In open pit mining, the most significant challenge is determining the optimum long-term production schedules that maximize the project value by providing ore to the processing plant at full capacity while satisfying all required constraints. For oil sands strategic mine planning and waste management, as mining advances in a specified direction, in-pit tailingscells’ dyke footprints are released for dyke construction. The dykes are constructed from overburden, interburden and tailings coarse sands dyke materials which come from the mining operation. The construction of in-pit (and ex-pit) tailings impoundment dykes therefore needs to be well integrated with the waste management strategy to ensure regulatory compliance and sustainable mining.
In this research, an uncertainty-based mathematical programming framework is developed based on mixed integer linear goal programming (MILGP) model for oil sands production scheduling and waste management. The effect of grade uncertainty on production schedules is investigated. The grade uncertainty financial risk associated with a production schedule is minimized using kriged estimates with a variance penalty scheme. This investigation is based on the concept of mean-variance analysis, which is the process of weighing variance (risk) against the expected net present value (NPV). Subsequently, the impact of organic rich solids (ORS) content on bitumen recovery during processing is also studied. ORS content is used to predict ore processability in addition to the traditional use of bitumen and fines contents, and its impact on NPV quantified. The developed MILGP model is implemented using new robust automated production targeting (APT) constraints that optimize the annual capacities for material mined and processed over the mine life. In addition, mining-cells are deployed for the generation of in-pit tailings-cells designs used as dedicated disposal areas for backfilling.
Two main waste management approaches are used to implement the developed model. Implementation A; where the ultimate pit is divided into predetermined pushbacks as tailings-cells within the ultimate pit limit (UPL), and Implementation B; where the ultimate pit is divided into mining-cells that are used to generate in-pit tailings-cells designs. To verify the research models, three oil sands case studies were carried out.
The first case study investigates the effect of grade uncertainty on production schedules. The technique applied is based on the concept of mean-variance analysis, which is the process of weighing risk (variance) against expected NPV. The model generates a range of NPV which represents the mining investment risk profile associated with grade uncertainty. The second case study explores the impact of ORS content on oil sands ore processability. The results showed a 3.46% overestimation of NPV arising from not taking into account the effect of ORS content on bitumen recovery during mine planning. The final case study examines the implementation of waste management strategies based on different size and number of unit mining-cells used in creating tailings-cells for backfilling. The results showed that, decreasing the volume of unit mining-cells used in creating the in-pit tailings-cells increases the NPV of the operation due to increased operational flexibility. Additionally, as the percentage of in-pit volume to be backfilled increases, more savings is generated from not sending tailings to external facilities at a higher cost. These results proved that the uncertainty-based MILGP model is a robust tool for optimizing oil sands long-term production schedules whilst taking into account grade uncertainty, ore processability and tailings-cells designs.