Surface data can reveal a progressive change in cutting efficiency, but it cannot measure cutter damage directly. A useful bit-wear assessment tracks a mechanics-based indicator against a formation- and state-matched baseline, checks reversible dysfunction, and validates the interpretation with post-run evidence.
Use work per foot as an indicator
One useful way to understand PDC wear is to consider how much loaded cutter travel occurs while drilling a unit distance. At high RPM and low ROP, the cutters travel around the bottom many times per foot; at the same RPM and higher ROP, fewer revolutions are required. That relationship is closely connected to depth of cut:
$$DOC \propto \frac{ROP}{RPM_{bit}}$$
Higher DOC means more axial advance per revolution. Lower DOC means more cutter revolutions are required to create the same footage. SPE-205844 used this concept to build a wear indicator combining:
- bit rotational speed,
- ROP,
- WOB.
A simplified form used in the study was:
$$Wear\ Indicator = 60 \cdot \frac{RPM_{bit}}{ROP} \cdot WOB$$
and equivalently:
$$Wear\ Indicator = \frac{WOB}{DOC}$$
under the unit convention used in that work.[1] The physical intuition is more important than memorizing the equation:
More loaded cutter travel per foot drilled implies greater opportunity for abrasive wear and thermal loading.
Lower depth of cut requires more loaded revolutions to drill the same footage, increasing conceptual cutter travel without directly measuring physical wear volume.
This qualification matters. The wear indicator described in SPE-205844 was explicitly not presented as a quantitative measurement of actual volumetric cutter wear.[1] It is an indicator. That is a very different claim. The cutters may physically experience:
- abrasion,
- thermal degradation,
- impact damage,
- structural overload.
Surface measurements see only the resulting drilling response. The indicator is evidence of changing cutting response, not a measurement of actual cutter wear. This distinction prevents false precision.
Track departure from a local baseline
A wear-indicator value of 4,500 means little without context. A stable progression of 2,100, 2,200, 2,300, 2,450, and 2,600 followed by 4,500 is much more informative. SPE-205844 therefore emphasizes departure from a baseline.[1] The useful question is whether the relationship between mechanical input and drilling response changed unexpectedly, not whether one value is universally high.
A persistent departure from the previous stand-level trend is more informative than an isolated high indicator value, but it still requires diagnosis.
Control state and bit RPM
The methodology also filtered the data to rotating on bottom. Sliding was excluded because the mechanical assumptions were more defensible during rotary drilling. In particular, the authors did not treat high surface WOB during sliding as a reliable formation-hardness indicator for this study.[1] Surface WOB in rotary and sliding modes should not be assumed to have identical physical meaning. Likewise, bit RPM during sliding may be dominated by the mud motor while surface RPM is zero. State filtering is not just data cleaning; it establishes the physical population to which the model applies. In motor-assisted drilling, the study estimated bit rotational speed as:
$$RPM_{bit} \approx RPM_{surface} + RPM_{motor}$$
with motor RPM estimated from:
$$RPM_{motor} = Flow \times Motor\ Speed$$
using the motor's revolutions-per-gallon specification.[1] With 120 rpm at surface and an estimated 100-rpm motor contribution, bit speed is approximately 220 rpm. Using only the surface value would materially underestimate cutter travel. During a slide, surface RPM may be zero while bit RPM remains substantial, another reason the calculation must preserve provenance.
Separate wear from its confounders
Suppose:
- WOB remains constant,
- RPM remains constant,
- ROP falls.
The wear indicator increases. One interpretation is bit deterioration. Another is harder rock. SPE-205844 explicitly warns that formation changes can alter the ROP-dependent indicator even in the absence of bit wear or damage.[1] That means a bit-surveillance system benefits greatly from:
- formation context,
- gamma/lithology,
- offset performance,
- rock-strength information where available.
