Artificial intelligence startup KoBold Metals uses artificial intelligence to comb through historical and scientific data to identify undeveloped mineral deposits. The company said this year that it had discovered a huge copper deposit in Zambia. KoBold said the $2 billion Mingomba mine will produce at least 300,000 tons of copper per year starting in the 2030s. Kurt House, co-founder and CEO of KoBold, said about 40% of the new funds will be used to develop existing projects into mines, of which the Zambian copper project will receive the “largest share.” The company uses OpenAI’s generative artificial intelligence technology as well as more traditional artificial intelligence technology. Ouse said KoBold plans to actively recruit and add data scientists with more traditional technical backgrounds to the team, as well as geoscientists who survey potential deposits and collect data.
Application Principles of AI in Mining
Mining exploration and development itself is a professional job of collecting data and knowledge-driven analysis of data. It is a process of collecting data, analyzing data with knowledge, forming prospecting and exploration conclusions, building geological models of mineral deposits, and optimizing mining plans.
From the perspective of data collection, it is to collect information through various means such as geology, geophysics, geochemistry, remote sensing, etc., and collect geological information data through exploration, trenching, drilling and other working methods to lay the foundation for data analysis for prospecting. Knowledge-driven means processing various geological data and influences based on the experience of experts in the field of geology and mining, and establishing a prospecting model to guide the exploration direction to analyze the location, size, and probability of discovery of the ore deposit, and display it through mapping.
AI is good at processing large amounts of data, analyzing geological surveys, satellite images and historical exploration data, and quickly establishing a visual geological mine model. Through machine learning models such as neural networks, patterns, anomalies and potential ore deposits that may not be recognized by traditional exploration methods can be quickly determined, thereby achieving a breakthrough in prospecting. In the field of mining, the role of AI is to replace manual labor. At present, it has been applied in prospecting, geological data and image processing, three-dimensional geological modeling, and disaster warning.
The Application of AI in Mining Has Achieved Initial Results
In terms of prospecting and exploration, AI systems can identify geological anomalies and mineralization characteristics and predict the location and scale of underground deposits. In terms of mining production, AI-based sorting systems can quickly identify valuable minerals in waste rock in real time, improving the recovery rate and utilization rate of mineral resources. At the same time, AI technology can achieve all-round monitoring of the mining production environment by integrating technologies such as video monitoring, intelligent recognition algorithms and big data analysis platforms, and timely discover and deal with unsafe factors.
In terms of data image processing, AI can quickly and accurately complete tasks such as data classification, data induction, and data feature extraction, such as three-dimensional inversion of geophysical data, automatic geological mapping of high-resolution remote sensing images, and three-dimensional geological modeling of mines.
In terms of geological disaster warning, AI is combined with remote sensing technology and Internet of Things devices to monitor geological environmental changes such as landslides, earthquakes, and groundwater level fluctuations in real time. Through rapid identification and intelligent analysis of abnormal data, the probability and possible impact range of natural disasters such as earthquakes, landslides, and mudslides can be more accurately predicted.
AI Technology Has Obvious Advantages Over Traditional Mining Exploration and Development Methods
It has effectively improved the accuracy and success rate of prospecting and exploration: As the current shallow and easy-to-mine mineral resources are gradually exhausted, mining exploration work has become more complex and difficult. At present, the exploration of traditional resources such as copper, gold and cobalt is already very difficult, with a success rate of only 0.5% for exploring new minerals, and more than 99% of conventional exploration projects have failed to become mines. Compared with traditional prospecting, AI technology can quickly and accurately process and analyze large amounts of data, greatly improve exploration efficiency, reduce the scope of exploration, delineate mineralization prospecting areas in greenfield exploration projects with relatively scarce data, and discover some prospecting information that has been ignored by traditional theories or methods.
Greatly reduced the manpower and time costs in the mining development process: In the mining exploration and development process, from the initial consultation of a large amount of literature, collection of field rock sample information, geological mapping, to processing physical and chemical remote drilling and other data, delineating the prospecting target area, establishing a three-dimensional geological model of the mine, and then to the production stage of mining, transportation, ore sorting, smelting, etc., as well as comprehensive production management, mine disaster warning, etc., each link requires professional knowledge and technical personnel to spend a lot of time and energy to complete, and the labor and time costs are huge. AI has efficient computing and information processing capabilities, and can automatically extract, integrate, analyze and model data, realize intelligent control of mining, ore dressing, and smelting equipment, greatly reduce labor and time costs, and improve production efficiency and resource utilization.
Greatly optimized all aspects of mining industry production management and reduced safety risks: In all aspects of mining production, the limitations of many factors such as human resources, equipment use, and material supply inevitably affect the production and resource utilization efficiency of mines. AI technology can accurately identify production bottlenecks and their causes through the comprehensive analysis of historical and real-time monitoring of mine production data, realize intelligent decision-making and control, and then promote the optimization and adjustment of mining production links, improve production management capabilities and resource utilization efficiency. At the same time, AI technology can realize mine safety monitoring and risk warning through various sensors and monitoring devices in mines, identify safety hazards and abnormal behaviors, and take early warnings and corresponding safety measures to reduce the occurrence of accidents.
AI Technology Still Has Many Bottlenecks in the Short Term
Geological data is difficult to obtain and process: the earth has a long history of geological evolution, and mineralization is an intricate geological process. Geological exploration and development data has the characteristics of heterogeneity, multi-scale, amorphous, and discontinuous. The biggest application premise of AI technology is big data, such as deep geological information, and a large amount of geological data, production data, equipment data, etc. involved in the exploration and development of mines with complex production environments. There are certain difficulties in obtaining and processing, and currently obtaining these data still relies on traditional means.
Data quality and result reliability need to be improved: At present, the intelligent application of AI in mining exploration and development is highly dependent on high-quality big data samples. In the process of collecting and collating actual geological data, such as the impact of thermal disasters and rock bursts during deep exploration, environmental noise and interference during mine development, and data missing and errors caused by human collection and collation of geological data, all will affect the quality of the data and lead to reduced accuracy and reliability of AI analysis results.
