Water & Waste Treatment

ODOURsim® modelling of wastewater treatment works improves predictions

Professor Tom Stephenson
First Published in Water & Waste Treatment magazine, September 2008.

In this article we focus on ODOURsim®, dynamic odour emission modelling software for wastewater treatment works (WwTW) developed by Water Innovate Limited. We describe the purpose of the model, detail its capabilities, and identify potential users who will find this software a valuable tool for treatment plant design and upgrade. Having grown from the internationally respected Centre for Water Science, Cranfield University, ODOURsim® is an example of an academic idea developed into a commercial product for the water sector.

ODOURsim is designed to provide atmospheric dispersion models (ADM) with more accurate input data by considering the impact that temporally and spatially variable parameters such as flow rate have on odour formation and emission. By modelling these variable formation and emission characteristics, ODOURsim can produce dynamic models that accurately reflect diurnal treatment cycles. Without this dynamic modelling, ADM would normally assume unrealistic steady-state odour generation rates from WwTW leading to simplistic and inaccurate odour impact predictions. We explain how sound science and expertise in the field of water and environmental waste underpins the ODOURsim model. The calculations used by this model are outlined, and the model outputs are described.

Successful trials of ODOURsim at a major UK utility WwTW are detailed. We describe model validation and ground-truthing at this site, and how the dynamic model output has reinforced the confidence in odour impact predictions here. The value of anticipating threshold exceedence and evaluating proposed abatement cost effectiveness at the site is explained.

Odour modelling software

ODOURsim is a Windows-based software package designed for use by environmental consultants, plant designers, civil engineers and software houses as a tool for their development of odour management and abatement strategies. Predictions can be used to support planning applications. The software facilitates an accurate assessment existing problem odour generation at a WwTW, and proposed plant improvements can be pre-tested to evaluate the effectiveness in reducing odour impacts and improving upon air pollution control.

Using wastewater characteristic data that are easy to measure or can be reliably estimated, dynamic emission modelling for an existing or proposed WwTW is achievable. These predictions are valuable in their contribution to the development of a robust and defensible odour management plan by operators. ODOURsim enables operators to assess the impact of a proposed change to a plant before capital is spent.

Using Hydrogen Sulphide (H2S) to model formation and emission of odours, ODOURsim interfaces with standard ADM with the unique feature of producing dynamic predictions that are specific to each WwTW. For example, odour predictions will vary temporally with changes of influent loading, treatment works hydraulics and meteorological conditions. Similarly, as the characteristics of different assets varies during the day, the spatial odour map will also vary. Realistic dynamic modelling informed by these wastewater variables produces accurate modelling of variable odour generation rates, allowing equally realistic diurnally varying contour plots of odour emissions to be generated. This is a significant improvement on simple ADM models that assume steady-state odour emission rates.

Dynamic odour generation modelling is vital in effectively predicting odour impact. For example, hydrogen sulphide generation rates increase significantly under hydraulic conditions in which anaerobic conditions can arise. Figure 1 illustrates how low influent loads with aerobic conditions are predicted to produced lower H2S odours compared to higher influent loads where anaerobic conditions can occur. The assumption of a constant odour emission rate is clearly inadequate.

Figure 1
ODOURsim model outputs show how changing hydraulic conditions at the same location can change odour generation rates. On the left, under high influent loads anaerobic conditions develop at sedimentation tanks, whilst at non-peak flow time (right), aerobic conditions mean H2S generation is significantly reduced.

Various WwTW technologies and processes can be incorporated into an ODOURsim model. The appropriate physical dimensions, design capacities and other process details are input to the model together with meteorological data. After computation, the model output of odour impact can be presented in several ways, including graphically or as exceedence contour maps identifying precise locations within WwTW where the odour sources are.

The ODOURsim® software comes with a powerful user-friendly Graphical User Interface (GUI). Detailed reports for each simulation undertaken can be generated, and graphical outputs can be tailored to suit the particular requirements of end-users. Data generated by the robust computation module can be integrated into data formats of existing ADM, and the software comes with a recovery feature enabling simulations to be resumed if unexpectedly interrupted, for example, by hardware failure.

The ODOURsim model

ODOURsim® models hydrogen sulphide (H2S) generation as an odour marker compound in two physical phases. Liquid phase simulation is based upon the Hvitved-Jacobsen sewer model with additional calculations to model H2S-to-sulphate oxidation. Both hydraulic and biological processes, including biofilms and suspended biomass, are thus modelled. Secondly, the mass transfer of H2S from the liquid phase to the gas phase (i.e. to the atmosphere) is modelled based on direct research of this process in wastewater. The impact of wastewater hydraulics on gas mass-transfer rates where flows are quiescent or turbulent, such as at drop-weirs, channels, trickling filters or during aeration, are incorporated into the calculations. In this way, specific plant where odour is generated within a WwTW can be identified. Calculations are made on a time-based iterative basis to achieve temporal dynamic modelling.

Users of the software describe their WwTW by breaking down its treatment processes into model process units to accurately reflect their specific treatment systems. For example, flow at an inlet channel can be defined in the model as process units reflecting split and subsequently recombined flows, each flow being linked to a series of treatment stages, such as drop-weirs, transfer to any number of primary settlement tanks, and further to any number of biological oxidation units and so on. Storm flow bypasses can be represented as a linked group of separate process units. Having defined the process units, each process forms a model node, which is built into a graphically connected network. Each upstream node calculation informs the next node downstream. This might sound complicated, but the software has been designed so that simple click and drop mouse actions can quickly build the process on screen.

