logo
                           
Majid Mohammadian, Ph.D., P.Eng. 
Associate Professor                      

Department of Civil Engineering 
 
         

Research Group on Desalination and Industrial Outfalls (RDIO)         

 Collaborators (uOttawa): Drs Nistor Rennie Seidou 


 Projects

1- Environmental impact assessment of major desalination plants in Qatar under changing climate

2-  Mixing study in North Saskatchewan River, Canada

     2.1  Mixing study of the outfall of Agrium Redwater (North Saskatchewan River, Canada)

     2.2  Mixing study of the outfall of Capital region Waste Water Treatment plant (North Saskatchewan River , Canada)

     2.3  Mixing study of the outfall of Goldbar Waste Water Treatment plant (North Saskatchewan River, Canada)

3-   Three Dimensional numerical simulation of submerged outfalls

4-   Three Dimensional numerical simulation of surface outfalls

5-   Three dimensional numerical simulation of thermal outfall of a Petrochemical Complex in Persian Gulf

6-   Numerical Modeling of Mixing and Transport of Ammonia in Ottawa River, Canada

7-   Nitrification Kinetics & Modified Model for the Rideau River, Canada

8-   Biosorption of Pb and Cu Using Fixed and Suspended Bacteria

9-   Hydrodynamic modeling of Mooney's Bay, Canada

 

Environmental impact assessment of major desalination plants in Qatar under changing climate

Project Director: Majid Mohammadian
Students: Hossein Kheirkhah, Vahid Pilechi, Amir Gharavi, Hamidreza Shirkhani
Collaborators: Gregory Lawrence, Ioan Nistor, Hazim Qiblawey, Colin Rennie, Ousmane Seidou


a. Project summary  

Desalinated sea water is the main source of potable water in Qatar. Nearly 60% of desalination plants in the world are located in the Persian Gulf, and some of the largest are located in Qatar. Evaluation of the environmental impacts of desalination plants is an important issue, as the increased temperature and salinity of the discharge may cause problems for the ecosystem. In order to protect aquatic life and the ecosystem, there are strict international regulations on the level of required dilution of the concentrated outfall of desalination plants. For example, strict criteria for discharge salinity have been set by the US Environment Protection Agency; the salinity within 50 metres of the discharge point must be within plus or minus 1.2 parts per thousand (ppt) of background levels. By the time the discharge is more than a kilometre offshore, salinity must be within 0.8 ppt of background levels. Similar criteria exist for the temperature of the discharge.

This research focuses on two selected plants in Qatar where field measurements, laboratory experiments, and numerical simulation are used to study dispersion and dilution of the concentrated brine. The results are analyzed and compared with the available regulations.
In order to evaluate the future environmental impacts of desalination plants under a changing climate, various climate change scenarios will be considered and the projected results of general circulation models will be downscaled and employed in numerical simulations.
 
  (Fig. 1) Dispersion of hypersaline brines (left) and upwelling of warm plumes (right)
 
b. Numerical modeling  

A three-level flow model was set up and run for the spring tidal conditions of February 2012. The model is based on the domain decomposition concept and uses two-way communications between different levels. The largest model (Level 1) is a 2D model which starts from the Strait of Hormuz and covers the entire Gulf by a Cartesian structured grid with 3km resolution. The water level boundary condition used in this model was obtained from IHO buoy data at Farur and Siri Islands, Lengeh port, and Ghanada estuary (Fig.1).
 
(Fig. 2) Model level 1 extension and the location of the available IAHR tidal buoys
 
The model results were calibrated and verified using IHO buoy data for Halul and Umm Said respectively.
 
(Fig. 3) Results of calibrated numerical model (continuous line) and IHO gauge data at Halul (left) and Umm Said (right) stations

The intermediate-level model (Level 2) is nested in Model Level 1 and covers Qatar’s eastern coasts (where the selected sites located) with 500m grids. This model is also 2D and is used to simulate the tidal currents in the region of interest (Fig. 4).
 
                                 
       (Fig. 4)
 
The intermediate-level model (Level 2) is nested in Model Level 1 and covers Qatar’s eastern coasts (where the selected sites located) with 500m grids. This model is also 2D and is used to simulate the tidal currents in the region of interest (Fig. 4).
The results were calibrated and verified using IHO buoy data for Jabal Al Fuwayrit and Al Wakrah respectively.
 
