Groundwater

Oral

397268 - Estimating Soil Hydraulic Parameters using Gradient Based Approach

Wednesday, June 6
10:30 AM - 12:00 PM
Location: Lakeshore B
Co-Authors: Shivam Tripathi, Kanpur – IIT Kanpur

The conventional way of estimating parameters of a differential equation is to minimize the error between the observations and their estimates. The estimates are produced from forward solution (numerical or analytical) of differential equation assuming a set of parameters. Parameter estimation using the conventional approach requires high computational cost, setting-up of initial and boundary conditions, and formation of difference equations in case the forward solution is obtained numerically. Gaussian process based approaches like Gaussian Process Ordinary Differential Equation (GPODE) and Adaptive Gradient Matching (AGM) have been developed to estimate the parameters of Ordinary Differential Equations without explicitly solving them. Claims have been made that these approaches can straightforwardly be extended to Partial Differential Equations; however, it has been never demonstrated. This study extends AGM approach to PDEs and applies it for estimating parameters of Richards equation. Unlike the conventional approach, the AGM approach does not require setting-up of initial and boundary conditions explicitly, which is often difficult in real world application of Richards equation. The developed methodology was applied to synthetic soil moisture data. It was seen that the proposed methodology can estimate the soil hydraulic parameters correctly and can be a potential alternative to the conventional method.

Pankaj Kumar Rai


IIT Kanpur

Presentation(s):

Send Email for Pankaj Rai


Assets

397268 - Estimating Soil Hydraulic Parameters using Gradient Based Approach



Attendees who have favorited this

Please enter your access key

The asset you are trying to access is locked. Please enter your access key to unlock.

Send Email for Estimating Soil Hydraulic Parameters using Gradient Based Approach