The specific wastewater characteristics at each WwTW are also used to inform the model. The biodegradable and soluble non-biodegradable fractions of the wastewater, and the suspended solids concentrations, are input variables since biological processes and gas-transfer rates are temperature and pH dependent, these characteristics are also inputted to the model. In practice all required parameters are usually routinely and readily measurable, however, if unavailable for a specific site, ODOURsim® can provide default values. These include COD, total suspended solids, effluent temperature, pH and wastewater flow rates.

Sophisticated time-based calculations of odour marker compound generation rates are made using data sources and process modelling that is highly realistic. The diurnally varying odour generation rates output by ODOURsim can be integrated into a number of widely used ADM to predict meteorological dispersion, resulting in a much improved odour impact assessment as compared to the simplistic constant odour source emissions otherwise utilised by ADM.

Using the model: a case study

A major UK water utility company commissioned Water Innovate Ltd. to model the current and future odour emissions from a WwTW. A key aim was to assess the reduction in odour impact if proposed plant improvements were commissioned. This study used ODOURsim to characterise the formation and emission of odours so that AERMOD ADM could accurately model the dispersion of those odours. Crucially, ODOURsim was able to evaluate the odour impact of two different scenarios. The study also included a robust ground-truthing validation of the ODOURsim model for the current operation. This was evaluated by monitoring actual H2S concentrations at several reference locations at the site and comparing these results to model predictions. A comparison between outputs from the dynamic odour modelling using ODOURsim with ADM outputs using a constant-source emission assumption was also reported.

The current treatment process was broken into a series of process units to construct the model. In outline, inlet flow passed through split-flow detritors. The recombined flow passed to a drop weir, on to high-rate trickling filters, then to primary settlement tanks (PST, the six tanks were modelled as a single settlement node to simplify calculations, although separate modelling is possible). Subsequent process unit nodes included sixteen secondary trickling filters (again modelled as a single node) and eight humus tanks.

The proposed upgraded plant included modifications to the pre-treatment and storm weir plant processes, reconfiguring the PST then splitting the flow from here between the high-rate trickling filters (8%), the secondary trickling filters (55%) and a new activated sludge unit (37%). The proposed improvements also included more extensive final settlement. These various alterations were input to ODOURsim as process nodes. Process unit physical dimensions were taken directly from “as-drawn” site maps, from on-site measurements, and from suppliers for certain proposed new plant items.

Wastewater characterisation included collecting soluble and particulate COD measurements from the crude sewage and return liquors, together with sulphide and pH data over a two-day period. Flow data were collected every 15 minutes via telemetry, permitting average hourly flow rates to be derived. Complete meteorological data for this two day model calibration exercise was purchased commercially from a site proximate to the WwTW. These data were used to inform both ODOURsim and AERMOD.

Ground-truthing the model was conducted by measuring H2S at a series of reference sites (at 1m above the ground) using a Jerome analyser and comparing monitored data with model predictions at the same reference height. Micrometeorological measurements at the reference sites conducted over the two-day calibration exercise showed that model predictions were in good agreement with measured H2S, and that ODOURsim accurately modelled H2S generation rates (Figure 2).

Figure 2
Ground-truthing of the ODOURsim model showed good correlation (C) between monitored H2S (B) concentrations at reference sites across the WwTW and ODOURsim (A).

Having demonstrated a satisfactory calibration, the model was re-run using the calculated diurnal odour generation pattern at the site to cover 8760 hours (1 year). These calculated odour generation rates were input to AERMOD, and 98 percentile odour contour plots were derived for both the current WwTW process and the proposed future scenario (Figure 3).

Figure 3
ODOURsim model simulations showing 98%ile contours of odours from the existing WwTW (left) and in future upon implementing proposed upgrades (right) to reduce odour impacts on local receptors.

Throughout the ground-truthing and calibration study, AERMOD predictions using UKWIR look-up tables (LUT) for constant source emissions for odour were shown to poorly reflect the reality (Figure 4), in contrast to the accurate predictions derived from ODOURsim dynamic modelling. Using LUT data, odour emissions from the WwTW were significantly over-estimated. Use of this standard approach to odour impact modelling would have resulted in a poor impact assessment, low confidence in any proposed plant improvement predictions and higher capital expenditure than necessary.

Figure 4
Emission plots based on LUT constant odour emission rates (B) are inaccurate against actual H2S measured (A) because the values do not correlate (C). Hence, using LUT there was a propensity to over-estimate odour emissions.


Reliable predictive computer odour emission modelling from industrial and municipal WWTW is an important aid to plant managers and designers who are required to assess and abate odour impact. Model outputs can inform decisions on the effectiveness of proposed plant improvements. Odour impact assessment is normally based on modelling the dispersion of an indicator compound assuming a constant emission rate of that compound from the source. This assumption is unsound since WwTW odour emission rates are often highly variable, invalidating such simplified ADM modelling and potentially eroding confidence in odour impact assessments.

ODOURsim is flexible dynamic modelling software that allows users to accurately model their specific WwTW and to calculate temporally variable odour generation rates at specific locations at that site. Underpinned by sound science in calculating the generation of H2S under varying hydraulic conditions which, in turn, influence biological oxidation rates and the mass-transfer of H2S gas to the atmosphere, variable odour source generation rates are more accurately simulated. Data generated can be interfaced with ADM to produce a better refined prediction of odour impact.

The improved predictive capacity of ODOURsim was demonstrated at the WwTW. An extensive ground-truthing exercise verified the model, demonstrating the more realistic modelling achieved by ODOURsim compared with results derived from constant-source LUT assumptions. ODOURsim software clearly indicated that proposed upgrades to the treatment process at the site would result in a significant reduction of its odour impact on local receptors.

Author Details

Professor Tom Stephenson is Technical Director at Water Innovate Limited and Head of the School of Applied Sciences at Cranfield University.

Back to top