 
                                  
(Fig. 5) Simulation result from Level 3 (left) for salinity
dispersion under tidal conditions
  (Fig. 6) Research Vessel Janan (Qatar University) used for field study
 
In the next step, a wave model was set up and coupled with the flow model. The wave model was set up in 2 Levels on the same grid as Level 1 and Level 2 of the flow model. The results of the coupled wave-flow model were also calibrated with the buoy data. We have also investigated the effect of a multiport diffuser on maximizing the near field mixing (Fig. 5).

c. Field work: The field work is currently carried out using following equipment  

A CTD is primarily designed to measure salinity, temperature, and depth by using conductivity, temperature, and pressure sensors respectively. Due to recent technological developments and the invention of new in situ sensors, the CTD utilized in the present study integrated an additional fluorometer and turbidity sensor. A CTD is lowered into the water and continuously records data from each of its sensors. The fluorometer used in the present study is suitable for the detection of Rhodamine WT concentrations ranging from 0.01 ppb to 230 ppb with 0.01 ppb sensitivity (i.e., resolution). Rhodamine WT concentrations were recorded with a sampling frequency of 4Hz.
 
          (Fig. 7) Seabird CTD SBE19plusV2                      (Fig. 8) ADCP
 
Depth-averaged velocity and bed elevation were measured using a SonTek M9 RiverSurveyor ADCP. An ADCP measures spatial averages of the three principal water velocity components in individual “bins” throughout a vertical column of water, and can be operated from a moving vessel. ADCPs use the Doppler principle to measure water velocities, utilizing four acoustic transducers. Each transducer is orthogonal to the others (spaced at 90º around the circle), and is angled from the vertical (at 25º for the present instrument). From each of the four transducers, an acoustic pulse (“ping”) is transmitted. The acoustic pulse scatters off particles in the fluid, which are assumed to be traveling at the velocity of the fluid. The frequency of the scattered sound is changed due to the Doppler shift related to the velocity of the particles. The backscattered sound is received by the transducer, and the along-beam component of the particle (fluid) velocity is estimated based on the observed Doppler shift. Furthermore, the backscattered sound is “range-gated”, meaning it is processed in sequences, which allows for determination of velocity at sequential depths (“bins”) below the instrument.  A coordinate transform allows for estimation of the three Cartesian velocity components at all depths below the instrument. A separate “bottom track” (Doppler sonar) pulse is used to measure the local water depth and the speed of the vessel.  Lastly, an on-board compass allows for rotation of the measured velocities into Earth coordinates (East, North, Up). 
A survey-grade dual frequency real-time kinematic (RTK) Global Differential Positioning System is employed on the boat to locate the measurements. The GPS was manufactured by Novatel, and included a Novatel DL-V3-L1L2 base receiver and a NovAtel ProPak LB Plus rover receiver with reported relative horizontal position accuracy of ±2 cm CEP (i.e., 50% of position estimates have an error < 2 cm). The precision of the RTK-DGPS system was previously evaluated by Rennie and Rainville (2006), wherein the average error of measured RTK-DGPS velocity equalled 2.6 cm/s. Position data were collected at 10 Hz and integrated into the CTD and ADCP data sets for correct positioning and synchronization of the CTD and ADCP data. 
 
                   (Fig. 9) GPS-RTK system                  (Fig. 10) Anemometer
 
Wind speed and direction were measured using a WindSonic anemometer. This instrument includes the anemometer sensor, a data logger, and a solar battery. The data were collected at 4Hz frequency with 0.01m/s and 1º resolution for velocity and direction respectively.

Mixing study in the North Saskatchewan River, Canada

Student: Vahid Pilechi
Supervisor: Majid Mohammadian
Co-supervisor: Colin Rennie
Collaborator: David Zhu

 
The objective of this project was to measure the spatially distributed tracer concentration and water velocity in the mixing zone so that dispersion patterns could be determined.Rhodamine WT was used a tracer of plant effluent. The tracer concentration was tracked from a boat moving both across and along the river using an in situ fluorometer integrated into a Conductivity-Temperature-Depth (CTD) instrument. Physical water samples were also collected at various locations in the river for subsequent laboratory analysis of the rhodamine concentrations. The measured physical sample rhodamine concentrations suitably verified the in situ fluorometer results.
 
