Recursive estimation in constrained nonlinear dynamical systems

Abstract

In any modern chemical plant or refinery, process operation and the quality of product depend on the reliability of data used for

Model based predictive control for energy efficient biological nitrification process with minimal nitrous oxide production

“…However for nonlinear systems, EKF is a widely used de facto approach which linearizes the nonlinear system about current mean and covariance by using a Taylor series and then estimates the unmeasured states by a prediction and correction approach [24,25]. This technique does not guarantee optimal estimation as there is loss of information during the linearization process [26,27] and performs poorly when presented with constraints [28]. Nevertheless, EKF is known to provide convergent results for varying degrees of parameter changes in the model and also known as de facto filter [27,29,25].…”

In‐Line Monitoring of Bulk Polypropylene Reactors Based on Data Reconciliation Procedures

“…For instance, EKF cannot handle inequality or equality constraints efficiently and can therefore give rise to infeasible estimates. Besides, EKF needs a linear or locally linearized model, which may lead to accumulation of modeling errors and bias (Vachhani et al, 2005) [8] . Finally, EKF algorithms are difficult to tune and can give poor estimates of unmeasured parameters and disturbance variables.…”

A recursive state estimator in the presence of state inequality constraints

“…Several researchers have worked on the design of state estimations when the states of the system are constrained either by equality constraints or inequality constraints or both [1][2][3][4][5][6][8][9][10][11][12][13][14][15][16]19,[21][22][23][24][25][26][27][28][29][30][31]. For example, Wen and Durrant-Whyte [30] reduced the model of the system by substituting the equality constraints into the model of the system.…”