Case Study — Instrumenting In-Pit Tailings Dykes at Muskeg River Mine (2013)¶
About this case study
This page is a digest and full Vietnamese translation of the 2013 Canadian Geotechnical Society conference paper Application of Geotechnical Instrumentation in Monitoring In-Pit Dykes Performance, by K.W. Soe Moe and K.W. Biggar (BGC Engineering Inc., Edmonton) with S. Martens and R. Hoda (Shell Canada Ltd., Calgary). The paper and its figures remain the copyright of the authors and of the Canadian Geotechnical Society; it is reproduced here for education and reference alongside an independent Vietnamese translation. Cite the original English paper in any formal work.
In-pit dykes are the least-discussed and one of the most demanding instrumentation problems in mine waste management. An in-pit dyke (IPD) is a tailings-retaining embankment built inside a mined-out pit, separating one tailings cell from the next. Unlike an external tailings dam, it is founded on a floor that has been unloaded by up to 90 m of excavation, it is built from whatever waste the mine is stripping at the time, and it has to be raised continuously while tailings are being deposited behind it — so the instruments buried in it must survive being extended upward through the fill, year after year, by the very equipment that is building the dyke.
This paper, from Shell Canada Energy's Muskeg River Mine (MRM) about 70 km north of Fort McMurray, Alberta, is one of the few published accounts of how that is actually done at scale: dykes 60–90 m high, built on a foundation that has been unloaded and reloaded, monitored well enough that the design itself could be relaxed as the data came in.
What the paper establishes¶
The dykes are designed by the Observational Approach (Peck, 1969), classified "Very High" consequence under the Canadian Dam Association (2007) guidelines. That classification is what makes the instrumentation load-bearing: the design assumes most-likely material parameters, and the monitoring programme is the mechanism that proves the assumption — or buys the time to act on a pre-planned remedy.
| # | Section | What it covers |
|---|---|---|
| 1 | Introduction | Alberta's three oil sands deposits; external vs. in-pit tailings containment; why the Observational Approach requires a well-planned instrument programme |
| 2 | Geologic setting | Quaternary cover over the Cretaceous McMurray Formation (Upper / Middle / Lower members, with the water-bearing LM1w basal sands), separated from the Devonian Waterways Formation (DW) by a sub-Cretaceous unconformity and its weathered Paleosol |
| 3 | Design considerations | CDA "Very High" classification; the five key design requirements; dykes founded on LM1w or DW at 200–230 masl against 290 masl original ground — hence 60–90 m dyke heights |
| 4 | Fill type classification | Type 1 (coarse oil sand), Type 2 (fine oil sand), Type 5 (overconsolidated clay) and Trafficable Waste; fluid content and rut criteria as construction quality control (Table 1) |
| 5 | Geotechnical instrumentation | Vibrating wire piezometers, slope inclinometers, magnetic extensometers, shaped accelerometer arrays, earth pressure cells and standpipes at 200–300 m spacing; mud-rotary vs. auger installation; 1 m × 1 m trench installation; instrument protection mounds; cell and radio modems (Table 2) |
| 6 | Pore pressure measurement | Construction pore pressure ratio ru (eq. 1, Table 3); foundation response ratio B (eq. 2, Table 4); why dissipation is only moderate, and why some fill piezometers read negative pressures |
| 7 | Deformation monitoring | 0.01–0.07 mm/day typical, up to 0.2–3.0 mm/day during nearby construction; distributed strain over 30–40 m zones with no discrete shear zones; the four-level caution limit table (Table 5) |
| 8 | Total stress monitoring | Earth pressure cells; Ko = (σh − u) / (σv − u) (eq. 3); measured Ko of 0.4–0.7 feeding a hydraulic-fracture assessment |
| 9 | Conclusions | Instrumentation as the enabling mechanism for safe construction and design optimisation |
Full documents (PDF)¶
Bản dịch tiếng Việt đầy đủ (19 pp., A4). Bản dịch trọn vẹn toàn bộ bài báo, gồm 5 bảng, 14 hình minh họa và toàn bộ danh mục tài liệu tham khảo. Thuật ngữ được chuẩn hóa theo TCVN (áp kế dây rung, ống đo nghiêng, thiết bị đo biến dạng sâu từ tính, hộp đo áp lực đất, ứng suất hữu hiệu…), trình bày bằng phông Times New Roman.