(Fig. 11)  Peristaltic pump for very slow injection of rhodamine into the outfall
 
The river hydrodynamics were also measured using a RiverSurveyor Acoustic Doppler Current Profiler (ADCP). The ADCP was mounted to the survey vessel after being specifically selected for surveying the relatively shallow depths of the North Saskatchewan River in October. An anemometer was also set up near the outfall to monitor wind speed and direction.
The CTD and ADCP data were collected simultaneously and synchronized by means of Global Positioning System (GPS) position and time-stamp data that were integrated into the data streams collected by the CTD and ADCP. The survey was conducted by collecting data along multiple transects in each reach.
The rhodamine concentration lessened gradually along the river path as the tracer mixed with ambient water.

Mixing study of the outfall of Agrium Redwater (North Saskatchewan River, Canada)

Student: Vahid Pilechi
Supervisor: Majid Mohammadian
Co-supervisor: Colin Rennie
Collaborator: David Zhu


Agrium Redwater is a fertilizer-producing plant with an approximate outfall discharge of 4600m3/day into the North Saskatchewan River.
 

The gradual mixing of effluent with ambient water within the surveyed distance is observed in the Agrium results. At 12 km from the outfall location, the river flow separates due to the presence of an island in the middle of the river section. The consequent complex flow pattern leads to a uniform rhodamine concentration distribution across the last surveyed section in the right-hand channel. However, the left-hand channel at 11.5 km did not display a uniform concentration distribution, and thus full mixing was not observed.
 
(Fig. 13) Mixing trend along the river path from the Agrium survey results, based on mean rhodamine concentrations in each section.  Location across the section is non-dimensionalized such that the left bank is at 0 and the left bank is at 1
 

Mixing study of the outfall of Capital Region Wastewater Treatment Plant (North Saskatchewan River, Canada)

Student: Vahid Pilechi
Supervisor: Majid Mohammadian
Co-supervisor: Colin Rennie
Collaborator: David Zhu

 
The Capital Region Wastewater Treatment Plant (WWTP) is the fourth-largest treatment facility in Alberta. It treats the wastewater and sewage of more than 200,000 people each day, and the current capacity of this plant is 140 ML/d.
 
The outfall is located on the right bank side of the river, so the rhodamine concentration was higher on the left bank side in sections closer to the outfall. The rhodamine concentration was 0 on the left bank side in the initial sections. It required about 6 km for the rhodamine to first reach the left bank.
Following the river path, the rhodamine gradually mixed and the difference between the rhodamine concentration on the left and right bank sides lessened. In the last surveyed section, at about 83 km from the outfall, the average rhodamine concentration on the left bank side was 0.83ppb versus 0.56ppb on the right bank side.
For the Capital Region WWTP the 100ppb difference between the rhodamine concentrations on the right and left bank sides decreased continuously along the river. As previously mentioned, because the rhodamine was injected from the right bank into the river, it was highly concentrated on the right bank side of the river and the measured concentration on the left bank side was 0 for the initial sections. The rhodamine reached the left bank at the section 6km from the outfall. The high rhodamine concentration difference between the left bank and right bank sides lessened from 100 ppb at the outfall to 0.27 ppb in the 83km section. However, this value is still 46% of the average rhodamine concentration in the 83km section, which means that the full mixing condition was not achieved.

 
(Fig. 15) Mixing trend along the river path (outfall to 6km section) from Capital Region WWTP survey results, based on mean rhodamine concentrations in each section. Location across the section is non-dimensionalized such that the right bank is at 0 and the left bank is at 1
 
(Fig. 16) Mixing trend along river path (9.4km section to 83km section) from Capital Region WWTP survey results, based on mean rhodamine concentrations at each section. Location across the section is non-dimensionalized such that the right bank is at 0 and the left bank is at 1

Mixing study of the outfall of the Gold Bar Wastewater Treatment Plant (North Saskatchewan River, Canada)

