Mining Industries Data Initiative – “MINDI” program
Mining companies collect and store tremendous amounts of data, often without resources and/or knowledge on how to use these data to improve productivity records. This program aims to understand and categorize the different types of data generated in the production process, with respect to its quality, transfer and analysis for increased control and productivity.
The program will start with a three year period, comprising a pre-set number of sub-projects prioritized by the partners. The work-packages will be run step-wise, implementing results of earlier steps in the following sub-projects, or in parallel, depending on the topic of the projects. After the first three year period, the decision will be made for another three year period with a similar arrangement.
To make sure that the full scale program will have the right focus, the program will start with a separate pilot study. The MINDI pilot study was started in January 2016 and will run for 6 months. Please read more about the MINDI pilot study and it’s objectives here.
The MINDI Program was initiated in 2017 and will continue to 2020. Read more about the MINDI program below.
Mining Industries Data Initiative
Short version of MINDI Project Program Definition
Today’s mining systems include a variety of sensors, data extraction and monitoring systems, automation systems, data transmission systems and systems for data analysis and visualization for different categories of personnel. As new machinery is provided with more and more sensing equipment, and as the number of sensors grow with the Internet of things (IoT) and cloud solutions, huge amounts of data of varying type and quality are becoming available.
When information is digitized, it can be more efficiently used, and becomes potentially accessible for refinement and for the creation of new services. However, the utilisation of data for information extraction and refinement is currently far from optimal in the mining sector. Large quantities of data are today stored but not actively used. There is a need for collaborative investigation to take full advantage of current technological and methodological advances in the management of data for the mining industry. One large project would not be enough, since there are many issues, many possibilities and also quite a lot of problems that need to be solved.
We therefore we suggest the “RTC Project Program model” where we solve the most important issues first and then work step by step depending on the parties’ priorities.
The MINDI program shall for mining companies and their machinery and system suppliers, contribute to increased mining productivity by developing knowledge and promoting the application of a data-driven approach e.g. in order to implement more timely and accurate monitoring of processes and sub-processes, with increased quality. The result should be shorter lead-time from event to action and faster access to the accurate data at the right time and place, for a larger part of the organisation.
- The overall objective for the Project program is to increase productivity in a mining operation. Other business objectives are:
- Participating mining companies have developed a strategy and architecture for handling data in their organisation.
- Participating companies have increased their knowledge and understanding of how they should utilise their data resources, which will enable them to make their operations more efficient.
- Richer and continuously updated geological models will give better understanding of the rock and ore body.
- All kinds of stakeholders will benefit from well-described principles for how data can be exchanged and who owns data, business principles, etc. between different organisations.
- All sub-projects are executed according to time and budget plans.
- A pilot for a data marketplace is implemented in a first user case.
- Strategies and architectures for all involved mining companies are described.
- A few sub-processes from an operation are selected and analysed from a data driven approach.
- An analysis of what data is needed and available in order to make the geological models richer and more valuable is performed.
- Bottlenecks and suitable data sources for monitoring are identified and data analysis is performed on data on one bottleneck.
- Analysis of a suitable set of maintenance data from selected equipments is performed.
The participants of the MINDI program have identified a number of possible activities – projects – in line with the purpose and objectives for the program.
Proposed projects are briefly described here, chapter 5.1 to 5.4, and preliminary project plans are attached. The first step for each project will be to engage participants from the MINDI partners and appoint a project manager. Thereafter shall respective project managers go through the proposed project plan and develop it to an complete project plan. The High Level Group must approve each project plan before the project can start. All budgets and time plans for projects must be developed within the limits of the total Project Program budget, see chapter 6 in this document.
Rock Tech Centre has conducted a pilot study, RTC MINDI pilot study, regarding how to benefit from the vast amount of data generated in the mining operations. The utilisation of data for information extraction and refinement is currently far from optimal. One thing that is needed in order to benefit from the potential is to investigate and develop solutions for sharing data between products and systems from different suppliers. The output from the pilot study was a number of project ideas, amongst them the one described her: an investigation into what is needed for sharing and utilising data from many different sources.
The overall purpose is for the suppliers, to share, process, refine and develop vital information that enables mining companies to cut costs and increase productivity. The suppliers will also benefit by being able to create better and more attractive products. The mining companies will get access to better products and information systems that enables to manage and control their processes in a more efficient way.
Since data exchange is fundamental for the entire MINDI program, an important purpose of this subproject is to create a foundation for the remaining subprojects in this program.
It would be of great benefit for the whole mining process if it was possible to collect and analyse data from different sub-processes in order to analyse and compare the outcome from the whole process with what happened in some of the sub-processes.
In summary, it is to collect, tag, sort and connect all data from all different sources, such as drill rigs, loaders, crushers etc. Then integrate, analyse, simplify and store the data for the entire life-of-mine. This should be an automated data management tool with easy-to-use interface and well-connected to mines database. The simplified information about each step of the operation should be easily accessible and it should be possible to cross-reference and sort them by blast, bench, coordinates, geology, fragmentation, dig ability, crushing performance, water level, etc. By this, a solid platform is made for future improvements in any bottleneck; historical data from the same mine in the vicinity of problematic area can be easily extracted and used for the betterment of the problem. Such system will act as a self-learning process that gives feedback of each unit operation in the mine. These feedbacks will be of great value for the next level/bench/pushback in similar geological conditions and can be adjusted to production requirements of the mine.
The overall purpose is to optimise the production process in the mining operation, in order to enable higher quality through optimisation of the production process parameters (e.g. drill hole patterns, Crushing parameters) using machine learning. The project will build a model in order to analyse and understand the relation between early steps in mining process, rock properties and the output from the mine.
Today’s 3D models used in mining are mainly free standing models produced for each discipline as geologists, rock mechanics and mine planners have different needs and applications for their models. These solitary models comprise their individual necessary data set without utilizing all available information within the mine operation. A shared model has the potential to contribute to a more efficient production and planning of e.g. rock reinforcement, production drilling and mineral processing. The current 3D-models often miss to incorporate all available data, reflecting a lack of interdisciplinary communication. A shared and integrated model of a mine and the related ore bodies could substantially improve efficiency through collaboration and communication between geologists, rock mechanics and mine planners, while widening the available data sources for all disciplines.
The purpose is to increase productivity and safety by implementing a multidisciplinary approach on data dissemination with increased communication between production processes using geological, geophysical and rock mechanical data. To achieve this, the project shall identify a so called “Common Earth model” (CE-model) that can be recommended to mining companies. A common model in a mining operation would with great probability lead to increased understanding and knowledge about the geological conditions, for geologists, rock mechanics and mine planners, which in turn should lead to significant better precision in e.g. production planning and planning of rock reinforcement. Furthermore the project shall, for the mining companies, identify and show the business value of the use of such a model.
The purpose is, for mining companies, equipment suppliers and maintenance service suppliers, to build knowledge by identifying and describe the need of new solutions, based on new methods for analysing large data volumes, which can be an enabler for implementing predictive maintenance. Further more will the project identify a user case and develop and test a first version of a solution using new analyse methods for large data volumes. A well-performed predictive maintenance will contribute to increased system availability and user availability.
The project will be executed in four steps with a decision point after every step. Step one will analyse the most important needs, problems and possibilities in the area of predictive maintenance and monitoring and identify a first case. If a case is identified will the project continue with step two.
Step two will include several activities. First identify data sources connected to the selected case, collect data and if necessary develop the structure for the data storage. Then develop an application, based on available analyse tools, for analyse of these data. Third, test the analyse tools on the target data and finally analyse the outcome, utility and advantage of the result. This should be done for one or several of the identified cases.