Front-End Loader Proof of Concept to assess the performance of an autonomous fragmentation analysis system.

Autonomous fragmentation measurement systems designed for shovels have been in operation for nearly two decades, however fragmentation systems capable of performing in the dynamic conditions of the front-end loaders (FEL) were only recently introduced to the market. This paper presents an in-depth evaluation of an automated digital fragmentation measurement system specifically designed to perform under the dynamic operational conditions of the FEL in two of Teck’s mine sites. As early adopters of this novel application, Teck embarked on a Proof-of-Concept (PoC) trial phase, defining a set of key performance indicators (KPIs) to qualify the performance of the digital solution in tough mining operating environments. The KPIs focused primarily on the hardware robustness, quality of image and segmentation, both day and night, availability of the system and sample ratio. The benefits expected from using this technology include increasing the safety of the current fragmentation measurement process, improving the quality of the analysis, increasing the sample representativeness, and using the attributes of the system to create an oversize detection alert to inform the loading operation process. A critical part of the evaluation was verifying the system’s effectiveness on the particle segmentation algorithm as this element often gets overlooked, especially when dealing with the large number of samples autonomous systems can generate; minor misclassifications due to inadequate segmentation are often dismissed as they might not be statistically significant, however systematic misclassifications can cause unacceptable deviation in the data and limit the intended application of the technology. The system’s performance was further validated through several tests that involved identifying objects of known sizes to ensure the reliability on the consistency of the particles’ segmentation.