A baseline established in one formation should not automatically be applied to another. Consider stick-slip. ROP can decrease. Effective bit speed becomes highly irregular. Surface torque may oscillate. A wear indicator can increase. But the immediate problem may be: drilling dysfunction rather than progressive cutter wear. The two can also interact. A hard stringer may initiate dysfunction. The dysfunction may damage cutters. The damaged cutters may then reduce ROP permanently. SPE-205844 explicitly notes that these events are not mutually exclusive.[1] This makes root-cause analysis a sequence problem. The engineer needs to ask:
What changed first?
A bit can experience gradual abrasive wear for thousands of feet. Failure implies a more consequential change in cutting structure or effectiveness. In the dataset studied in SPE-205844, likely failure onset was often associated with a distinct departure from the prior wear trend followed by ROP that did not recover.[1] Once severe damage began in that hard, abrasive formation, deterioration sometimes progressed rapidly. Those observations are dataset-specific and should not become universal footage rules.
| Pattern | Wear indicator | ROP or supporting response |
|---|---|---|
| Normal abrasive progression | Gradual increase with depth | Gradual response |
| Sudden damage event | Sharp departure from baseline | Step deterioration or failure to recover |
| Formation transition | Change aligned with geological context | Formation-matched response |
| Dysfunction | Change with mechanical instability | May recover when dysfunction clears |
These patterns can overlap, but separating them improves the diagnosis.
Close the loop with bit forensics
One of the strongest aspects of SPE-205844 is that the surface-data interpretation was compared with actual PDC dull photographs.[1] That matters because field analytics needs a truth source. Without post-run validation, it is easy to tell convincing stories from historical traces. For example:
ROP fell here, so this must have been when the bit failed.
A dull photograph may reveal:
- severe ring-out,
- shoulder damage,
- thermal wear,
- surprisingly light wear.
The physical evidence can:
- support,
- modify,
- reject
the hypothesis derived from the surface data. This is the scientific feedback loop.
A qualified surveillance workflow
A defensible workflow is compact but disciplined:
| Stage | Practice |
|---|---|
| Define the population | Separate rotary, slide, and other states according to the model assumptions. |
| Establish rotational context | Distinguish surface, estimated motor, and estimated bit RPM. |
| Calculate and aggregate | Use a documented, unit-consistent indicator and a robust interval such as stand-level median. |
| Build the baseline | Match formation, bit/BHA context, and operating state. |
| Detect departures | Look for gradual progression, step changes, and failure to recover. |
| Diagnose | Review torque, WOB, RPM, differential pressure, MSE, formation, and reversible dysfunction. |
| Preserve uncertainty | Report competing explanations when formation and bit degradation both remain plausible. |
| Validate after the run | Compare the inferred history with the dull grade, photographs, and observed cutter damage. |
Surface data can be useful for identifying:
- Changing cutting efficiency: The same or greater mechanical input produces less useful footage.
- Departure from a previous bit-run trend: A stand-level relationship changes materially.
- Timing of a suspected degradation event: The likely interval can be narrowed for closer investigation.
- Supporting evidence of dysfunction: Torque, pressure, or other channels may reveal a plausible mechanism.
Those are valuable outputs. Surface data generally cannot guarantee:
- Exact cutter damage: It does not directly see the PDC cutting structure.
- Exact wear volume: A mechanics indicator is not a physical wear gauge.
- Unique root cause: Formation and dysfunction can mimic degradation.
- Absence of downhole dysfunction: Some damaging dynamics may have weak surface signatures.
Those limitations should remain visible in the interpretation.
The wear indicator remains beside its source channels and offset context; the display supports surveillance without turning surface data into a direct cutter-damage measurement.
Related technical resources
- When Should You Pull a Drilling Bit? A Decision Under Uncertainty
- Mechanical Specific Energy: Powerful Metric, Incomplete Diagnosis
- Formation-Aware Drilling Optimization: Interpreting Performance Across Changing Rock
References
- Witt-Doerring, Y., Pastusek, P. P., Ashok, P., and van Oort, E. Quantifying PDC Bit Wear in Real-Time and Establishing an Effective Bit Pull Criterion Using Surface Sensors. SPE-205844-MS, SPE Annual Technical Conference and Exhibition, 2021.