Evaluation of Sampling Facilities and Process Analytical Technologies in Applied Biotechnology: Process Sampling, Representativeness, and Acoustic Chemometrics
Author
Boland, Ludovic
Term
4. term
Education
Publication year
2008
Pages
252
Abstract
This thesis evaluates sampling systems and their efficiencies in lab-scale, applied biotechnology setups. A planned evaluation of ACABS’ bioreactor system could not be carried out due to Force Majeure, but the international TENIRS system was fully assessed. The results were surprising: the TENIRS reference sampler is inaccurate and highly sensitive to operating conditions, yet it is almost perfectly reproducible—consistently wrong in the same way (a systematic error, bias). We tested model systems of polymer in water: light particles, heavy particles, a mixture of both, and a synthetic manure, covering realistic dry matter levels. Bias increases with the complexity of the multiphase system but decreases as the circulation speed in the TENIRS measuring loop increases. At low speed (20% of maximum), composite samples (10 increments) of synthetic manure showed bias up to 320%, while light pellets were around 85%. At higher speed (60%), the bias for synthetic manure was 15%, but systems with heavy particles were strongly underestimated (about −70%). These findings point to a clear need to redesign the TENIRS reference sampling system. Acoustic chemometrics was used to model dry matter content. All models were validated on independent test sets. For synthetic manure, we achieved a prediction-to-reference slope of 0.89 and r² = 0.89, indicating that acoustic chemometrics is a promising Process Analytical Technology (PAT) approach for physical response variables. However, more work is needed on this type of applied biotechnological system. The thesis also provides a review of PAT modalities with a focus on applied biotechnology.
Dette speciale evaluerer prøvetagningssystemer og deres effektivitet i laboratoriebaserede, anvendte bioteknologiske systemer. Den planlagte evaluering af ACABS’ bioreaktorsystem kunne ikke gennemføres på grund af Force Majeure, men det internationale TENIRS-system blev evalueret fuldt ud. Resultaterne var overraskende: TENIRS’ referencesampler er upræcis og meget følsom over for driftsbetingelser, men den er samtidig næsten perfekt reproducerbar—altså konsekvent forkert på samme måde (en systematisk fejl, bias). Vi testede på modelsystemer af polymer i vand: lette partikler, tunge partikler og en blanding af begge, samt en syntetisk gylle. Tørstofindholdet dækkede realistiske niveauer. Biasen stiger med systemets flerfase-kompleksitet, men falder, når cirkulationshastigheden i TENIRS’ målesløjfe øges. Ved lav hastighed (20 % af maksimal) gav sammensatte prøver (10 delprøver) af syntetisk gylle bias op til 320 %, mens lette pellets lå omkring 85 %. Ved højere hastighed (60 %) var bias for syntetisk gylle 15 %, men for systemsimuleringer med tunge partikler var biasen kraftigt negativ (omkring −70 %), dvs. en stærk undervurdering. Disse fund indikerer et klart behov for at redesigne TENIRS’ referencesamplingssystem. Akustisk kemometri blev brugt til at modellere tørstofindholdet. Alle modeller blev valideret på uafhængige testdata. For syntetisk gylle opnåedes en hældning på 0,89 og r² = 0,89 mellem forudsagt og reference, hvilket peger på akustisk kemometri som en lovende metode inden for procesanalytisk teknologi (PAT) til fysiske responsvariabler. Der er dog behov for mere arbejde på denne type anvendte bioteknologiske systemer. Specialet rummer også en oversigt over PAT-metoder med fokus på anvendt bioteknologi.
[This apstract has been rewritten with the help of AI based on the project's original abstract]
Keywords