Model description:
Coupled nonlinear differential equations describing a process involving a continuous flow stirred tank reactor are given by
$$\begin{align*} \dot{C}_1 &= -C_1u + C_1(1-C_2)e^{C_2/\Gamma} \\ \dot{C}_2 &= -C_2u + C_1(1-C_2)e^{C_2/\Gamma}\dfrac{1+\beta}{1+\beta-C_2}. \end{align*}$$
In these equations, the state variables $C_1$ and $C_2$ represent dimensionless forms of cell mass and amount of nutrients in a constant volume tank, bounded between zero and unity. The control $u$ is the flow rate of nutrients into the tank, and is the same rate at which contents are removed from the tank. The constant parameters $\Gamma$ and $\beta$ determine the rates of cell formation and nutrient consumption; these parameters are set to $\Gamma$= 0.48 and $\beta$ = 0.02 for the nominal benchmark specification.
Type:
Form:
Model order:
Time domain:
Linearity:
Publication details:
Title | Neurocontrol of nonlinear dynamical systems with Kalman filter trained recurrent networks |
Publication Type | Journal Article |
Year of Publication | 1994 |
Authors | Puskorius, G.V., and Feldkamp L.A. |
Journal | IEEE Transactions on Neural Networks |
Volume | 5 |
Issue | 2 |
Start Page | 279 |
Pagination | 279-297 |
Date Published | 1994 |
ISSN | 1045-9227 |
Accession Number | 4685633 |
Keywords | filtering and prediction theory, Kalman filters, nonlinear control systems, Nonlinear dynamical systems, recurrent neural nets |
Abstract | Although the potential of the powerful mapping and representational capabilities of recurrent network architectures is generally recognized by the neural network research community, recurrent neural networks have not been widely used for the control of nonlinear dynamical systems, possibly due to the relative ineffectiveness of simple gradient descent training algorithms. Developments in the use of parameter-based extended Kalman filter algorithms for training recurrent networks may provide a mechanism by which these architectures will prove to be of practical value. This paper presents a decoupled extended Kalman filter (DEKF) algorithm for training of recurrent networks with special emphasis on application to control problems. We demonstrate in simulation the application of the DEKF algorithm to a series of example control problems ranging from the well-known cart-pole and bioreactor benchmark problems to an automotive subsystem, engine idle speed control. These simulations suggest that recurrent controller networks trained by Kalman filter methods can combine the traditional features of state-space controllers and observers in a homogeneous architecture for nonlinear dynamical systems, while simultaneously exhibiting less sensitivity than do purely feedforward controller networks to changes in plant parameters and measurement noise. |
DOI | 10.1109/72.279191 |