Original English paper — Application of Geotechnical Instrumentation in Monitoring In-Pit Dykes Performance (9 pp.), Soe Moe, Biggar, Martens & Hoda, Canadian Geotechnical Society conference paper, September 2013. Reproduced here as published, for reference.
The three results worth carrying away¶
1. The fill behaved better than designed — and the design was relaxed¶
Initial design pore pressure ratios were ru = 0.5 (core), 0.3 (shell) and 0.4 (toe berm). Measured values overwhelmingly sat below 0.3. Because oil-sands fill is placed at its natural fluid content and compacted by fully-loaded haul trucks, it is denser and drier of optimum than conventional embankment fill — so it generates less construction pore pressure than civil-side experience predicted. Monitoring the construction let the team select ru values appropriate to the actual stress range, which the paper describes as "an overall improvement in the fill utilization of the designs."
2. The foundation barely responded — because it had been unloaded¶
Measured B values were 0.05–0.15 in the McMurray members and 0.1–0.3 in the DW Paleosol. The paper attributes this to stress relief from excavating up to 90 m of material before the dyke was built, and warns that a higher B response should be expected once dyke elevation passes original ground level. The high-permeability LM1w basal water sands showed no response at all to dyke construction — but did respond to fluid levels in the tailings cells.
3. Caution limits were tied to a rate, not a displacement¶
Table 5 is the transferable artefact. It sets four independent, rate-and-strain-based bands with an explicit monitoring and construction response attached to each:
| Rate of movement (mm/day) | Total shear strain (%) | Level | Response |
|---|---|---|---|
| < 0.07 | < 0.5 | Green | Normal construction; bi-weekly monitoring |
| > 0.07 – < 0.2 | > 0.5 – < 1.0 | Yellow | Increase monitoring to weekly / twice weekly |
| > 0.2 – < 1.0 | > 1.0 – < 2.0 | Orange | Increase to twice weekly / daily |
| > 1.0 | > 2.0 | Red | Stop dyke construction in that area until the rate falls |
Note the structure: the trigger is a velocity, and the top band's action is a construction stoppage, not an alarm. That is the same logic the site's ADAQS trigger-action framework is built on, and it is why the paper reports that no discrete shear zones ever developed — construction was stopped or slowed before one could.
Why this paper matters in this knowledge base¶
- It is the mine-side companion to Tailings Dam Safety and Tailings Dam Safety (ICOLD Bulletin 194), which treat external tailings facilities. In-pit containment has a different foundation history and a different construction sequence.
- Its caution-limit table is a worked, published example of the trigger logic described generically in Getting Started.
- Its total-stress and Ko section is one of the few field datasets linking instrumented earth pressure cells to a hydraulic fracture assessment — a failure mode that matters wherever low-stress zones develop near steep abutments.
- Its Observational Approach framing (Peck, 1969) is the same spine that runs through Dunnicliff and the FHWA reference manual.
How to use this case study¶
- Start with §5 (instrumentation and protection) if you are planning a programme on a structure that will be raised during monitoring — the protection-mound and "no vertical extensions in the core zone" rules are practical and rarely written down.
- Use Table 5 as a starting template for your own trigger table, then re-derive the numbers for your materials, your construction rate and your consequence classification. Do not copy the thresholds.
- Read §6 before specifying a pore pressure design value for compacted oil-sands or similarly placed fill; the conventional Proctor-based assumption is conservative here.
- Cross-link into the Mining Geotechnics overview for slope radar, GNSS and microseismic methods that complement the buried instruments described here.
Safety and currency. The data in this paper describe IPD 1 and IPD 2 as designed at the time of writing in 2013; the authors note the parameters were being updated as new data became available. Remote monitoring, GNSS and distributed fibre-optic sensing have all advanced since. Treat the observational logic and the caution-limit structure as durable, and derive your own thresholds from current site data, current CDA/MAC guidance and applicable regulation before using them for design or a regulatory submission.
Featured Video & Practical Resources¶
Featured Keynote Lecture (Rocscience): Dr. Evert Hoek Lecture Series — Rock Slope Engineering: Failure Mechanisms, Water Pressure Effects, and Pit Dyke Stability.
Complementary Practical Field Resources from RST Instruments & Measurand (Terra Insights)
- RST Instruments (@RSTgeotechnical): RSTAR Affinity Quick Dashboard Tour
Unified IIoT sensor management platform for real-time geotechnical telemetry and dyke monitoring networks. - Measurand (@MeasurandVideos): Automated ShapeArray System
Automated 3D subterranean deformation tracking connecting to centralized dyke monitoring networks.