-
Monte Carlo simulation: By using multiple random sampling and simulation, various possibilities of groundwater flow can be obtained and the uncertainty of groundwater flow can be quantified to improve the accuracy and reliability of model prediction.
-
Simulation method:
- Define the Probability Density Function (PDF) of the parameters, which can usually be established based on empirical formulas.
- Randomly sample the defined PDF to generate realizations and obtain the corresponding groundwater flow field.
- Repeat steps 1 and 2.
- For the n simulation results obtained in step 3, conduct statistical analysis such as calculating the mean, standard deviation, percentiles, etc., to obtain a statistical distribution that represents the simulation results.
- Based on the obtained statistical distribution, conduct risk assessment or other related applications, such as calculating the probability distribution of a certain water level or flux, or calculating the worst-case scenario under different water level or flux conditions.
-
Tools used:
- Numerical groundwater simulation tool: FEFLOW 7.3
- Scripting language: Python 3.8.6
-
Use the hydraulic conductivity (K) as the uncertain parameter in the experiment.
-
Define the PDF of K based on empirical formulas
-
Use the stats module in Python to generate n random samples of K.
-
Conduct Monte Carlo simulation:
- Set the sampled K values to the model using ifm.
- Each time the K value is set, conduct simulation and record the velocity.
- Repeat the above n times.
-
Visualize the results of the velocity obtained from n simulation runs. (
matplotlib,seaborn,plotly, ...)
