Process Capability Analysis in Manufacturing Systems

Process capability analysis assesses the inherent variation of manufacturing processes relative to specified tolerance limits. It quantifies whether a process consistently produces components within design requirements and underpins quality assurance, risk management and continuous improvement. Classical capability indices, such as C_p and C_pk, assume normally distributed output and relate process mean and standard deviation to specification limits. In practice, many processes exhibit non-normal behaviour due to factors such as tool wear, raw-material variability or complex assembly steps. To address this, researchers have developed transformations to approximate normality, quantile-based indices for skewed data and robust measures that rely on median absolute deviation or Gini’s mean difference. Advances also encompass interval estimation via bootstrap or analytical confidence limits and multivariate methods for correlating multiple quality characteristics. Recent work highlights application of these methods to high-precision electronics, automotive components and biomedical devices, demonstrating global significance in reducing scrap, improving yield and enhancing sustainability through resource efficiency. The integration of capability analysis into digital platforms and real-time monitoring further extends its practical impact across Industry 4.0 environments.

Research from Nature Portfolio

Recent studies have refined robust third-generation capability indices for non-normal processes modelled by Weibull distributions. Efficient dispersion measures such as median absolute deviation, interquartile range and Gini’s mean difference were systematically evaluated under varying asymmetry. Findings indicate that median absolute deviation performs reliably across low to moderate skew, while quantile-based indices suffer bias under high asymmetry. Bootstrap confidence intervals tailored to each dispersion measure have been proposed, with bias-corrected percentile intervals recommended for quantile indices and percentile-t intervals for median-based measures. Validation on industrial datasets confirmed superior performance of the robust methods, offering practitioners improved tools for assessing capability of components subject to fatigue-related and failure-time distributions.

Process Capability Analysis in Manufacturing Systems publication trend

The graph below shows the total number of articles in process capability analysis in manufacturing systems across all publications each year (not limited to Nature Index journals).

Technical terms

Process capability: The ability of a manufacturing process to produce output within specified limits, accounting for variation.

Process capability index (PCI): A numerical measure (e.g. C_p, C_pk) relating process spread and centring to tolerance limits.

Non-normal distribution: A statistical distribution that deviates from a normal (Gaussian) shape, often requiring specialised analysis.

Power transformation: A mathematical technique (e.g. Box-Cox) applied to data to stabilise variance and approximate normality.

Bootstrap confidence interval: A resampling-based method to estimate the uncertainty bounds of capability indices without strict parametric assumptions.

References

1. A Power Transformation for Non-Normal Processes Capability Estimation. IEEE Access (2024).
2. Robust process capability indices Cpm and Cpmk using Weibull process. Scientific Reports (2023).
3. Process-Quality Evaluation for Wire Bonding With Multiple Gold Wires. IEEE Access (2020).
4. Estimating and Testing Quantile-based Process Capability Indices for Processes with Skewed Distributions. Journal of Data Science (2021).