Search

In this work, we propose a sensitivity-based approach to construct reduced-order state estimators based on recurrent neural networks (RNN). It is assumed that a mechanistic model is available but is too computationally complex for estimator design and that only some target outputs are of interest and should be estimated. A reduced-order estimator that can estimate the target outputs is sufficient to address such a problem. We introduce an approach based on sensitivity analysis to determine how to select the appropriate inputs and outputs for data collection and data-driven model development to estimate the desired outputs accurately. Specifically, we consider the long short-term memory (LSTM) neural network, a type of RNN, as the tool to train the data-driven model. Based on it, an extended Kalman filter, a state estimator, is designed to estimate the target outputs. Simulations are carried out to illustrate the effectiveness and applicability of the proposed approach.

The Richards equation plays an important role in the study of agro-hydrological systems. It models the water movement in soil in the vadose zone, which is driven by capillary and gravitational forces. Its states (capillary potential) and parameters (hydraulic conductivity, saturated and residual soil moistures and van Genuchten-Mualem parameters) are essential for the accuracy of mathematical modeling, yet difficult to obtain experimentally. In this work, an estimation approach is developed to estimate the parameters and states of Richards equation simultaneously. In the proposed approach, parameter identifiability and sensitivity analysis are used to determine the most important parameters for estimation purpose. Three common estimation schemes (extended Kalman filter, ensemble Kalman filter and moving horizon estimation) are investigated. The estimation performance is compared and analyzed based on extensive simulations.