At NTWIST, we've learned that issues with ore classification rarely stem from a single source. Cumulative inaccuracies in drilling, modeling, blasting, and hauling can produce material misclassification so severe that it renders stockpile routing and metallurgical reconciliation ineffective. If your stockpiles aren't producing material as intended, your issue may not be with you reclaim plan - it may originate well before the mill. In this article, we'll explore the causes of ore misclassification and how it can be intercepted before it becomes an irreversible issue.
Orebody modeling uses sparsely placed, interpolated, and smoothed drill data to develop the foundation of grade control planning. But what causes misclassification between model and mill?
Sampling uncertainty: The combination of drill spacing, sampling bias, and interpolation can introduce unquantifiable error
Blast displacement: Unless corrected by vector or modeling, blasts can displace material by meters
Generalized dig boundaries: Dig boundaries are generalized in favor of broader classifications
According to a 2021 MDPI study, errors from sampling and blasting cause up to 19% ore loss and 20% classification error at the open pit (MDPI, 2021).
Misclassified material has downstream effects on processing, and the longer it goes without correction, the more material is affected. If a 1.1 g/t Au zone is mistakenly logged as 0.8, this will affect the makeup of any subsequent stockpiles and the dilution of material. Even more concerning is that misclassification often goes unnoticed until weeks or months later when metallurgical variance becomes apparent.
A 2016 SciELO study of South African gold operations found that misclassification and dilution accounted for up to 30% ore loss when upstream planning failed to account for in-mill conditions (SciELO, 2016).
To combat misclassification, operators need a system that can identify variance within the production cycle and adapt accordingly. Here's how we've helped operators reduce classification error:
Geological model validation: Comparing predicted and produced grades to identify model drift
Blast offset modeling: Using blast movement and in-situ vectors to improve dig boundary classification
Shovel-scale classification: Using reclassification tools to identify variance at the face
Dynamic stockpile routing: Using new classification to update truck routing and prevent misrouting
This creates a self-updating system of classification that adapts as ore is drilled, blasted, and extracted. It enables operators to identify and correct misclassification at every stage of the production cycle.
Modeling is only as good as the correction layers applied to it. If you're not accounting for blasting and sampling variance, you're misclassifying material every day.
Feedback loops fix models faster: Systems that can update based on mill performance can identify and fix classification issues quicker.
Misclassification is a system issue, not an operator issue: The right tools can identify what operators may have missed.
If you're experiencing recovery losses, mill instability, or stockpile unpredictability, the issue may lie in your classification process. Most operators don't have a routing issue - they have a resolution issue. At NTWIST, we help operators achieve precision in classification and routing by uniting geological models, dispatch systems, and metallurgical feedback. We're here to help you stop guessing where things have gone wrong and start knowing where they are.
Ore misclassification occurs when material is incorrectly classified as ore, waste, or a certain grade. This can lead to dilution, ore loss, incorrect routing, and processing issues.
Common causes of ore misclassification include sampling uncertainty, inaccurate geological models, blast displacement, insufficient drill data, and oversimplified dig boundaries.
Ore misclassification can cause ore loss by incorrectly classifying material as waste or low-grade ore. This can lead to incorrect routing material and diluting high-grade ore with low-grade ore.
Mines can reduce ore misclassification through geological model validation, blast movement modeling, real-time classification, shovel-scale grade control, and dynamic stockpile routing.
Blast movement can affect ore classification by displacing material. This can make it difficult to distinguish between ore and waste and lead to incorrect classification.
Pramodh Chandrashekar is an industry and technology expert and Senior Researcher at NTWIST. His expertise includes industrial AI, mining and manufacturing technology, digital transformation, and operational optimization.
He focuses on applying AI and data-driven technologies to improve operational efficiency and business performance.
LinkedIn: Pramodh Chandrashekar