NTT DATA, Kipu Quantum and Komatsu developed a proof of concept (PoC) to predict critical failures in electric shovel hoist cables used in mining operations. The initiative applies quantum machine learning to generate alerts up to seven days in advance, enabling intervention before failures affect critical operational assets. The approach has demonstrated the potential to reduce costs by as much as one-third.
The initiative addresses one of the mining sector’s most costly challenges. An unexpected failure in an electric shovel cable can bring operations to a halt, drive up maintenance costs and disrupt business continuity. In these environments, unplanned downtime can cost up to three times as much as preventive maintenance.
The project focused on a challenge that conventional methods have struggled to solve: identifying early signs of wear in complex data. Sensor signals contained significant noise, failure events were rare and operating conditions varied across equipment. These factors limited the performance of traditional machine learning models in production environments.
The companies addressed this challenge with Digitized Quantum Feature Mapping (DQFM), a technique that transforms sensor data into feature representations that make degradation patterns easier to identify than with classical methods. Better data representation enabled stronger predictive performance, even with simpler, more robust models that are easier to maintain. The approach also reduced correlation among variables, indicating greater stability and consistency in the results.
The project also incorporated a cost-based decision framework using the Expected Maintenance Cost (EMC) metric to translate predictions into actionable operational decisions. This framework helps balance proactive component replacement against the risk of failure, providing a clearer economic basis for maintenance decisions.
“Quantum machine learning helps us gain deeper insight into the data, uncover previously unseen relationships and turn those insights into simpler, more reliable operational decisions,” said Juan José Miranda, Head of the Innovation Center at NTT DATA Peru. “The value of this pilot extends well beyond mining. Sectors such as energy, transportation and manufacturing face similar challenges in managing critical assets and can apply this approach to optimize operations and reduce costs.”
The pilot demonstrates how quantum computing can deliver practical value in complex industrial environments today. The project combines industry expertise, advanced analytics and emerging technologies to turn complex data into timely, reliable operational decisions at scale.