01 / Reclaim
ROM stockpile reclaim
Decide which ROM stockpiles to reclaim from and when, with stockpiles categorised by properties such as grade, arsenic and talc content.
Customer case study · Gold
Gold Mill Feed Scheduling Optimisation
BlendOpt is a configurable optimisation platform for mining value chains. Configured for gold mill feed scheduling, it uses mathematical optimisation to decide which ROM stockpiles to reclaim from, what to send to the crusher and mill each day, and how to blend it, so feed stays within grade constraints and crusher and mill capacities.
Customer resultGold mill feed schedule
Powered by BlendOpt
Outcome
30+ recovery calculations
Moved from a complex Excel workbook into a BlendOpt digital twin of the crusher, mill, stockpiles and recovery.
Human planning errorEliminated
Recovery chartsHIG, Float, CIL
ScheduleDaily, one month
Gold · mill feedCase study below
What BlendOpt optimises
What BlendOpt optimises in gold mill feed scheduling
Recovery calculations are modelled in readable formulas inside a digital twin of the operation, replacing complex planning spreadsheets that are hard to check and maintain.
02 / Schedules
Crusher and mill feed schedules
Suggest the best crusher and mill feed schedules based on available inventory and predicted variability.
03 / Grade
Grade constraints
Keep mill feed blends within grade constraints so the plant receives feed it can process as planned.
04 / Capacity
Crusher and mill capacity
Respect crusher and mill capacities calculated from planned outages in every period of the schedule.
05 / Recovery
Recovery calculations
Model recovery formulas, including HIG, Float and CIL calculations, in human-readable notation that is easy to check and maintain.
06 / Scenarios
Scenario analysis
Run a range of scenarios to evaluate different operating conditions and compare the resulting schedules in interactive charts.
Customer results
Gold mill feed scheduling results
Results BlendOpt customers have measured in mill feed, stockpile and reclaim decisions.
30+ calculations
Excel recovery calculations replaced by a BlendOpt digital twin of crusher, mill, stockpiles and recovery
Read the gold case study below →
Cu:S 0.4 → 0.03
copper mill feed variability after new ROM stockpile definitions and optimised reclaim
Copper mine-to-mill optimisation →
+$24M / yr
forecast revenue from new stockpile definitions and a moderate capacity change
Stockpile optimisation case study →
The whole value chain
Gold mill feed in the context of the whole value chain
What reaches the mill depends on the ROM ore the mine plan delivers, which stockpiles it is sent to and when it is reclaimed.
BlendOpt takes the mine plan as an input and models ROM stockpiles, crusher, mill and recovery together, so a mill feed decision is judged by its effect on the whole operation rather than one day in isolation. See how this works in BlendOpt configured as gold mining software, across metals mine-to-mill optimisation, for copper mill feed and in integrated mine planning and scheduling.
Gold mill feed value chainOne schedule. No spreadsheet.
BlendOpt model
- Mine planROM: direct crush or stockpile
- ROMgrade, arsenic & talc
- Crusherplanned outages
- Millfeed blend
- GoldHIG, Float & CIL recovery
Optimisation objective
More gold ounces within grade and capacity limits
Mill feed blends are controlled across the schedule while complying with grade constraints and crusher and mill capacities.
Customer case study · Gold
Case study
Gold mill feed scheduling
30+ recovery calculations moved from Excel into a BlendOpt digital twin
A gold mining company replaced its complex mill feed scheduling spreadsheet with a BlendOpt digital twin of its crusher, mill, stockpiles and recovery calculations.
- 30+recovery calculations
- Dailyschedule over one month
- POCproof of concept first
Customer outcomes
- Prescriptive analytics
- Optimised gold recovery
- Optimised crusher and mill utilisation
01
The challenge
A gold mining company was using a complex Excel spreadsheet to schedule ROM stock levels, mill feed blends, throughput and recovery on a daily time granularity for one month.
