Outliers Identification

Evaluation Scope and Metrological Standards

In small-scale additive manufacturing and rapid batch runs, an anomalous measurement value can skew your entire statistical process control (SPC) assessment. Outliers identification isolates true fabrication anomalies caused by momentary mechanical glitches, material contamination, or thermal shifts from expected standard deviations. When managing tight mechanical tolerances across rapid production runs, differentiating between a single erratic data point and an emerging process drift ensures your calibration efforts address real root causes rather than random measurement artifacts.

Standard statistical filters like the Interquartile Range (IQR) rule and Grubbs' test provide clear mathematical criteria to determine if a geometric deviation qualifies as an actionable outlier. Applying these thresholds directly to critical physical dimensions—such as bore concentricity, wall thickness, and outer diameter—prevents premature retooling while pinpointing physical issues like sudden filament diameter variation or temporary nozzle clogging before full-run failure occurs.

Process Variation Breakdown

Systematic classification separates outliers into assignable causes and natural variation. Identifying whether a dimensional spike originates from slicing artifacting, transient stepper motor resonance, or ambient temperature fluctuation dictates the corrective protocol for subsequent runs.

• Statistical filtering using Grubbs' test at critical value G > 2.41 (N=12 specimen threshold, α = 0.05).
• Geometric isolation protocol separating isolated Z-axis layer stepping from uniform extrusion swelling.
• Immediate thermal and optical inspection of mechanical beds following detected 3-sigma deviations.