Student: Vahid Pilechi
Supervisor: Majid Mohammadian
Co-supervisor: Colin Rennie
Collaborator: David Zhu
 

The Gold Bar WWTP is located in the North Saskatchewan River Valley. It treats the wastewater for 700,000 people. The current capacity of this plant is 82MGD (310 ML/d).
(Fig. 17) Gold Bar WWTP site locationleft picture from Drainage Information Systems Drainage Services, and site location (right) (right picture adopted from Google Earth)

For the Gold Bar WWTP the outfall location is located on the right bank, and thus in the initial sections the rhodamine was concentrated on the right bank side of the river, and the concentration was 0 on the left bank side. The first section in which the rhodamine concentration was observed on the left bank side of the river was located at approximately 11.6km from the outfall location.
Downstream of where the effluent reaches the left side of the river, the measured concentration continues to be higher on the right bank side compared to the left. The high difference in concentration between the right bank and left bank at the outfall (about 13 ppb) lessened to 0.11 ppb in the 80 km section. For the 92 km section, the difference lessened to 0.05 ppb. However, the 95% confidence intervals for concentrations at the right bank and left bank in the 92km section do not overlap, so it appears that full mixing still has not been achieved. Still, the difference across the section of 0.05 ppb is only 20% of the average sectional concentration of 0.23 ppb. 

 
(Fig. 18) Mixing trend along the river path (11.6km to 92km section) from the Gold Bar WWTP survey results, based on mean rhodamine concentrations in each section. Location across the section is non-dimensionalized such that the right bank is at 0 and the left bank is at 1

Three-dimensional numerical simulation of submerged outfalls

Student: Hossein Kheirkhah
Supervisor: Majid Mohammadian
Co-supervisor: Ioan Nistor

 
Effluent discharges from industrial power plants into coastal and river waters have increased dramatically during the past few decades. It is a major concern in developing countries located in the Persian Gulf region, where most use desalination plants to provide sufficient potable water for their populations. Environmental impacts of effluent discharges on the recipient ecosystem are unavoidable. Therefore, the mixing and dispersion characteristics of the different jets resulting from the outfalls have to be studied accurately. As a part of the Qatar Project, the near-field area of the discharge point has been numerically modeled using the OpenFOAM CFD package. This model works based on the Finite Volume Method (FVM). Various RANS turbulence models have been applied in the numerical model to find the best ones to predict the geometrical and mixing characteristics of both positively and negatively buoyant jets issued at an angle into ambient water. Very simple turbulence models such as standard k-ε have been examined as well as more advanced models like the Reynolds Stress Model (e.g. LRR model), which have not been investigated for similar problems before. The numerical results showed good agreement with the experimental data for the models using realizable k-ε and RSM turbulence models.
 
(Fig. 19) Velocity vector map of LRR model for a positively buoyant jet. (a) tank width=0.4 m.( b) tank width=0.8 m. (c) tank width=1.2 m.
           
(Fig. 20) Velocity vector and concentration contour maps for a:30° and b:45° inclined jets (realizable k-ε turbulence model)

Three-dimensional numerical simulation of surface outfalls

Under construction

Three-dimensional numerical simulation of thermal outfall of a Petrochemical Complex in Persian Gulf

Students: Peter Bradley, Joel Hernberger, Desmond Lam, Jeffrey Miller, Michael Pelletier
Supervisor: Majid Mohammadian

 
 (Fig. 21) Dispersion of thermal plumes of a plant in the Persian Gulf

Numerical Modeling of Mixing and Transport of Ammonia in the Ottawa River, Canada

Student: Ivana Vouk
Supervisor: Majid Mohammadian
Co-supervisors: Colin Rennie, Robert Delatolla
           
(Fig. 22) 
 
(Fig. 23)  
 
(Fig. 24) 
 
(Fig. 25)

Nitrification Kinetics & Modified Model for the Rideau River, Canada

Student: Lianmiao Zhao
Supervisor: Robert Delatolla
Co-supervisor: Majid Mohammadian

