Drill bit failure is a prominent concern in the drilling process of any mine, as it can lead to increased mining costs. Over the years, the detection of drill bit failure has been based on the operator’s skills and experience, which are subjective and susceptible to errors. To enhance the efficiency of mining operations, it is necessary to implement applications of artificial intelligence to produce a superior method for drill bit monitoring. This research proposes a new and reliable method to detect drill bit failure in rotary percussion drills using deep learning: a one-dimensional convolutional neural network (1D CNN) with time-acceleration as input data. 18 m³ of granite rock were drilled horizontally using a rock drill and intact tungsten carbide drill bits. The time acceleration of drill vibrations was measured using acceleration sensors mounted on the guide cell of the rock drill. The drill bit failure detection model was evaluated on five drilling conditions: normal, defective, abrasion, high pressure, and misdirection. The model achieved a classification accuracy of 88.7%. The proposed model was compared to three state-of-the-art (SOTA) deep learning neural networks. The model outperformed SOTA methods in terms of classification accuracy. Our method provides an automatic and reliable way to detect drill bit failure in rotary percussion drills.

This study introduces a novel method for analyzing vibration data related to drill bit failure. Our approach combines explainable artificial intelligence (XAI) with convolutional neural networks (CNNs). Conventional signal analysis methods, such as fast Fourier transform (FFT) and wavelet transform (WT), require extensive knowledge of drilling equipment specifications, which limits their adaptability to different conditions. In contrast, our method leverages XAI algorithms applied to CNNs to directly identify fault signatures from vibration signals. The signals are transformed into their frequency components and then employed as inputs to a CNN model, which is trained to detect patterns indicative of drill bit failure. XAI algorithms are then employed to generate attention maps, highlighting regions of interest in the CNN. By scrutinizing these maps, engineers can identify critical frequencies associated with drill bit failure, providing valuable insights for maintenance and optimization. This method offers a transparent and interpretable framework for analyzing vibration data, enabling informed decision-making and proactive maintenance strategies to enhance drilling efficiency and minimize downtime. The integration of XAI with CNNs facilitates a deeper understanding of the root causes of drill bit failure and improves overall drilling performance.

Rotary percussion drills, specifically top hammer types, are essential in resource extraction and exploration. These drills frequently suffer damage at the drill bit, necessitating effective decision-making to address issues. Traditionally, on-site operators relied on intuition, but recent advancements promote automated diagnostic systems to enhance safety and failure detection in challenging mining environments, such as those with high dust levels and intense vibrations. This study uses Convolutional Neural Networks (CNN) to develop an automated diagnostic system for detecting drill bit abnormalities. Previous models classified five distinct bit conditions but faced challenges in generalizing performance due to variations in the frequency bands used for judgments. This study refines the approach by simplifying the system to detect two states: normal and abnormal. Using Gradient-weighted Class Activation Mapping (Grad-CAM), the system identifies critical frequency bands, enabling a unified frequency filter. By training the model to recognize differences in rock types, it adapts to untrained drill bit types, achieving an accuracy of 83.3%, significantly higher than the 44.8% in previous models. This innovation allows the system to generalize across different rock and bit conditions, improving drilling efficiency and minimizing downtime in diverse environments.

Monitoring drill bit wear during drilling improves operational efficiency. Wear in drill bits is influenced by factors related to drilling parameters and rock properties and can be categorized into types based on the cause of failure. Several studies have used machine learning to detect drill bit wear, mainly using classification models. However, these models require an extensive amount of data that includes different possible drill bit failures. This presents a challenge as acquiring data is difficult due to the scarcity of anomalies. This study introduces a novel approach utilizing convolutional autoencoders (CAE) for both anomaly detection and drill bit damage assessment within rotary percussion drilling. The proposed CAE model was trained on the vibrations of normal data, learning to accurately reconstruct input patterns. A dual threshold was set based on the reconstruction errors of the validation data, serving as a criterion to distinguish between normal and anomalous data. An innovative method for assessing drill bit degradation was developed, utilizing reconstruction errors and thresholds. These findings highlight the significance of adopting advanced machine learning techniques to optimize drilling performance and enhance operational efficiency within the mining field.

Rotary percussion drills, specifically top hammer types, are essential in resource extraction and exploration. These drills frequently suffer damage at the drill button, necessitating effective decision-making to address issues. Traditionally, on-site operators relied on intuition, but recent advancements promote automated systems to enhance safety and failure detection in challenging mining environments. This study employs Convolutional Neural Networks (CNN) to develop an automated system for detecting drill bit abnormalities. In a prior investigation, a CNN model was established to discern normal functioning and four distinct abnormalities. To enhance generalization performance, researchers endeavored to create a filter using Gradient-weighted Class Activation Mapping (Grad-CAM), an inverse analysis technique. However, challenges arose as the frequency bands contributing significantly to judgments varied among the five identified states. In response, this study adopts a refined approach, constructing a model that categorizes drill performance into two states: normal and abnormal. Leveraging Grad-CAM, the research identifies the frequency band with a high contribution rate to judgment and applies a filter tailored to this specific band. Notably, by training the CNN model to recognize differences in rock types, the resulting model exhibits adaptability to changes in drill bit types, achieving an accuracy rate of 83.3%.