The dynRAS model is a model developed to simulate recirculating aquaculture systems (RAS) with a particular emphasis on alkalinity and pH control. It is build on prior research effort (Pedersen 2018 and Wik 2008) with new functionalities, including the incorporation of the carbonate system to accurately represent alkalinity and pH fluctuations. Developed within the scope of the RASTOOLS project, the model aims to investigate the dynamic of RAS. The RAS components included in dynRAS are simplified to the fish tank, biofilter, and degasser, based on an experimental set-up by Jafari et al. (2024). The model operates on Python 3.8 and necessitates the scipy, numpy, pandas, math and matplotlib libraries.The model operates on Python 3.8 and requires the scipy, numpy, pandas, math, and matplotlib libraries. The preprint version of the scientific article related to the model can be found on bioRxiv: https://biorxiv.org/cgi/content/short/2024.06.28.600787v1

Model Components and Files

The model can be found within the folder DynRAS_model and includes the following files:

• Class_definition_and_solver.py: The main script to solve the system of differential equations describing the changes in the system over time.
• ChemODE_BIO.py, ChemODE_DGS.py, ChemODE_Fish.py: Contain the system of differential equations for the biofilter, degasser, and fish tank, respectively. ChemODE_BIO.py can be modified to simulate different dosing scenarios.
• Fish_growth.py and Growth_Bacteria.py: Include the differential equations (ODE) for the growth of fish and bacteria (AOB and NOB).
• TAN_prod.py: Contains a function to simulate Total Ammonia Nitrogen (TAN) excretion from fish.
• Biomass_function.py: Simulates fish removal using a threshold for the maximum biomass present in the system. This function is called every 14 days.
• params.py: Lists the parameters for the model.

Running the Model

Default Simulation

To simulate the experimental set-up from Jafari et al. (2024), simply run the Class_definition_and_solver.py file. The default scenario is set to match the experimental conditions.

Scenario 1

1. For Alkalinity 200:

   • In Class_definition_and_solver.py, comment out lines 94, 96, 97, 98; uncomment line 95, dY = chemODE_BIO_HCO3(self, params, t), uncomment out lines 168, 187, 200, comment out lines 169, 188, 201.
   • In ChemODE_BIO.py, comment out line 107 and uncomment line 106.
   • Run Class_definition_and_solver.py.

2. For Alkalinity 70:

   • In ChemODE_BIO.py, uncomment line 107 and comment line 106.
   • In Class_definition_and_solver.py, comment out lines 168, 187, 200, uncomment lines 169, 188, 201.
   • Run Class_definition_and_solver.py.

Scenario 2

1. For Alkalinity 200:

   • In Class_definition_and_solver.py, comment out lines 94, 95, 97, 98,169, 188, 201; uncomment line 96 dY = chemODE_BIO_NaOH(self, params, t) and lines 168, 187, 200.
   • In ChemODE_BIO.py, uncomment line 136 and comment line 137.
   • Run Class_definition_and_solver.py.

2. For Alkalinity 70:

   • In ChemODE_BIO.py, uncomment line 137 and comment line 136.
   • In Class_definition_and_solver.py, comment out lines 168, 187, 200; uncomment lines 169, 188, 201.
   • Run Class_definition_and_solver.py.

Scenario 3

1. For Alkalinity 200:

   • In Class_definition_and_solver.py, comment out lines 94, 95, 96, 98,169, 188, 201; uncomment line 97 dY = chemODE_BIO_alk_control(self, params, t) and lines 168, 187, 200.
   • In ChemODE_BIO.py, uncomment line 168 and comment line 169.
   • Run Class_definition_and_solver.py.

2. For Alkalinity 70:

   • In ChemODE_BIO.py, comment line 168 and uncomment line 169.
   • In Class_definition_and_solver.py, comment out lines 168, 187, 200; uncomment lines 169, 188, 201.
   • Run Class_definition_and_solver.py.

pH Control Instead of Alkalinity Control

• In Class_definition_and_solver.py, uncomment line 98 and comment lines 94, 95, 96, 97 to switch to pH control.
• In ChemODE_BIO.py, to control pH with $HCO_3^-$, uncomment lines 204 and 211, and comment lines 212 and 206.
• To control pH with NaOH, uncomment lines 212 and 206, and comment lines 204 and 211.
• Adjust the pH threshold on line 202 if necessary.
• Initial conditions in Class_definition_and_solver.py might need adaptation.

When running the Class_definition_and_solver.py script, it generates two datafram that will be saved in two csv file 'results.csv' and 'dosing.csv'. The 'results.csv' file contains the simulation results for each species in each compartment of the system, first column is the time ('Time') and the name of the other columns can be found from line 267 to line 271. The name of the columns is build with the name of the species first, followed by the compartment '_FT' for the fish tank, '_B1' for the biofilter and '_DGS' for the Degasser. (e.g. 'CO2aq_FT' is the $CO_2$ result in the fish tank). Time is in second, the chemical species concentration are in $mmol.L^{-1}$ you can multiply the species concentration by their respective molecular weight given from line 274 to line 279 to get the concentration in $mg.l^{-1}$. The 'dosing.csv' contains 3 columns with the first column containing the simulation time, second column the dosing rate of OH and the dosing rate of HCO3 in $mmol.l^{-1}.s^{-1}$. To prevent overwriting this file with each run, change the filenames on line 281 and 314 of the script. This will ensure that each simulation's results are saved to a new file. In addition, from line 288 you have exemple on how to plot the results from the model. The simulation will automatically plot results for $CO_2$, alkalinity and TAN in mg per litre from the fish tank as well as the dosing rate of the two supplements over the simulation time.

Numerical issues

While fully functional, the model still needs improvement. Notably, certain numerical issues may still arise in specific scenarios, particularly when controlling the pH through $HCO_3^-$ . In these cases, while the system can still be integrated numerically, the solving time may increase. Adjustments such as increasing the maximum dosing rate by increasing the $HCO_3^-$ concentration (M) or the flow rate (Q) will fix the time issue. Additionally, it is important to account for chemical equilibria when setting initial conditions. For example, setting a very high pH, a very high $CO_2$ and a very low $HCO_3^-$ can lead to numerical instabilities.

Generating Figures

Data

To generate the same figures as in the paper, use either the output from the model simulations or the data available in the Code_figure directory. This directory contains:

• Experimental_data: Experimental data collected during the trial.
• Simulation_result: Simulation outputs for each scenario presented in the paper.

Creating Figures

The scripts used to generate the figures are in the Code_figure directory:

• Jafari_et_al_2024.py: Generates figures for model validation.
• Figure_scenario_1_2.py: Generates figures for scenarios 1 and 2.
• Scenario_3.py: Generates figures for scenario 3.

Running each file will create the figures and store them in a directory named figure, which contains subfolders for the figures corresponding to each script.

License

This project is licensed under the CC-BY-NC-ND 4.0 license.