 
Improving the kinetic modeling of nitrification in rivers is of growing importance due to yearly increases in the anthropogenic release of nitrogen into rivers around the world. The use of water quality models can lessen the expense of water quality monitoring while enabling the user to predict trends of variation. Data collected from a series of laboratory kinetic experiments were used to calculate the rate of nitrification in the Rideau River, Canada, and to modify the nitrification algorithms used in the traditional water quality model Qual2E. The modified model relates the reaction rate coefficients with a simple biomass concentration measurement of volatile suspended solids (VSS) in the river and subsequently introduces biomass growth functions directly into the kinetic algorithm. Furthermore, this modified model includes a nitrate-nitrogen assimilation pathway. The model demonstrates an improved correlation to nitrogen parameters observed in river water samples compared with the classical water quality model Qual2E. The inclusion of bacterial concentrations based upon the simple measurement of VSS plays a critical role in the reactions of the nitrification system, and nitrate-nitrogen assimilation is an important pathway at low ammonia-nitrogen concentrations.
        (Fig. 26) Schematic of two-step nitrification reaction process
           
(Fig. 27) Schematics of bench-scale batch reactors, (a) model calibration experiments and (b) model validation experiments.
 
(Fig. 28) Model validation
 

Biosorption of Pb and Cu Using Fixed and Suspended Bacteria

Student: Ryan Black
Supervisor: Majid Mohammadian
Co-supervisor: Majid Sartaj

 
Biosorbents are a low-cost option for heavy metal (HM) remediation. To compare HM adsorption capabilities between fixed and suspended morphologies of bacteria, batch adsorption tests of Cu and Pb were performed at 22oC and a neutral pH. A mixed bed bioreactor (MBBR) and a batch reactor were used to develop the fixed biofilm bacteria and the suspended planktonic bacteria, respectively. The dominant bacteria were identified by 16s RNA sequencing as Enterobacter ludwigii, Zoogloea ramigerais, and Comamonas testosteroni. The two morphologies exhibited significant differences in percentage of dry weight and in EPS content.
The fixed bacteria, grown on carriers in a MBBR, displayed 29% EPS content by weight compared to 9.5% EPS found on the suspended bacteria which were grown without substrate in a batch reactor. The adsorption data fit the Freundlich isotherm for all cases (R 2 values were 0.89 and 0.97 for Pb and 0.96 and 0.99 for Cu for fixed and suspended biomass, respectively).
The adsorption data fit the Langmuir isotherm for all cases except Pb on suspended biomass (the R2 value was 0.91 for Pb on fixed biomass, and the R2 values were 0.88 and 0.99 for Cu on fixed and suspended biomass, respectively). The Langmuir adsorption data for Cu revealed that the q max was greater for fixed bacteria, at 9.80mg g -1, compared to 6.52mg g -1 for suspended bacteria. The isotherm constants suggest that the fixed bacteria exhibit a greater binding affinity for both HMs and that both morphologies have a degree of heterogeneity amongst binding sites.
 
a) Image of fixed bacterial biomass community taken at 40x magnification detached from the k1 carrier b) Image of Zoogloea ramigerais rod-shaped bacterial cells taken at 100x magnification. c) Image of Enterobacter ludwigii rod-shaped bacterial cells taken at 100x magnification. d) Image of Comamonas testosteroni rod-shaped bacterial cells taken at 100x magnification. Bacteria colony isolates grown on nutrient agar.

 
(Fig. 29) Freundlich isotherms comparing the adsorption capacity (q eq), of suspended biomass and fixed biomass at different equilibrium HM concentrations (c eq). Fig a) is Cu adsorption and Fig b) is Pb adsorption

Hydrodynamic modeling of Mooney's Bay, Canada

Student: S. Kheradmand
Project Director: Majid Mohammadian
Collaborators:  O. Seidou, C. Rennie, T. Odyemi

 
A shallow-water model was developed for Mooney’s Bay on the Rideau River. Due to the complex geometry of the area, an unstructured grid has been employed. A high-resolution upwind finite volume method based on characteristic decomposition is used to solve the depth-averaged equations. A field survey has been also performed, and velocity field and bathymetry have been measured using an Acoustic Doppler Current Profiler (ADCP). The flow characteristics in the bay have been analyzed and the numerical results compared with measured data. The present model is the first step towards a coupled flow and ice development simulation for Mooney’s Bay.
  
(Fig. 30)