ADM1 Sensitivity Analysis via Functional Principal Components

A methodology paper from Lafayette College demonstrates a global sensitivity analysis technique for Anaerobic Digestion Model No. 1 (ADM1). The work transforms time-dependent model outputs using functional principal component analysis (fPCA) for input into Morris' screening design. Results indicate that 95-99% of output variation can be captured by principal components, with the first PC sufficient to represent model outputs.

Use Cases

• Ranking stoichiometric parameter sensitivities based on calculated Morris indices.
• Identifying dominant mechanisms affecting kinetic responses in anaerobic digestion models.
• Informing mathematical model modification by highlighting least sensitive parameters.
• Applying functional principal component analysis to reduce dimensionality of time-series model outputs for sensitivity studies.

Strengths

• Methodology validated on the widely-used Anaerobic Digestion Model No. 1 (ADM1).
• Results show 95-99% of time-dependent output variation can be captured by principal components.
• Analysis identifies parameters that dominantly affect kinetic responses.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

Dhan Lord B. Fortela, Lafayette College

Collection Method

Methodological research involving Morris' screening design and functional PCA on ADM1 outputs.