Identification Of Kinetic Models Research Articles

Identification of kinetic models of heterogeneously catalyzed reactions

Three case studies present different aspects of the identification of kinetic models in heterogeneous catalysis. The demonstrations are based on the real examples of laboratory catalytic hydrogenation experiments. The common target of all the studies is to demonstrate the possibilities of maximizing the reliability of parameter estimates for the limited experimental sets of data. The topics of the case studies include the comparison of Langmuir–Hinshelwood (LH) and non-stationary kinetic models, a discussion concerning the effects of invalid assumptions for catalyst surface coverage, and issues related to the design of experiments.

Testing novel short-cut methods for complex kinetic characterisation in early safety assessment of a chemical process

One of the hazards of most concern in the chemical industry is the thermal hazard, which is governed by the thermodynamics and kinetics of the usually complex chemical reactions. To develop risk analysis indices the approximate knowledge of the reaction path, kinetics and thermal effects are crucial elements. While standardised tests to experimentally determine the ‘thermal stability data’ and assessment procedures for runaway reactions are developed, their interpretation is difficult to be generalised because they have been developed in the context of a specific process, and they often contain empirical as well as rationally deduced elements and rules. This paper illustrates major advantages and result improvements in thermal risk analysis when two short-cut techniques are combined: the experimental ‘Differential Scanning Calorimetry’ ( DSC) and the numerical ‘Modified Integral Transformation Procedure’ ( MIP). The DSC is useful in investigating the n-th order kinetics and overall thermal effect of a reaction following an elaborated strategy; attempts to fit more complex kinetics fail especially in the model discrimination step. The MIP ( Maria and Rippin, 1997) performs a short-cut kinetic model identification by considering previous information stored in databanks and incomplete information about the process. By coupling the DSC and MIP techniques quick identification of a (reasonable) complex chemical kinetics becomes possible. Exemplification is made for an auto-catalytic process with incomplete data by obtaining: i) a more exact evaluation of the safety index TMR ad compared with those derived from classical DSC; ii) a more robust estimate of the kinetics by using the databank information in overcoming poor-conditioning due to incomplete data.

Expert system for ode chemical kinetic model identification by using a transfer of information numerical strategy

Recently developed numerical data analysis techniques could be worthy instruments to check/validate chemical reaction paths and to build/reduce (Nonlinear) Kinetic Models ((N)KM). The developed Expert System (ES) includes a number of numerical procedures (concentration trajectory analysis, transfer of information numerical strategy for initial guess generation, Concomitant Parameter Estimation and Model Reduction (CPEMR)) interconnected to ensure high reliability in identifying the more suitable Ordinary Differential Equations (ODE) - NKM and in finding the global Parameter Estimate (PE) solution.

Expert system for ODE chemical kinetic model identification by using a transfer of information numerical strategy

Expert system for ODE chemical kinetic model identification by using a transfer of information numerical strategy

The identification of kinetic expressions and the evolutionary optimization of specialty chemical batch reactors using tendency models

This paper presents a method for the identification of kinetic models for a wide variety of batch/semibatch processes related to the production of fine and specialty chemicals. The developed methodology is particulary suitable for those industrial processes where a detailed understanding of process fundamentals (reaction mechanisms, process kinetics) is not available. It is also readily applicable to those processes where it is difficult to obtain time dependent concentration data.The modeling effort is aimed at developing a low-order, nonlinear “Tendency Model” which is descriptive of the qualitative and approximate quantitative behavior of the overall process. the physical/chemical insight gained from a set of two-level multifactor experiments in primary operating variables (e.g. temperature, catalyst concentration, initial concentrations of key reactants, promotors, etc) is used to select an approximate functional form (viz. power-law kinetics, Langmiur-Hinshelwood, etc.) for various rate equations in the kinetic model. A priori process knowledge and process understanding gained from statistical analysis of the data collected is used to aid in the model identification. The final values of reaction orders and the remaining kinetic parameters are determined by minimizing a prespecified model-fitting function.A factorial analysis of the model's prediction for process responses is used to determine the accuracy of the model. An inaccurate model is modified to account for any observed discrepencies. The developed “Tendency Model” is used for the optimization of the process in an evolutionary manner. As the process operates and additional data becomes available, the model parameters are updated using the information on the sensitivity of the plant-model mismatch with respect to various reaction orders. The developed methodology is illustrated by means of experimental data for an example process related to the production of fatty acid epoxides which are used as stabilizer/plasticizers for PVC resins.

Kinetic model identification in mixed populations using continuous culture data

AbstractThe kinetic behavior of heterogeneous microbial populations of sewage origin was studied in a single‐stage, isothermal, continuous flow, completely mixed aeration tank. A series of experiments were carried out at various dilutions rates using glucose as the limiting substrate. The cell dry weight and substrate concentration in terms of chemical oxygen demand (COD) were continuously monitored. The results indicate that reproducible steady‐state conditions can generally be obtained; however, multiple steady states were observed at dilution rates near washout. At low dilution rates (below about 0.1 hr−1) the contribution of microorganism decay became appreciable. Using the multiresponse data of cell dry weight and COD, the parameter values in various existing growth models were estimated. The analyses of variance and residuals revealed that models proposed by Moser, Monod, and Contois, each with a decay term added, were significantly better than the other models which were tested.