4^th AMM-2030 Conference: On the Road to Autonomy!

Presentation of the conference programme

As part of the Industry 4.0 and digital transformation, the mining industry is striving to increase productivity to meet the growing demand for metals, while facing challenges such as labour shortages and climate change. Technological advances such as predictive maintenance and autonomous vehicles are transforming the way mining is done, but there are barriers to their adoption.

The presentation will focus on the results of a research project that has highlighted the factors which facilitate and hinder the adoption of innovative and sustainable technologies when evaluating private mining investment projects. It will also look at the impact of responsible finance and ESG transparency on investment decisions in the mining sector. This project is based on a synthesis of current knowledge about the benefits and costs of adopting various innovative technologies in the mining industry, and on around twenty interviews in Quebec with key players of this ecosystem: equipment manufacturers, mining companies, investors and the government.

Mining companies are increasingly integrating new digital technologies into their mining operations and in all stages of the mining cycle. These include the Internet of Things (IoT) and artificial intelligence. These changes are having an undeniable influence on the skills of the human resources responsible for running the industry’s operations. To ensure a better correspondence between jobs and training, it is essential to analyze which skills are required by new technologies in order to adapt the training to the needs of the sector. This presentation reviews the main types of skills needed for Industry 4.0, with a particular focus on those required for mining operations. The results of two case studies carried out as part of the AMM-2030 project will be presented as examples. The first case concerns the Internet of Things technology aimed at improving the intelligent monitoring of a mining ventilation system. The second technology analyzed concerns an integrated management system for mining data (Neuromine), enabling optimal decision-making. There is a certain multidisciplinarity for most of the positions where digital knowledge is increasingly required. The presentation will conclude with a number of recommendations arising from these case studies.

As part of the AMM-2030 project, MISA worked with Cadence Consultants to map over 125 mining-related processes. These processes were modelled using the Pix4 cloud application. This is the first phase in the implementation of operational intelligence in the mining sector. Understanding all our operational processes is the starting point for optimizing them.

Mineral liberation is a key indicator of the performance of the grinding process and can account for performance problems during separation. The literature shows that its use in a control context is financially advantageous. The degree of liberation is typically determined by scanning electron microscopy combined with X-ray spectrometry (SEM/XRS). However, the time required to prepare and process samples using this method severely limits its application for daily monitoring in factories. Laser-induced breakdown spectroscopy (LIBS) is used to measure liberation and to overcome the problem of sampling frequency, albeit at the expense of analysis resolution. The size of the beam makes it impossible to distinguish particles smaller than 50 µm. Combining these two analysis techniques with the data reconciliation method yields optimal results, i.e. an estimate of the degree of liberation at a fast rate and with high resolution. Three cases will be presented. The first is the baseline scenario using raw SEM/XRS and LIBS data. The second takes into account SEM/SRX and LIBS measurements along with the results of a liberation model based on a beta distribution. Finally, the last case examines a recursive application using lower weighting for the SEM/XRS data and the model in order to avoid estimates drifting over time.

In order to plan more accurately and improve the efficiency of short- and medium-term mining operations in underground mines, it is crucial to have detailed knowledge of the travel times of ore haul trucks – an energy-intensive activity on which other mining tasks usually depend.

However, despite the critical nature of this transport operation, it is difficult to estimate precise travel times when GPS access is not possible, operational conditions vary and new routes are created as the mine develops.

To tackle this problem, the mining industry has invested in the installation of underground vehicle detection beacons, the data from which could be used to trace and time travel routes. Exploiting this data, possibly in combination with other telemetry data, through machine learning algorithms is a promising avenue.

However, the way these beacons are managed and the detection data is collected make data cleanup a titanic task, even though it is necessary for its exploitation.

The predictions obtained after a superficial cleanup could therefore become much more accurate if the management and maintenance of these beacons were conducted more seriously.

My research work has enabled me to formulate recommendations for the implementation of a detection database that can be exploited by algorithms significantly simpler and faster, and producing high-quality results.

During this presentation, the following points will be discussed:

- Connected fleet management tools;

- Interconnections between systems (access control/ERP/computer-assisted maintenance management);

- Case study presentation (GHG and fuel consumption reduction);

- Collaboration with MISA: Car-sharing - Fleet sizing using AI - Customizable and dynamic profiles.

Oil analysis is a practical and useful tool that enables mines to monitor the condition of their motorized equipment, the degradation of the oil and the wear and tear of the various components. This monitoring method can be seen as a short-term tool. After years of use, mines have acquired a significant amount of information. Even if these databases contain input errors or are incomplete, they reflect the wear and maintenance of a piece of equipment, or even a mining site. By analyzing the quality and consistency of these databases, it is possible to ensure more effective maintenance of equipment.

The first objective of this research is to explore in depth the oil analysis database of a Canadian mine, in order to better understand oil degradation in relation to equipment, oil type and components. As oil does not break down, the aim is to develop an equipment degradation model to better predict the time between oil changes. The result should be presented in the form of a time window enabling the additional costs generated by oil changes carried out outside these periods to be estimated.

In the second stage, this research will attempt to implement this time window in a new optimizer software for maintenance cycles.

This research is part of a broader project to digitize maintenance tools in the mining industry.