How would you select representative and worst-case sampling points?

Why

In a large tank, material may mix differently near the feed port, wall, bottom and outlet. Temperature gradients or hold-time effects can also create risk at different phases of a campaign. Use flow modeling, development data, engineering runs and empirical mapping to identify representative and scientifically defensible worst-case locations and times.

Prespecify the sampling plan, method, acceptance criteria and rationale. Ensure the sampling action does not itself change the system or introduce contamination. For biologics, account for material attributes that differ across phases such as product concentration or aggregation risk.

A single fixed “worst-case” location should not be assumed for all failure modes. Uncertainty calls for expanded coverage and later refinement using PPQ and continued verification evidence.