Real-Time Mining is a research and development project within the European Union\'s Horizon 2020 initiative and consists of a consortium of thirteen European partners from five countries. The overall aim of Real-Time-Mining is to develop a real-time framework to decrease environmental impact and increase resource efficiency in the European raw material extraction industry. The key concept of the research conducted is to promote a paradigm shift from discontinuous to a continuous process monitoring and quality management system in highly selective mining operations. The Real-Time Mining Control Cockpit is a framework for the visualization of online data acquired during the extraction at the mining face as well as during material handling and processing. The modules include the visualization of the deposit-model, 3D extraction planning, integrated data of the positioning-system as well as the visualization of sensor and machine performance data. Different tools will be developed for supporting operation control and optimized decision making based on real-time data from the centralized database. This will also integrate results from the updated resource model and optimized mine plan. The developed Real-Time Mining cockpit software will finally be integrated into a wider central control and monitoring station of the whole mine.

The project Viable Alternative Mine Operating System (¡VAMOS!) develops a new safe, clean and low visibility mining technique for excavating raw materials from submerged inland mines. During operations, the perception data of the mining vehicle can only be communicated to the operator via a computer interface. In order to assist remote control and facilitate assessing risks a detailed view of the mining process below the water surface is necessary. This paper presents approaches to real-time 3D reconstruction of the mining environment for immersive data visualisation in a virtual reality environment to provide advanced spatial awareness. From the raw survey data a more consistent 3D model is created using postprocessing techniques based on a continuous-time simultaneous localization and mapping (SLAM) solution. Signed distance function (SDF) based mapping is employed to fuse the measurements from multiple views into a single representation and reduce sensor noise. Results of the proposed techniques are demonstrated on a dataset captured in an submerged inland mine.

The application of sensor technologies for raw material characterization is rapidly growing, and innovative advancement of the technologies is observed. Sensors are being used as laboratory and in-situ techniques for characterization and definition of raw material properties. However, application of sensor technologies for underground mining resource extraction is very limited and highly dependent on the geological and operational environment. In this study the potential of RGB imaging and FTIR spectroscopy for the characterization of polymetallic sulphide minerals in a test case of Freiberg mine was investigated. A defined imaging procedure was used to acquire RGB images. The images were georeferenced, mosaicked and a mineral map was produced using a supervised image classification technique. Five mineral types have been identified and the overall classification accuracy shows the potential of the technique for the delineation of sulphide ores in an underground mine. FTIR data in combination with chemometric techniques were evaluated for discrimination of the test case materials. Experimental design was implemented in order to identify optimal pre-processing strategies. Using the processed data, PLS-DA classification models were developed to assess the capability of the model to discriminate the three material types. The acquired calibration and prediction statistics show the approach is efficient and provides acceptable classification success. In addition, important variables (wavelength location) responsible for the discrimination of the three materials type were identified.

This paper presents the development and implementation of a theoretical mathematical-statistical framework for sequential updating of the grade control model, based on a support vector machine learning algorithm. Utilising the Zambujal orebody within the Neves-Corvo Cu deposit in Portugal, parameters that can be measured in real time, used in visualisation, modelled for resource estimation, and used for process control visualisation and optimisation are considered. The methodology broadly comprises of three steps. Firstly, the provided dataset is used to develop a virtual asset model (VAM) representing the true 3D grade distribution in order to simulate the mining method. Then ore quality parameters are established simulating real time monitoring sensor installation at: (a) stope development and rock face monitoring (face imaging and drillholes); and (b) transport monitoring (muck pile, LHD/scooptram). Next, the acquired data was assimilated into the models as part of the sequential model update. Two different mining methods and the monitoring information that can be acquired during the ore extraction are analysed: (a) drift and fill mining and (b) bench and fill mining, which are widely implemented at the Neves-Corvo mine. Selected study zones were chosen such as to contrast mining through the high/low grade zones with different degrees of heterogeneity, which demonstrate the performance of resource estimation and classification models developed in heterogeneous mining stopes. The grade accuracy and error in the resource model, and high/low grade ore classification accuracy and error are evaluated as performance metrics for the proposed methods. In drift and fill mining, drillhole and face sampling data collection was simulated in a real-time manner and fed into the support vector machine (SVM) regressor to update the resource estimation model in both a high grade and low grade drift scenarios. In each scenario, six drift and fill mining steps were simulated sequentially and the posterior resource models, after integrating real time mining data, have shown significant improvement of bias correction in both updating planned resources and reconciling extracted ore. In bench and fill mining, grade classification based on random sampling data from muck pile was demonstrated, considering scoop by scoop derived monitoring data. Three different classifiers (mean, median, and Bayesian) were tested and shown very good performance. In the case study presented here, a sequence of 15 blasting steps was simulated with each step requiring 112 scooping operations to transport the blasted ore. Using the real time monitored information, it was shown that at each blasting step over 85% of the scoops can be labelled correctly using the proposed methods and with an accuracy of over 95%.

