<p>Owing to the rise in fossil fuel prices, overall energy security concerns, and the current push towards green engineering; renewable and green fuels have seen an increase in interest in recent years. Two notable technologies in this green movement are the production of biodiesel from microalgae and the production of biogas from anaerobic digestion of waste biomass. Production of biodiesel from microalgae was studied extensively in the 80s through the early 90s and found to be economically infeasible given the technology of the time. However, recent literature has suggested that one possible method to improve the feasability of the process would be to combine it with an anaerobic digestor to provide nutrient and biomass recycling. For such a system, having accurate models of each process would be highly advantageous for optimal design and control. To this end this thesis moves towards this overall goal by examining and modelling the anaerobic digestion of the microalgae <em>Chlorella vulgaris</em>.</p> <p>Starting with a set of experimental data (anaerobic digestion of <em>Chlorella vulgaris</em>) provided by LBE-INRA, the minimum number of kinetic equations needed to predict the data are found using principal component style analysis. This number is found to be two to three reactions. Using this as a basis for model development, a mass balance model is developed around both two and three reaction cases. To date there is very little literature on the modelling of anaerobic digestion of microalgae and so kinetic laws are selected from the general anaerobic digestion models ``Anaerobic Digestor Model 1'' (ADM1) and ``Acidogenesis/Methanogenisis Model'' (AM2). Given that the kinetic laws were derived from general literature, model fitting is a must. To faciliate this process a novel systematic parameter identification procedure to locate identifiable parameter subsets within each model is presented. Applying this novel procedure to the provided data is seen to lead to promising identification results. Through these identification trials it is shown that the three reaction model best captures the dynamics of the system. This three reaction model serves as the basis for subsequent steady state optimality and sensitivity analysis. From these efforts it is shown that the predicted optimal curves match literature data very well but uncertainty in certain key parameter estimates lead to highly sensitive model predictions (and therefore low confidence). This leads to the conclusion that the developed model is capable of predicting the kinetics of <em>Chlorella</em> digestion but additional trials are needed to further refine the model fitting results.</p> <p>Coupling an anaerobic digester to a microalgal culture is currently considered one of the most promising avenues towards the production of renewable bioenergy, either in the form of biodiesel or biogas. Accurate mathematical models are crucial tools to assess the potential of such coupled biotechnological processes and help optimize their design, operation and control. This paper focuses on the compartment of anaerobic digestion of microalgae. Using experimental data for the anaerobic digestion of <em>Chlorella vulgaris</em>, a grey-box model is developed that allows good prediction capabilities and retains low complexity. The proposed methodology proceeds in two steps, namely a structural and a parametric identification steps. The fitted model is then used to conduct preliminary optimization for the production of biogas from <em>Chlorella vulgaris</em>. The results provide some insight into the potential for bioenergy production from the digestion of microalgae and, more generally, the coupled process.</p> / Master of Applied Science (MASc)
| Identifer | oai:union.ndltd.org:mcmaster.ca/oai:macsphere.mcmaster.ca:11375/11913 |
| Date | 04 1900 |
| Creators | Cameron, Elliot T. |
| Contributors | Chachuat, Benoit, Dr. Thomas A. Adams II, Dr. Carlos Filipe, Chemical Engineering |
| Source Sets | McMaster University |
| Detected Language | English |
| Type | thesis |
Page generated in 0.0014 seconds