An integration scheme for stiff solid–gas reactor models

Self Cite

The paper addresses e¢ cient methods for parameter sensitivity analysis and ranking in large, nonlinear, mechanistic models requiring examination of many points in the parameter space. The paper shows how orthogonal decomposition and permutation of the sensitivity derivative is an intuitive and structured method for automatic ranking of the parameters within a candidate set. Provided the model error is Gaussian, and with the problem on a triangularized form, the additional variance associated with each parameter can easily be found. Ranking according to additional variance is therefore another option. The methods are tested on an industrially used simulator model.

Section: Examplementioning

confidence: 99%

Methods for Parameter Ranking in Nonlinear, Mechanistic Models

Lund¹,

Berntsen²,

Foss³

2005

Self Cite

“…The reactor is modeled by introducing two major reaction zones. The upper zone is treated as a counter current reactor while the lower zone is treated as a stirred tank (Halvorsen et al, 1991;Halvorsen, 1992;Schei et al, 1998) and (Foss and Wasbø, 2001). The original PDE model represents a hyperbolic PDE system and proved to be very difficult to solve numerically.…”

Section: Introductionmentioning

confidence: 99%

Passivity Based Control and Optimization of a Silicon Process

Ruszkowski

Read

Kaiser

et al. 2005

Abstract:In this paper we develop a Real Time Optimization (RTO) strategy for a silicon reactor. The reactor dynamics are described using parabolic partial differential equations with countercurrent liquid, gas and solid flow with chemical reaction. The model equations are converted to a system of differential algebraic equations using the method of lines and up-winding approximation for the spatial derivatives. The model parameters are matched to industrial data by minimizing a quadratic objective function using a novel interior point SQP method. The dynamic response has been verified and we have shown that passivity based inventory control can be used to stabilize the process at its optimal operating point

“…Simod is especially detailed with respect reaction kinetics and the thermodynamics of the furnace. The model is a nonlinear di¤erential algebraic equation system, see (Foss and Wasbø, 2001). Simod was originally developed for teaching and training purposes of plant operation personnel in Elkem.…”

Section: Introductionmentioning

confidence: 99%

System Analysis of Complex Reactor behavior - A Case Study

Lund

Foss

Løvåsen³

et al. 2004

Self Cite

The paper examines the nonlinear behaviour between carbon coverage (input) and silicon production (output) in a submerged arc silicon furnace. A silicon furnace is a highly endothermic chemical reactor. Carbon coverage is the main input for control of the silicon yield and is determined manually. This is a di¢ cult task since the dynamic response varies with the carbon level in the furnace. Internal interconnections between the reactions in the process cause the dominant zeros and poles in the response to move. For high input values and high silicon yield, inverse responses and …nally unstability result. The paper describes how the dominant poles and zeros move and identi…es the internal interconnections in the reactor that cause the change in dynamics.Keywords: Reactor dynamic behaviour, zero and pole drift. Submerged arc silicon furnace.