Technical services believed that the spreadsheet had become overly complex, hard to interrogate assumptions, easy to introduce errors and generally difficult to maintain, and was looking for an alternative. Ore that was not directly crushed was sent to ROM stockpiles categorised by various properties including grade, arsenic and talc content. Crusher and mill capacities were calculated based on planned outages. The operation had an overarching aim to increase gold ounces and expected improvements were possible through better control over feed blend conditions.
Over 30 recovery-related calculations were embedded within the production planning Excel workbook, which were difficult to interpret and maintain, and technical services was losing confidence in the resulting decisions emerging from the workbook. Furthermore, ROM stockpiles modelled within the workbook resulted in negative values due to the optimisation process generating infeasible plans.
Technical services assessed a number of options for replacing their spreadsheet and chose the BlendOpt value-chain optimiser to construct a digital twin of their operation. This decision was made following a proof of concept (POC) project which demonstrated BlendOpt’s combination of prescriptive analytics and mathematical optimisation was well suited to solving their gold mine-to-mill planning problem.
02
The solution
A digital twin of the customer’s operation was modelled within BlendOpt, including mill, crusher, stockpiles and recovery calculations. BlendOpt was used to run a range of scenarios evaluating different conditions.
03
The value unlocked
- BlendOpt eliminated human error in planning decisions. When combined with more frequent and rapid data updates in BlendOpt, this resulted in greater trust and confidence in planning forecasts.
- Formula modelling within BlendOpt is human-readable, expressed as it would be if the formulas were written down in the end user’s preferred notation. This improved reliability and confidence in planning decisions by introducing maintainable and easily understood recovery calculations.
- BlendOpt suggested the best crusher and mill feed schedules to users based upon available inventory and predicted variability, while complying with grade constraints and crusher and mill capacity constraints.
- BlendOpt’s prescriptive analytics also included interactive charts which presented the user with HIG, Float and CIL calculations over time.
If you’re interested in BlendOpt’s capabilities in a gold operation context, read our article detailing the value that BlendOpt can provide from gold mine to mill.
FAQ
Gold mill feed scheduling FAQ
How BlendOpt schedules gold mill feed within grade and capacity constraints, and replaces the planning spreadsheet.
What is gold mill feed scheduling optimisation?
Gold mill feed scheduling optimisation is deciding which run-of-mine (ROM) ore to crush and send to the mill in each period, and which ROM stockpiles to reclaim from, so the operation gets the most gold from its ore. The schedule has to respect grade constraints and crusher and mill capacities at the same time. BlendOpt uses mathematical optimisation to search for the best schedule rather than relying on manual trial and error.
Why replace a mill feed spreadsheet with optimisation software?
Spreadsheets used for mill feed scheduling often grow overly complex, with assumptions that are hard to interrogate and formulas that are easy to break. In this case study the workbook held over 30 recovery related calculations and produced negative ROM stockpile values from infeasible plans, and technical services was losing confidence in its results. BlendOpt replaced it with a digital twin that is easier to maintain and removed human error from planning decisions.
How does BlendOpt handle grade and capacity constraints?
Grade constraints and crusher and mill capacities are modelled directly in the BlendOpt digital twin, with capacities calculated from planned outages. BlendOpt then suggests crusher and mill feed schedules based on available inventory and predicted variability that comply with those constraints. Different conditions can be tested by running a range of scenarios.
Can BlendOpt model gold recovery calculations?
Yes. Recovery formulas are written in BlendOpt in human-readable form, as they would be written down in the user's preferred notation, which makes them easier to understand and maintain. Interactive charts present HIG, Float and CIL calculations over time so planners can see how the schedule affects recovery.
Does BlendOpt only optimise the mill, or the whole mine-to-mill chain?
BlendOpt can model the whole value chain downstream of the mine. It takes the mine schedule as an input and, starting from ROM, covers ROM stockpiles, crushing, milling and recovery. Stockpile definitions and reclaim can be optimised together with mill feed, which is how a copper operation cut Cu:S mill feed variability from 0.4 to 0.03. For more on gold operations, see the BlendOpt gold mine-to-mill overview.