🎓 Lesson 18
D5
Lab-Field Data Discrepancy Diagnosis: Sample Disturbance and Moisture Hysteresis
Lab tests sometimes give different results than field measurements because soil or rock samples get disturbed during collection or their moisture content changes between site and lab.
🎯 Learning Objectives
- ✓ Analyze moisture retention curves to identify hysteresis loops and quantify suction differences between drying and wetting paths
- ✓ Explain how sampling method (e.g., Shelby tube vs. block sampling) affects strength and compressibility parameters for cohesive soils
- ✓ Apply the Bjerrum correction factor to undrained shear strength (su) derived from lab vane or triaxial tests on sensitive clays
- ✓ Calculate degree of sample disturbance using the ratio of field-measured N60 (SPT) to lab-derived N1,60 and interpret its implications for design safety
📖 Why This Matters
In mining and open-pit slope design, a 20% overestimation of shear strength—due to uncorrected sample disturbance—can lead to non-conservative slope angles, increased risk of failure, and costly remediation. Similarly, ignoring moisture hysteresis in waste dump or leach pad liner design may underestimate infiltration rates by up to 3×, violating environmental compliance. This lesson equips you to diagnose *why* your lab data disagrees with field performance—and fix it before design decisions are locked in.
📘 Core Principles
Sample disturbance manifests as loss of natural fabric, pore pressure dissipation, and stress path deviation—especially critical in soft clays, glacial tills, and weathered shales. The degree of disturbance is quantified by the ratio of field stiffness (e.g., Gmax from seismic cone) to lab stiffness (e.g., Gmax from resonant column), where ratios < 0.5 indicate severe disturbance. Moisture hysteresis arises from contact angle hysteresis and air-entry value differences between drying (main wetting curve) and wetting (main drying curve) paths; this creates two distinct soil-water characteristic curves (SWCCs) for the same soil at identical matric suction. Diagnosis requires integrating field context (e.g., recent rainfall, excavation rate, groundwater fluctuations) with lab test history (e.g., saturation protocol, consolidation stress path).
📐 Bjerrum Correction for Undrained Shear Strength
The Bjerrum correction adjusts lab-measured undrained shear strength (su_lab) to reflect field conditions for sensitive clays, accounting for sample disturbance and stress history. It is essential before using su in limit equilibrium slope stability analysis.
💡 Worked Example
Problem: A lab vane test on a soft marine clay yields su_lab = 28 kPa. The clay has a liquidity index (LI) of 0.72 and was sampled using a 75-mm-diameter thin-walled Shelby tube. Field vane testing (in situ) at adjacent location gives su_field = 19 kPa.
1.
Step 1: Determine sensitivity ratio St = su_field / su_lab = 19 / 28 ≈ 0.68.
2.
Step 2: Use Bjerrum’s empirical chart (or equation) — for LI = 0.72, the recommended correction factor λ ≈ 0.55 (from Figure 3.12, Lambe & Whitman, 1969).
3.
Step 3: Compute corrected strength: su_corrected = λ × su_lab = 0.55 × 28 = 15.4 kPa.
4.
Step 4: Compare with field value: 15.4 kPa is conservative relative to 19 kPa, reflecting typical underestimation of field strength when using uncorrected lab data.
Answer:
The corrected su is 15.4 kPa, which falls within the safe design range for preliminary slope stability (typically 12–18 kPa for similar clays per SME 2022 guidelines).
🏗️ Real-World Application
At the Highland Valley Copper mine (BC, Canada), bench-scale slope failures occurred despite conservative lab-derived c’ and φ’ values. Post-failure investigation revealed that rotary-sampled core used for direct shear tests had lost >40% of its natural structure due to drilling-induced vibration and handling. Concurrent TDR moisture monitoring showed hysteresis-driven suction differences of 12–25 kPa between pre-rain (drying path) and post-rain (wetting path) states in the overburden clay-silt layer—causing unexpected reductions in effective stress. Revised design incorporated Bjerrum-corrected su, SWCC hysteresis modeling in SEEP/W, and switched to block sampling for future campaigns—reducing predicted failure probability by 63%.