20 September 2018: A new site project (P5-044) is on track to deliver better options for grade control and accurate metal reconciliation.
|Data fusion model for integrated sensor trial|
This article was published in CRC ORE News - September 2018
A new site project is on track to deliver better options for grade control and accurate metal reconciliation. In April 2018, BHP commenced a project with CRC ORE to evaluate a broad range of sensing techniques at their Escondida copper porphyry mine in Chile. The aim of the project is to identify the contribution of each of these technologies on material destination optimisation.
To achieve this, there is a sharp focus on the collection of quality data. A variety of disparate blasting, grade, metallurgical and mining data is being integrated, covering both spatial and temporal realms. BHP will also provide cross belt analyser data to provide an additional validation source in conjunction with flotation feed grade information. This will be supplemented by information from 15 run-of-mine samples which are being collected at site over a 6-month period.
The collected data will be used to evaluate three key hypotheses:
|Hypothesis 1:||The combination of Sodern’s Blast Hole Tool (BHT), Blast Induced Material Movement (BIMM) and dispatch data will allow a precise estimation of copper grade coming into the plant at shovel scale comparable to ShovelSense predictions.|
|Hypothesis 2:||The combination of BHT, BIMM and dispatch data can inform and therefore improve the shovel scale grade measurements algorithm, particularly in highly heterogeneous areas.|
|Hypothesis 3:||Vertical heterogeneity does not have a significant impact in the grade estimation at shovel scale.|
Each of the hypotheses will be evaluated against current grade control practices applied on site to quantify the contribution of each of the technologies on copper grade prediction at shovel scale.
The analysis of the complex and diverse data collected in the trial will require development of a comprehensive data analytics platform.
CRC ORE is building a tool to link the disparate data sets on a common platform to allow the efficacy of the combined technologies to be determined. This dedicated software application is being designed to link the diverse data together, on a blast by blast basis, across the spatial and temporal domains demarcated by the primary crushers. This will allow the hypotheses to be tested in a regimented and user-friendly manner without having to resort to complex spreadsheet analysis.