Uncertainty and sensitivity analysis of algae production models

M.S. Stefanov, P.M. Slegers, A.J.B. van Boxtel

Research output: Contribution to conferenceAbstract


Lab scale studies have shown great promise in biodiesel production by algae. Yet, knowledge on large scale systems is needed to advance the technology. Credible feasibility, technical, techno-economical, and life cycle assessment studies require accurate biomass productivity estimates at various production locations, system layouts, algae species, weather conditions, etc. Models for making such estimates have been developed in the scenario study of (Slegers, et al. 2011) for flat panel photobioreactors. There is a desire to evaluate a priory the predictive power of the models of (Slegers, et al. 2011) in the face of parametric and modelling uncertainty. This projects aims at discovering 1) what the likelihood of the biomass productivity estimates is, 2) at what conditions the productivities are most accurate, 3) what needs to be researched to improve the accuracy, and 4) what the optimal conditions for model validation and calibration are. A quasi-random Monte-Carlo method is the main tool for the analyses. The first step is to reduce the scope of the problem without affecting the validity of the analysis. It was found that the biomass productivity is a monotonic function of the inputs and parameters. Figure 1 shows the reduction of the yearly biomass productivity variance for all the factors if the parameters and inputs deviate within 50% and one of them is fixed in that range. The decision variables were varied and fixed in the ranges explored in the work of (Slegers et al. 2011). It can be seen that biomass productivity is most sensitive to the kinetic parameters, adsorption coefficient, and decision variables.
Original languageEnglish
Publication statusPublished - 2012
EventNBC 14 - Ede, Netherlands
Duration: 16 Apr 201218 Apr 2012


ConferenceNBC 14


Dive into the research topics of 'Uncertainty and sensitivity analysis of algae production models'. Together they form a unique fingerprint.

Cite this