What is Grade Control?
Grade control is the operation where truckloads are flagged as ore or waste based on blast hole, drill hole and geologic information. Classification of material in the field basically.
Why does Grade Control Matters?
Classification errors generate expense. If a block of ore is misclassified, a net loss ($) occurs. Such misclassification errors can never result in a net $ gain.
What is the Actual Lost?
The actual loss ($) associated with each type of misclassification error is simply the potential value of the SMU less the actual value recovered.
What is the principle of Minimizing loss due to misclassification
The principle here is that any value selected other than the (unknown) true value will incur a loss.
Case Study: Gold Mine
A mining polygon is classified into one of four types: Waste, ROM ore, Agg. ore, Mill ore.
With the following lost table:
What are the 5 steps for finding the minimizing loss classification from simulation
*heavy
- Simulations give grade distribution for each block. For each block, z(x) is th eunknown grade of the block to be classified
- Go through each cut-off range defined from waste and ore types and find the probabilities of z(x) for each category
- Calculate the loss for each type of misclassification by finding the average (expected) value within each ore category
- Plug values average grade values inside the loss table to find the loss function for each type of missclassification
- Multiply the probability of each category found earlier to get the loss per column
Describe the content of the following table
- Zero Loss in the diagonal corresponds to rightfull classificatino
- Actual Loss Function due to misclassification of ore
Maximizing profit classification
Discuss Overestimation and the Correct classification scenarios
The approach covers overestimation and the correct classification scenarios, being the difference between revenue and cost.
Discuss the concept of Maximizing Profit Classification (Profit/Loss approach)
Involves the incorporation of all possible classification scenarios, and instead of basing our decision only on expected loss, we will select the classification that maximises the expected profit or undiscounted cash flow. Method covers both overestimation and underestimation.
Maximizing profit classificaiton
Discuss Underestimation scenatios
Underestimation scenarios and reflects the opportunity cost of losing the difference between the potential and recovered value
What are the step for Maximizing profit classification
same as Minimizing loss classification🖕
Minimizing loss classification
In step 5.
5: (Multiply by the probability of each category found earlier for that block to get loss per column. )
Which value below, should we pick for classification the block in order to minimize loss
0.82, the lowest
Maxmixzing profit classification
In step 5.
5: (Multiply by the probability of each category found earlier for that block to get loss per column.)
Which value below , should we pick for classificatino of the block in order to maximize profit
0.43, the highest
Case Study: Grade Control comparisons in a Controlled environment, the W
With a Process of:
1. Generate point simulations
2. Change support and generate block simulation
3. Apply the functions (OK, Min/Loss, Max/Profit) and compare to True values
What are the three (3) conclusions of the comparisons?
- The OK estimates recovered approximately 70% of the total dollar value, Inferior method to Min/Loss
- The Min/Loss classification recovered 73.6% or 3.6% more then by using OK, Superior method to OK estimates
- The Max/Profit classification recovered approximately 68.3%. Inferior two both other methods. Too aggressive bad, no conservatism.
Discuss Integrating grindability to improve ore classification and comminution
-Improve grinding circuit performance, resulting in the reduction of costs, time and energy consumption.
- Different grinding characteristics have different processing costs, for example, harder ores have increased comminution time and large recycling loads within comminution circuits. à Therefore, classification should be done according to grindability too.
Case Study: Ore control, Optimality & decision making under uncertainty
Good mill vs. grade control results using the traditional, but wants to do better.
What are the results of trying Sequential Indicator Simulation?
Initially:
Its makes up to 7% increase of gold
How is ‘Integrating Grindability’ related to ore classification
Changes in processing costs impact ore/waste definition. Therefore, grade control can utilize grindability at a smaller scale to impact block classification.
When the grindability of each block is considered, the effect of the cost of processing on the classification as ore or waste may… (3 Scenarios)
- Upgrade a block from waste to ore
- Downgrade a block from ore to waste
- Leave the classification unchanged, when compared with the base-case scenario
Case Study: Economic functions & comparisons of the Walker Lake deposit
Comparison of the methods using a fully known deposit:
Simulation techniques: SGS, SIS
Economic functions: Min/Loss, Max/Loss
What are the Profit results for the three scenarios (high,med,low cut-offs)
- Low Cutoffs: SGS and Max/prof makes more profit
- Medium Cutoffs: SGS and Min/loss makes more profit
- High Cutoffs: SIS and Min/loss makes more profit
