The following is a summary of the thesis entitled 'Optimisation of Biogas
Production By Advanced Process Monitoring and the Effect of Mixing Frequency on Biogas Reactors With and Without Microbial Support Media' submitted to the School of Chemistry, University of Bangor, Wales, UK in fulfilment of the degree of Doctor of Philosophy.
In the first chapter, current literature was reviewed and specific areas where
research was lacking were highlighted. Some of theses research areas were selected and formed the basis for the subsequent chapters. A shortened verson of the review has been accepted for publication in the international journal Bioresource Technology.
The second chapter described the construction of pilot scale anaerobic digesters,
including a four-stage design, two two-stage designs and four single-stage designs. The four-stage digester was operated continuously for ten months under a variety of loads to gain an insight into the behaviour of digesters during start-up, failure, recovery and stable operation. The hydrodynamics of the system were also examined by a tracer study. This gave valuable information on the mixing and flow patterns within the digester vessels. The single and two-stage digesters were built for investigation into the effect of mixing frequency on digesters with and without microbial support media fitted.
In chapter 3, data collected from the four-stage system was statistically analysed.
Curve fitting non-linear regression models were created to estimate the optima of each measured parameter in terms of both methane production and yield. Total alkalinity was found to be a good indicator of process stability and also gas production rate with an R2 of 64.3 % and standard error of observations of 0.254 L.L -1 d-1•
Chapter 4 concentrated on a regression model which could estimate total
alkalinity through data obtained by cheap and simple to maintain pH, redox and
conductivity probes, i.e. a software sensor. The selected prediction model had an R2 value of 71 % and root mean square error of 1441 mg.L ·The software sensor estimate was used by a rule based supervisory control system which controlled the system stability and gas production by modulating the input feed rate to maintain an optimal alkalinity.
Chapter 5 examined an alternative method of estimating bicarbonate alkalinity
using Fourier Transform Near Infrared Spectroscopy (Fr-NIRS). A good estimate of
total alkalinity was made by optimisation using spectral pre-processing and partial least squares analysis of the alkalinity titration data, with an R2 of 87 % and root mean square error of prediction of 1230 mg.L·', suggesting this method could also b~ developed into an on-line monitoring system.
In chapter 6 the effect of mixing frequency on single and two-stage anaerobic
digesters with reticulated polyurethane foam biomass support media was investigated.
The work included quantification of extracellular polymeric substances (EPS) which form the structural component of microbial aggregates was performed in addition to measuring the effect on the methane production rate. It was found that mixing frequency had the most effect on methane production with as much as 26 % increase, suggesting that simple optimjsation experiments in industrial scale processes could be economically beneficial.