This research is part of the European Union funded Real Time Mining project, which aims to develop a new framework to reduce uncertainties during the extraction process in highly selective underground mining settings. A continuously self-updating resource/grade control model concept is presented and aims to improve the raw material quality control and process efficiency of any type of mining operation. Applications in underground mines include the improved control of different components of the mineralogy and geochemistry of the extracted ore utilizing available “big data” collected during production. The development of the methodology is based on two full scale case study, the copper-zinc mine Neves-Corvo in Portugal and Reiche-Zeche mine in Germany. These serve for both, for the definition of method requirements and also as a basis for defining a Virtual Asset Model (VAM), which serves for artificial sampling as benchmark for performance analysis. This contribution introduces to the updating concept, provides a brief description of the method, explains details of the test cases and demonstrates the value added by an illustrative case study.

Efficient data collection, analysis and preservation are needed to accomplish adequate business decision making. Long-lasting and sustainable business operations, such as mining, add extra requirements to this process: data must be reliably preserved over periods that are longer than that of a typical software life-cycle. These concerns are of special importance for the combined on-line-on-mine-real-time expert system SOLSA (http://www.solsa-mining.eu/) that will produce data not only for immediate industrial utilization, but also for the possible scientific reuse. We thus applied the experience of scientific data publishing to provide efficient, reliable, long term archival data storage. Crystallography, a field covering one of the methods used in the SOLSA expert system, has long traditions of archiving and disseminating crystallographic data. To that end, the Crystallographic Interchange Framework (CIF, [1]) was developed and is maintained by the International Union of Crystallography (IUCr). This framework provides rich means for describing crystal structures and crystallographic experiments in an unambiguous, human- and machine- readable way, in a standard that is independent of the underlying data storage technology. The Crystallography Open Database (COD, [2]) has been successfully using the CIF framework to maintain its open-access crystallographic data collection for over a decade [3,4]. Since the CIF framework is extensible it is possible to use it for other branches of knowledge. The SOLSA system will generate data using different methods of material identification: XRF, XRD, Raman, IR and DRIFT spectroscopy. For XRD, the CIF is usable out-of-the-box, since we can rely on extensive data definition dictionaries (ontologies) developed by the IUCr and the crystallographic community. For spectroscopic techniques such dictionaries, to our best knowledge, do not exist; thus, the SOLSA team is developing CIF dictionaries for spectroscopic techniques to be used in the SOLSA expert system. All dictionaries will be published under liberal license and communities are encourage to join the development, reuse and extend the dictionaries where necessary. These dictionaries will enable access to open data generated by SOLSA by all interested parties. The use of the common CIF framework will ensure smooth data exchange among SOLSA partners and seamless data publication from the SOLSA project.

Short-term mine plans are the key operational basis for ore production targets ranging from shift to weekly or monthly targets. Short-term plans cover detailed operational subprocesses such as development, extraction and backfill schedules as well as materials handling and blending processes. The aim is to make long-term goals feasible by providing a constant plant feed that complies with quality constraints. Short-term mine planning highly depends on the accuracy of the resource model as well as the current production status and equipment fleet. Most of these parameters are characterized by uncertainties due to a lack of information and equipment reliability. At the same time, concentrate production and quality must be kept within acceptable ranges to ensure productivity and economic viability of the operation. Within the EU-funded Real-Time Mining project, the reduction of uncertainty in mine planning is carried by using real-time data. Ore and rock characteristics of active faces and equipment data are iteratively integrated in a simulation-based optimization tool. Therefore, predicted processing plant efficiencies can be met by delivering constant ore grades. Hence, a constant concentrate quality is ensured and long-term targets can be fulfilled. Consequently, a more reliable exploitation plan of the mineral reserve is facilitated.

This paper proposes a system for the ingestion and analysis of real-time sensor and actor data of bulk materials handling plants and machinery. It references issues that concern mining sensor data in cyber physical systems (CPS) as addressed in O’Leary et al. [2015].

Recent development of hyperspectral snapshot cameras offers new possibilities for ore analysis. A method for generating a 3D dataset from RGB and hyperspectral images is presented. By using Structure from Motion, a reference of each source image to the resulting point cloud is kept. This reference is used for projecting hyperspectral data onto the point cloud. Additionally, with this work flow it is possible to add meta data to the point cloud, which was generated from images alone.

In mining operations, the time delay between grade estimations and decision about the scheduling of stopes mining can result in seriously outdated information and, consequently, a substantial mined reserves bias. To mitigate this gap between the grade estimation of an orebody and its exploitation, this paper proposes a new method of speedily updating resources and reserves integrated into the concept of real-time mining. This consists in the continuous and swift update of mine reserves, which requires a continuous and fast stream of the measurements of stopes in an underground mine rather than the chemical lab analysis of core samples or chip/face samples. Here we propose using portable for the swift monitoring of ore grades. However, this “fast” data be highly uncertain. For this reason, the first step consists of creating a bidistribution function between “uncertain” XRF and the corresponding “hard” measurements, based on empirical historical data. Following this, the uncertainty of the XRF measurements is derived from those bi-distributions through the conditional distribution of real values given to the known XRF measurement.The second step involves updating the reserves by integrating this uncertain XRF data, which has been quantified by conditional distributions, in the grade characterization models. For this purpose, a stochastic simulation with point distributions is applied. A case study of a sulphide copper deposit illustrates the proposed methodology.