized Blast Design and Execution

Key Takeaways

• AI blast design optimization is available today: Strayos uses drone photogrammetry and ML or genetic-algorithm optimization, trained on historical blast outcomes, to recommend burden, spacing and timing before you drill.
• O-Pitblast and Deswik.Blast are established commercial alternatives with strong simulation support for the same design step.
• Orica BlastIQ is the cloud platform connecting drill data through to blast execution. It standardizes how design intent reaches the field.
• Explosive loading itself (physically charging the holes) has the least AI penetration. Current systems mostly add QA/QC checks rather than optimization.
• Fragmentation and vibration prediction is trained on your own historical blast data, so accuracy improves the more blasts you feed back in.

TL;DR

Blast design optimization with AI starts with drone-captured bench topography and your historical blast results. An AI platform such as Strayos, O-Pitblast or Deswik.Blast turns them into an optimized pattern before drilling. You then execute and track that design through a connected platform like Orica BlastIQ.

What Is Blast Design Optimization?

Every open-pit blast follows a design. The engineer sets hole diameter, burden (distance to the free face), spacing between holes, stemming, sub-drill, explosive type and delay timing. Together these control powder factor, which is how much explosive energy goes into each tonne of rock. The goal is rock broken to the right size, moved the right amount, with vibration and airblast inside limits.

Traditionally, engineers set these values from empirical formulas, such as the Kuz-Ram fragmentation model, and rules of thumb. They then adjust after each blast based on the results. Blast design optimization replaces much of that trial and error with prediction. A model forecasts fragmentation, movement and vibration for a candidate design, then searches for the design that best meets your targets.

How Do I Optimize Blast Design and Execution With AI?

The workflow has three parts: capture the bench, optimize the design, then carry it into the field.

Capturing the bench as a digital twin

Strayos builds a 3D digital twin of the bench from drone photogrammetry and smart-drill data. A good drone survey gives accurate face profiles, so burden on the front row is measured rather than guessed. Drill data adds information about rock hardness along each hole. Tools like BLASTDOG blast-hole logging can add rock mass data hole by hole.

Running blast design optimization on the pattern

Next, Strayos applies machine learning and genetic-algorithm optimization, trained on your site’s historical blast outcomes. It recommends burden and spacing, differential energy loading and timing. It also predicts fragmentation, material movement and vibration or airblast before you drill a single hole.

O-Pitblast (via its O-Pitsurface product) and Deswik.Blast cover similar ground with strong simulation and vibration prediction. They represent the more established commercial end of blast-design software. They rely less on AI trained on your data and more on configurable simulation with decades of empirical models behind it. Your choice often depends on what already integrates with your mine-planning stack.

Carrying the design into execution

Once a design is final, execution runs through a connected platform. Orica BlastIQ is the dominant cloud-based option. It stores and shares blast data across the drill-to-blast workflow, so the field crew works from the design the engineer optimized. Post-blast results then flow back in to improve the next cycle. However, the loading step itself sees little AI. Orica’s electronic detonator line (WebGen, for instance) adds QA/QC checks and logging rather than optimization.

What to Expect From Blast Design Optimization in Practice

These tools are only as good as the historical blast data you feed them. A site with a few years of consistent fragmentation and vibration monitoring will get real value from ML recommendations. By contrast, a site starting from scratch initially gets a well-built simulation tool more than a trained optimizer.

Therefore, the feedback loop matters as much as the design tool. Blasting evaluation measures fragmentation and ore movement after each blast. Those results train the next design. Better fragmentation also pays off downstream in shovel productivity and crusher throughput.

FAQs

What does AI change in blast design?

AI predicts fragmentation, movement and vibration for a candidate design before drilling. It then searches for the burden, spacing, loading and timing that best meet your targets. Traditional design relies more on empirical formulas and after-the-fact adjustment.

Which software supports blast design optimization?

Strayos offers AI-native optimization trained on site blast history. O-Pitblast and Deswik.Blast provide established simulation-based design tools. Orica BlastIQ connects the final design to drilling, loading and execution in the field.

How much data does blast design optimization need?

Ideally, you need a few years of consistent fragmentation and vibration records linked to each blast design. With less history, the tools still work as simulators. Their recommendations improve as each new blast is recorded and fed back.

Can AI optimize explosive loading?

Only to a limited extent today. Loading has the least AI penetration in the blasting workflow. Current systems mainly add QA/QC checks and logging to make sure the crew loads holes as designed.

Try It With Geocluster

An optimization tool needs a coherent dataset to learn from: drone surveys, drill logs and historical blast performance. Assembling that dataset is a research problem before it is a blast-design problem. Geocluster is built to support exactly that work if you are building a blast-optimization feedback loop. You can also read how the Geocluster Research Harness works.