Blast fragmentation size distribution is one of the most critical factors in evaluating the blasting results and affecting the downstream mining and processing operations in open-pit mines. Image-based methods are widely applied to address the problem but require heavy user interaction and experience. This study deployed a deep learning model Mask R-CNN to develop an automatic measurement method of blast fragmentation. The model was trained using images captured from real blasting sites in Nui Phao open-pit mine in Vietnam. The trained model reported high average precision scores (Intersection over Union, IoU = 0.5) 92% and 83% for bounding box and segmentation masks, respectively. The results lay a solid technical basis for the automated measurement of blast fragmentation in open-pit mines.

Among the slope hazards caused by rainfall, not all of them occur directly during storm washout, and the wetting–drying cycles’ effect on the rainfall–evaporation process is an important cause of shallow slope instability. In this study, taking the slope of the open-pit coal mine dumps in Cam Pha, in the Quang Ninh region of Vietnam, as the research object, we carry out experiments on the physical properties of the rock body under different wetting–drying cycles, as well as numerical analyses. The results show that the wetting–drying cycles significantly affect the physical and mechanical parameters and permeability of the rock body. In the process of the wetting–drying cycle, a transient saturated zone occurs on the surface of the slope, and the range of the unsaturated zone inside the slope body decreases with the increase in the number of wetting–drying cycles. Moreover, the infiltration line keeps moving downward, but the rate of downward movement is slowed down by the decrease in the gradient of matrix suction affected by rainfall. Under the influence of the wetting–drying cycles, the slope displacement, plastic zone, and maximum shear strain increment range gradually approach the slope surface with the wetting–drying cycles, and the displacement peak gradually increases. A dump is a site for the centralized discharge of mining waste, formed by the crushing and stockpiling of the original rock formation. Bang Nau is the name of the dump considered in this study. After multiple rainfall events, the slope stability under five wetting–drying cycles decreases from 1.721 to 1.055, and the landslide mode changes from a whole landslide to a single-step shallow landslide, with a certain landslide risk. It is necessary to strengthen the slope stability as the landslide risk is very high, and it is necessary to strengthen the monitoring and inspection of the slope.

The success of a cement production project depends on the supply of raw materials. Long-term quarry production scheduling (LTQPS) based on resource models is essential to maintain a consistent supply to cement plants. Geological uncertainty is inherent due to sparse exploration data in resource models and significant risk factors for not achieving production targets. This research proposes a stochastic framework for LTQPS that considers the impact of geological uncertainty on raw material supply. A clustering algorithm uses multiple simulated deposit models to aggregate blocks into mining cuts. A new stochastic mixed-integer programming model is formulated with two objectives: to minimise the cost for developing the raw mix and the risk of not meeting production targets. The proposed framework is implemented successfully in a limestone deposit in Southern Vietnam, resulting in an increase of 5 million tons (Mt) and a 30% reduction in unit cost over the deterministic mixed-integer programming model.

Precise and reliable prediction of blast fragmentation is essential for the success of mining operations. Over the years, various machine learning models using artificial neural network have developed and proven to be efficient in predicting the blast fragmentation. In this research, we design multiple-output neural networks to forecast the cumulative distribution function (CDF) of blast fragmentation to improve this prediction. The model architecture contains multiple response variables in the output layer that correspond to the CDF curve’s percentiles. We apply Monte Carlo dropout procedure to estimate the uncertainty of the predictions. Data collected from a Nui Phao open-pit mine in Vietnam are used to train and validate the performance of the model. Results suggest that multiple-output neural network models provide better accuracy than single-output neural network models that predict each percentile on a CDF independently. Whereas, Monte Carlo dropout technique can give valuable and relative reliable information during decision making. Article highlights: • Precise and reliable prediction of blast fragmentation is essential for the success of mining operations. • A predictive model based on the multi-output neural network and Monte Carlo dropout technique was designed to predict the fragmentation CDF curve in the blasting operation of an open-pit mine. • The predictive model was proven reliable and provided better accuracy than models based on a single-output neural network.

Building geological models of stratified mineral deposits is still challenging. In particular, there is a lack of practical and automatic methods for constructing models integrated with specialized stratiform geological disciplines. This study develops an engineering geological structure modeling method for lithological predictions of stratified deposits using exploration borehole data. This approach first constructs engineering features to automatically capture the specific stratigraphic and geometric rules of stratified deposits. They are then fed into Ensemble Machine Learning models for training and optimization. Furthermore, the developed method is employed to predict lithologies of the Ta Thiet limestone deposit, which is extracted for cement production. Compared with the conventional geostatistical method, Indicator Kriging, the machine learning models demonstrated better performances. They also accurately respected the geological shape and distribution of the deposit and efficiently supported the mine planning and raw materials blending for cement production.