Operations Data
In the last decade, a variety of sensors and smart chips have been integrated across mining operations. All those sensors and connected devices generate a massive amount of data in real time. The variety of generated data includes geological data, equipment monitoring data, operational performance data, environmental data, among others. An avenue of opportunities has been created with the latest development in Wi-Fi, 4G, and 5G technologies. Drone systems are applied for a variety of activities, such as to measure geotechnical stability, mining mapping, tailings dam monitoring, among others. Real-time insights can be delivered through high-performance computers, artificial intelligence, and machine-learning advanced models.
When working with sensors, standards and procedures must be clearly defined and implemented. Defects or anomalies in instruments can provide inaccurate measurements, produce errors into downstream calculations, impact process control efficiency, and affect user's confidence in information systems [35].
Data provided by FMSs are used to calculate strategic KPIs such as availability, utilization, effective utilization, production, productivity, unproductive time, productive time, mean time between failures, and mean time to repair. In Ref. [36], the author reviewed the consistency of an ordinary FMS extracting a four-month dataset. Haul truck activities' cycle times were assessed, and the samples indicated an inconsistency rate of 21% in the haulage time. Shovel loading times presented data quality issues in 29% of the records. It is essential to mention that the FMS considered is this research is configured to receive manual inputs for most of the assignments, especially cycle time stages (loading, hauling, queuing, dumping, returning, maneuver), maintenance states (corrective and preventive), and unproductive time (shift change, refueling, meals, road maintenance, etc.). This limitation naturally implies a higher tendency for errors when compared with more sophisticated systems developed to detect the majority of trucks and shovels activities automatically.
A similar conclusion was previously reported by Hsu 2015, which assessed datasets from two different FMS vendors in two separate open-pit mine sites. The primary quality problem observed was inconsistent labeling (assignment) of activities to time categories. In terms of quantitative assessment, both systems presented a surprisingly high percentage of concise duration stages, which suggests either data corruption (software/hardware) or human errors (operator input). Additionally, the research identified a mismatch in the maintenance intervention records. All quality issues reported above can potentially undermine any data analytics initiative or continuous improvement project.
A general trend for mining operations data management in the BD era is the integration of different data sources and systems from geological mapping and block model update routines to transportation and disposal of ore (crushing plant and stockpiles) and waste (waste dump). The current dispersed approach loses the value of connecting mine planning software, which contains information about the ore body, FMSs that monitor operational performance and assets' health, enterprise resource planning, which are the general repository of company's information. However, despite all the advantages obtained by enhancing BD capabilities in the mining industry, some topics require further discussion. Data privacy, security, and archiving are significant issues that need serious consideration [37,38].
