Abstract:

Fleet Management System (FMS) plays a crucial role in large mining operations. From second-by-second data recording to making real-time near-optimal decisions on dynamic scheduling and dispatch of mobile mining equipment. We present a workflow to proactively turn historical mine-dispatching operational data into knowledge for better short-term mine planning decisions. The main objective of establishing this feedback methodology from operational data into short-term planning practice is to allow for timely adjustments within production planning practices. In addition, the feedback loop from dispatch's historical data into the planning practice ensures future short-term mine planning results are consistently within acceptable uncertainty limits established by historical operational data. We focus on establishing practical workflows by comparing what has been achieved with how to continually refine short-term planning practices to improve the alignment of planned versus actuals.

We use one year of recorded dispatch data to guide our study in establishing the best practices for short-term mine planning with a bi-weekly resolution. We use clustering to group recorded dig and dump locations in the database into mining polygons allocated to shovels and centroids of waste dump locations throughout the planning time horizon. We used the haulage road profiles to estimate the loaded and empty haul cycle times based on rimpull curve calculations, which usually overestimate the fleet productivity. Furthermore, to correct the rimpull curve estimation error - we present a modified haulage cycle-time estimation method - which honours the recorded speed distributions of the loaded and empty hauls with more accurate haulage cycle time estimation. The established workflow allows the planner to assess the impact of changing operational scenarios such as stockpiling, different sizes of mixed-fleet trucks, and introducing new haul roads into the mine plan in the presence of operational uncertainty.

Menu

Shahrokh Paravarzar