Complex Dynamic Numerical Model Development for Simulating the Anaerobic Co-Digestion of Fruits, Vegetable and Fatty Oil in a Semi-Continuous Stirred-Tank Reactor (CSTR)

  • Hatem Yazidi

Student thesis: Doctoral ThesisDoctor of Philosophy (PhD)

Abstract

Municipal solid waste (MSW) generation is a global challenge. As part o f its main national strategy for attaining a sustainable future, the Sultanate o f Oman government encouraged, via its research council, the development o f scientific solutions towards the minimization o f MSW.

This thesis investigated the anaerobic co-digestion process, which is considered to be a reliable medium to transfer organic components o f MSW into energy. To this end, the development o f a numerical estimation and optimisation platform led to the creation o f a flexible, user-friendly tool. Furthermore, the complex dynamic model (CDM) developed in this thesis permitted the creation of a useful prediction tool to investigate biogas production.

The mathematical approaches frequently used to estimate anaerobic co-digestion model parameters are based on both one-at-a-time parameter analysis and local sensitivity analysis. These methods do not provide an accurate assessment o f the model uncertainty’and sensitivity as, by default, they are based on the one parameter at a time approach(they keep all other parameters constant). Using the multi-dimensional parameter space approach, described within, it has been demonstrated how this can be applied globally so all uncertainties can be identified. Moreover, in this work, interesting aspects such as the development of a numerical platform for a systematic identification o f the parameters contributing to the anaerobic co-digestion process stability was divulged. The development o f a stand-alone graphical user interface in this thesis made easy the simultaneous use o f all developed program codes without compromising the quality of the results.

The calculation of the significant (oc= 0.05,) partial rank correlation coefficient (PRCC) has shown that the following parameters have contributed the most to the uncertainty of the complex model’s estimate o f cumulative methane production: (1) Polymer hydrolysisrate (Kfl (2) Gas-liquid transfer constant (Klafl and (3) Maximum acetoclastic methanogens growth rate (PmaxaJ- However, the application o f sensitivity analysis (eFAST method) demonstrated that the cumulative methane production is mostly sensitive to the Maximum acetoclastic methanogens growth rate (Umax4).

The comparison between CDM and ADM1 (The International Water Association Anaerobic Digestion Model No.l) showed high level o f similarities. Also, the numerical estimations o f cumulative methane production, using both models CDM and ADM1 fitted well the experimental data with statistically significant Pearson correlations ranging between 0.9 and 0.95 (1.5e-25<p-value<3.2e-15). furthermore, both models were able to predict reasonably well volatile fatty acids, or precisely, propionic acid and acetic acid.
Date of Award2019
Original languageEnglish
Awarding Institution
  • Glasgow Caledonian University
SupervisorGeraint Bevan (Supervisor), Ole Pahl (Supervisor) & Colin Hunter (Supervisor)

Cite this

'