Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)


Civil Engineering


P.L. Dold


The objective of this study was to establish the general framework of a comprehensive dynamic mechanistic mathematical model for predicting the behaviour of an activated sludge plant treating petroleum/petrochemical industrial wastewaters. Specifically, the aim was to develop two activated sludge mechanistic models which predicted the behaviour of systems treating: (1) petroleum refinery wastewater, and (2) petrochemical refinery wastewater. The wastewater treatment plants at two industrial facilities in Ontario were used to develop these models.

A review of the literature identified that mechanistic models were not available for industrial activated sludge plants, but there was extensive information on the development and application of such model to municipal treatment systems i.e. the ASMI model developed by the IAWPRC (Henza et al., 1987). The development of models for the municipal systems served as a valuable resource for developing these models, but extensive differences between the two types of wastewaters meant that it could not be applied directly to predicting the behaviour of these systems.

The premise for developing these two industrial models was based on the fact that the types of contaminants treated in petroleum and petrochemical refineries have similar characteristics. Both types of systems are based on the processing of crude oil or its derivatives. Therefore, if the contaminants have similar properties, it was hypothesized that the removal mechanisms of the different compounds would also be similar. For example, removal precesses such as volatilization, sulphur oxidation, degradation of oils and greases, and degradation of inhibitory compounds such as phenolics are characteristic of these industrial facilities. As a result it was possible to develop a model structure which would apply t to the activated sludge treatment of various petroleum/petrochemical wastewaters. This study illustrates these points as they apply to the two selected facilities.

In general, development of a mechanistic model involved:

(1) Identifying model components; and, (2) Identifying model processes.

Once these model elements were determined, the model was formulated into a mathematical structure. In terms of the development of an activated sludge mechanistic model the following steps were identified as necessary parts of the process and were applied to the development of each of the activated sludge models in this study:

(1) Selection of model structure and complexity, (2) Gathering of experimental data, (3) Mathematical model formulation, and (4) Model calibration and verification.

The first stage in the study was the development of the petroleum refinery model in order to establish the model structure. In an earlier study (Baker, 1993), an experimental programme designed especially to provide the data necessary for model development was conducted. Using this data the model structure was developed for activated sludge treatment of petroleum refinery wastewaters. Once the model structure was in place, the data were used to calibrate the model predictions to the known system behaviour monitored during the experimental programme. This involved determining values for the different stoichiometric and kinetic coefficients included in the model. By adjusting the values of the parameters until they match observed system behaviour, a single set of values was established which enabled the model to predict adequately system behaviour.

With the experience gained during development of the petroleum refinery model, development of the petrochemical model followed a different approach. The petroleum refinery model provided a structure on which to base the petrochemical model. Therefore, only limited experimentation was conducted to collect the necessary data to determine the important components and processes to include in the model. This experimental programme was performed at the petrochemical facility, and data collected during this study were supplemented with data monitored on the full-scale system. The core of the experimentation involved performing a series of batch tests in which activated sludge was mixed with individual influent streams and the response of the mixture with respect to OUR, filtered COD, and filtered TKN was measured over time. These tests provided data on the kinetic response of the petrochemical system. This data was used to calibrate the model to the behaviour of the petrochemical facility. The result was a model capable of predicting the response of the petrochemical system with reasonable accuracy.

A final stage of the study was to illustrate the advantages of having modelling tools of these systems. One key advantage is to be able to simulate system behaviour on a computer scale before testing operating strategies on a full-scale system. To illustrate a practical application of these models, the petrochemical model was used to determine those modes of operation which provided better system response in terms of effluent COD and VOC air emissions for this particular plant. A range of operating configurations and operating conditions were simulated in the model. Another application of activated sludge models which was not considered in this study was to assist in the design of new wastewater treatment systems.

In conclusion, this study developed two activated sludge models for the treatment of (1) the petroleum refinery system, and (2) the petrochemical refinery system. It illustrated the methodology required to develop models of these type of industrial wastewater treatment facilities. With the framework established for the treatment of petroleum based wastewaters, the development of a model for a different petroleum/petrochemical treatment systems can be accomplished with limited experimentation. The approach applied to the development of the petrochemical activated sludge system model can be utilized. This was a valuable contribution of this work.

In achieving these objectives, the study also consolidated current knowledge on the response of these systems. As a result, a better understanding of system behaviour and the important removal mechanisms associated with treatment of petroleum/petrochemical wastewaters was attained. The differences between municipal wastewater treatment and petroleum wastewater treatment became apparent during the course of the study and confirmed the need for separate models from the IAWPRC ASMI model.

Files over 3MB may be slow to open. For best results, right-click and select "save as..."