Abstract. The following article is a preview of the study and LATAM Benchmark of digital mining operation, with more than 6 medium and large mining operations, in which patterns are detected, linking different levels of operational management and their relationship with decision-making. At the same time, the use of Artificial Intelligence remotely and in the field is discussed. Some repetitive findings and  gaps are presented in different contexts of mining sites and some concepts and solutions for transformation and automation processes connecting it with human teams and how to reduce risks in a dynamic and proactive way.

Abstract. This article is a preview of the study and LATAM Benchmark of digital mining operation, with more than 6 medium and large mining operations, in which patterns are detected, linking different levels of operational management and their relationship with decision-making. At the same time, the use of Artificial Intelligence is discussed remotely and in the field. Some repetitive findings and gaps are presented in different contexts of mining sites and some concepts and solutions for transformation and automation processes connecting it with human teams and how to reduce risks in a dynamic and proactive way.

Latent challenges

Today’s technological advances have made different industries find themselves in a constant process of innovation, and mining is not far behind applying new technologies to face challenges in matters of safety and health, productivity and sustainability, reducing operating costs. , working on energy efficiency, care for the environment and quality of life for workers, among others.

● Despite the incorporation of various technologies to meet these challenges, there are still gaps   in KPIs or key indices of business success. Having good technologies and first-rate professionals, why do they keep happening? Is it possible to generate a change in how the operation is managed?

In relation to the above, INDIMIN has sought to reduce operational gaps through the use of technology and advanced data analysis, transforming processes with a systemic perspective.

For more than half a decade, and in a pioneering way, it has developed the bases to have Artificial Intelligence in mining, digitization with a user-centered design, co-designing in the field and integrating multiple disciplines, such as engineering, data science, science social and design, among others, linking them in areas such as Planning, Processes, Mine and Plant Operations, Predictive maintenance, mineral comminution and recovery.

Integrate and analyze in a pragmatic way, online the millions of data that are generated daily, as well as aim its use to self-management, decrease variability and dynamically learn from the best practices of the same coworkers, are some of the infinite possibilities improvement and transformation.

For the purposes of this chapter, and based on the experience described and what was applied in the field, an analysis focused on Mine Operation has been carried out, taking as a basis more than 2.5 million data, more than 30,000 hours of operation and 2,760 shifts of 6 Medium and large mining sites Mining in Latin America, producing copper, iron and coal, finding important differences in the different indicators that directly influence the productivity of the operation.

The dimensions to be analyzed are diverse. However, in this case we focus on human behavior and its relationship with Safe Operation, Productivity and Efficiency, and on how they strategically impact the day-to-day operation, often going unnoticed operationally in the SIC or short interval control, by not using advanced analysis technologies, focusing on the use of averages for daily operational management.

Urgent and important, but still invisible in day-to-day management

If we delve into the case of loading equipment operators, we have found up to a 31% difference between yields (ton / hr) of operators of the same equipment model (electric shovel) and more than 4,000 hours of loss in operation due to to cycle times that are outside the ranges and objectives established by the mining companies. This means losses of up to 5.5M tons per year per site, equivalent, and this is only the ‘tip of the iceberg’ of more opportunities for improvement.

Heat maps that relate by loading operator the tonnages loaded per truck and the loading time. Compared operators have operated on the same type of material, same loading front, same loading equipment and same number of hours worked

.

Some critical alerts from the study.

From best practices to sustained frustration of people

In addition, a great variability in the performance of operators was observed, reaching differences of up to 68% in set-up times and 100% in unloading times for operators working in the same conditions and locations.

These patterns are not only repeated in operational efficiency, but also in other aspects that are a relevant part of the operation: such as safety and equipment care.

Regarding safety issues, close to 900 overspeeds were identified per operator in one year of operation, which represents a constant concern in the mining industry, where one of its fundamental pillars is to maintain the safety of the operation and avoid incidents that may become fatal.

Likewise, the care of the equipment plays a relevant role to fulfill the planning and obtain the desired reliability. Up to 1,700 overloads were observed by operators in transport equipment, which may mean higher expenses in maintenance, tires, fuel consumption, among other associated consequences.

All the operational gaps presented are added to the low proximity to the operators’ headquarters and the lack of constant visibility in their performance, where they normally must wait for annual evaluations to be aware of them, losing the opportunity to take corrective actions in the field. timely and objective way. In all this there is great loss of value for the business. According to a study by PwC Chile (2016), more than 48% of workers in the mining industry feel undervalued and with low understanding of the objectives. The following sentence obtained in the field, partly reflects the relevance of this area “We do not want to be treated like robots, or be replaced by machines, we want technology to work with us and help us improve …” asserts a person with a long history in large mining..

Beyond the technical challenge

A relevant question could be: How do we close these gaps and empower our human team? Is this possible to do?

Other relevant findings of the PwC Chile study (2016), which covers more than six medium and large open pit mining sites, have to do with frustrations, collaboration and the objectification of the results in the human team.

Without going any further, the phrase “we do not want to be treated like robots, we want technology to help us to be better”, deeply reflects the feeling of a human group that is not necessarily supported in their development and job satisfaction, both in their daily work , as in its development and obtaining constructive and objective feedback.

