An experimental and modelling investigation of wet oxidation of municipal biosolids
This thesis describes the development and validation of detailed kinetic models describing the wet oxidation of municipal biosolids.
The treatment of municipal biosolids, which are produced by wastewater treatment plants, is becoming increasingly important because current disposal methods are not sustainable. This is driving the focus for alternative treatment processes which address the unique problems that municipal biosolids present, in particular their high water content and the presence of pathogens.
Wet oxidation is particularly suited to treat municipal biosolids as it is a liquid phase process, and therefore does not require water removal prior to treatment. The high temperatures involved, above 200°C, also kill the pathogens present, and sterilise the material. While the wet oxidation process has been around for many years, and has been the focus of numerous studies, there are still relatively few models that describe the kinetic behaviour of the wet oxidation of municipal biosolids. Models that characterise the numerous intermediate and reaction end products are extremely useful for prediction and process optimisation, however most kinetic models available in the literature contain just a single intermediate compound.
An experimental programme was developed and three sets of experiments were performed to characterise the municipal biosolids produced by the Rotorua District Council wastewater treatment plant. Their behaviour under wet oxidation was quantified using standard water quality tests. Computer image analysis of the intermediate liquid samples was also performed to determine whether useful information could be obtained from the series of photographs taken of the liquid samples. This analysis showed that the normalised image intensity index that was computed for each sample closely followed the measured and modelled change in total COD.
Using the concentration results from the experimental investigation, kinetic pathways were proposed as part of a kinetic model to describe the degradation observed under wet oxidation. The developed kinetic model contained the states of the liquid and gaseous intermediate and end products that were of particular interest for this application which were soluble Chemical Oxygen Demand (sCOD), Volatile Fatty Acids (VFAs) and Dissolved Organic Nitrogen (DON). The model compounds were chosen because they represented participating chemical species which have significant influence on the economics and operability of the downstream biological wastewater treatment process in this study. Without an accurate prediction of these chemical species, the balance of the downstream biological system can be upset. This can cause the microorganisms in the treatment plant to die off, which causes the process to turn septic and stop functioning.
Regressing the kinetic parameters from the experimental data initially produced suboptimal results since the model could not replicate the observed output in response to changes in environmental conditions. However after adding additional constraints, and a careful analysis of the solution process and kinetic pathways, this kinetic model was extended such that it could predict the key compounds of interest under different reaction conditions, as well as the prediction of oxidant consumption.
The developed kinetic model was regressed using experimental data from the lab scale wet oxidation system which used a 600ml stirred batch reactor. Then, the previously regressed model was validated with experimental data from a pilot plant wet oxidation system incorporating the substantially larger 300l bubble column reactor. The model still gave good results using the fitted parameters from the lab scale experiments. The quality of fit for both systems indicates that the kinetic model has successfully captured the reaction kinetics and the effects of important environmental conditions for this specific feed material, with an overall R2 of 0.925. This was further reinforced by a correlation and statistical analysis which confirmed that the model is statistically significant and not over-parametrised.
The detailed dynamic kinetic model from this research allows the user to establish optimal operation. The model was used to identify optimal operating conditions and showed that for a 2 hour residence time, a temperature of 230°C and 16 bar oxygen partial pressure maximised the concentration of VFAs in the reactor effluent.