Faced with this problem, INDIMIN creates Smart Mining Coach, an intelligent and personalized digital assistant, for productivity in Mining, which allows a complete view, online and remotely, of what is happening with the main KPIs of people, processes and assets of the operation, providing live and timely information for a better and faster decision-making process, promoting the development of people, their learning, motivation and performance oriented  in results and clear objectives.

This digital assistant detects online patterns and operational deviations, predicting what will happen in the operation, delivering recommendations enabled with artificial intelligence that anticipate and suggest optimized actions that seek to prevent risks and losses of operations and increase productivity, working with incentive , safety and reliability.

Smart Mining Coach works on the path of process transformation towards the digitization and automation of processes and operations, where Artificial Intelligence plays a key role. It focuses on collaboratively empowering, discovering, and reinforcing desired behaviors, best practices, and superior goals for sustained operational excellence. It has customized modules to support the team in the operation and in remote centers, such as Managers, Superintendents, Remote Control Centers, Shift Managers, Dispatchers, Instructors and Operators.

Mining, Friendly and Actionable Artificial Intelligence?

Bringing the complex to the simple and everyday

Multiple technologies are available today. However, for various reasons, its use ‘dies’ or is diluted by users in the field or remotely. Why? The so-called Technology Push helps to Transfer new Technologies from the laboratory or academy, however, in the context of continuous operation, either in the field or remotely, not all the factors influencing innovation are considered, to focus, co-design, iterate and lead continuous improvements that encourage users to be the protagonists of the generation of value with maximum efficiency and productivity.

A real example of how Artificial Intelligence has been adapted to be able to accompany the different key users of the operation in the field and remotely, is the production prediction model at the end of the shift, developed by INDIMIN. The challenges of how confidence is achieved in everyday use and carried out in an actionable way are remarkable insofar as they have been co-designed in such a way that it encourages and encourages its use, making AI transparent, friendly and motivating to encourage automation, increasing assertiveness and value in decision making.

It is important to note that a few years ago, the first versions of these models had accuracy of approximately 80%, but it was exceeded more than 94% by integrating more sources of information, combination and research of different AI algorithms, which also invisibly they integrate areas and break down silos of operation for better decision making.

Prescription capabilities have been added to these models. That is, find the reasons for where the operating losses, high variability and impact are to suggest what can be done to improve, since the reasons behind to achieve a reliable and safe operation are dynamic and multivariable.

Usability, a transformation challenge

This model indicates to the different users, such as Remote Centers, Shift Managers, or Superintendents during the first hours of the shift, and with up to 94% accuracy, what the production level will be at the end of the day, based on historical data. of the key variables of the operation.

It has been demonstrated that the machine learning model is able to predict the end of the shift several hours before, at the beginning of the shift and with greater precision than a linear projection model, allowing advance decisions to be taken to improve the management and productivity of processes, operations, assets and people.

Additionally, the models dynamically detect gaps in efficiency and operational losses, delivering recommendations online and remotely, to take action on them at the right time. For this, combinations of neural networks have been used to predict and, at the same time, genetic algorithms to optimize and achieve searches of what to do to improve.

This prediction of what will happen and recommendations of what to do to improve productivity are hosted in Smart Mining Coach, showing the results in a simple and easy-to-understand way for the end user, in order to leverage the use of technology. by people. The key in this is to achieve trust in the end users. In other words, AI shows better results, in a simpler, more didactic and sustainable way. Analogous to navigation in route maps, such as Google Maps or Waze, the algorithm is only relevant for the final results, but the safe and timely interface and navigation is the same or more relevant from the point of end users, being the AI something that happens to work in the background for the end user.

● Based on the above, Smart Mining Coach has achieved successful results in the usability of the tool, reaching up to 90% loyalty of end users and 8.3 average sessions per user per day. These results are given thanks to the simple and easy-to-use interface, co-designed with the end users, in order to leverage usability and achieve concrete and high-impact actions.

In the case of operators, there is the Coach Operator module, a mobile application for the operator of mining equipment that was co-designed with operators, in response to the need for a simple tool that functions as a direct communication channel between different areas. of the business and the operator in a bidirectional way. An important point to mention is that from the same interest of operators and unions the Coach Operator is considered as a practical, faithful and motivational tool regarding the measurement of results and objectives in a context where technology is viewed as an asset and critical support for the development of people and the organization, and not as a threat or competition.

In Coach Operator, all the relevant information is made available on a daily basis: such as your assignment, security notices, access to relevant documents, goals and objectives of the shift, added to feedback and personalized interaction shift by shift according to the performance patterns  found by the operator in its main operation and safety indicators, which range  from alerts for poor performance, to acknowledgments, congratulations and motivational messages regarding good  performance, in order to promote learning, self-management, and the empowerment that ultimately lead to better operational performance.

The existing data in the mining companies gives us the opportunity to change the way we do things, question, correct and improve our processes, if we can apply the new analysis capabilities that support the management of the operation and online decision making in a timely and reliable manner. What do we expect to transform our mining processes hand in hand with people and new technologies?

Bibliographic references:

• PwC Chile, 2016, “Modelo de Medición y Evaluación de Factores de Productividad del Capital Humano”.