Ch. Venkateswarlu

Optimal state estimation of multicomponent batch distillation

Various methods based on extended Kalman filter, adaptive fading Kalman filter and steady state Kalman filter have been presented for inferential estimation of compositions from temperature measurements in multiple fraction multicomponent batch distillation. This dynamic model based state estimators incorporate component balance equations together with thermodynamic relations, that include bubble point temperature computation and avoid constant relative volatility concept. A performance criterion with multiple performance indices is used to assess the composition estimators. The sensitivity of the estimators is studied with respect to the effect of number of measurements, measurement noise, and filter design parameters including initial state covariance, process and observation noise covariance matrices. The results of the state estimation methods are evaluated by applying them for composition estimation on all trays, reboiler, condenser and products of a ternary hydrocarbon system. The results show the better suitability of the method of extended Kalman filter for composition estimation in multicomponent batch distillation.

Brief Communication: Classification and identification of mosquito species using artificial neural networks

An artificial neural network method is presented for classification and identification of Anopheles mosquito species based on the internal transcribed spacer2 (ITS2) data of ribosomal DNA string. The method is implemented in two different multi-layered feed-forward neural network model forms, namely, multi-input single-output neural network (MISONN) and multi-input multi-output neural network (MIMONN). A number of data sequences in varying sizes of different Anopheline malarial vectors and their corresponding species coding are employed to develop the neural network models. The classification efficiency of the network models for untrained data sequences is evaluated in terms of quantitative performance criteria. The results demonstrate the efficiency of the neural network models to extract the genetic information in ITS2 sequences and to adapt to new data. The method of MISONN is found to exhibit superior performance over MIMONN in distinguishing and identification of the mosquito vectors.

SOFT SENSOR BASED NONLINEAR CONTROL OF A CHAOTIC REACTOR

Abstract Abstract Control of nonlinear systems exhibiting complex dynamic behavior is a challenging task because such systems present a variety of behavioral patterns depending on the values of physical parameters and intrinsic features. Understanding the behavior of the nonlinear dynamic systems and controlling them at the desired conditions is important to enhance their performance. In this work, a soft sensor based nonlinear controller strategy is presented and applied to control a chemical reactor that exhibit multi-stationary unstable behavior, oscillations and chaos. In this strategy, an extended kalman filter is designed to serve as a soft sensor that provides the estimates of unmeasured process states. These states are used as inferential measurements to the nonlinear controller that is designed in the framework of globally linearizing control. The results evaluated for stabilizing the reactor for different conditions including deterministic and stochastic disturbances show the better performance of the soft sensor based nonlinear control strategy over that of a PID controller with modified feedback mechanism.

Estimating biofilm reaction kinetics using hybrid mechanistic-neural network rate function model.

This work describes an alternative method for estimation of reaction rate of a biofilm process without using a model equation. A first principles model of the biofilm process is integrated with artificial neural networks to derive a hybrid mechanistic-neural network rate function model (HMNNRFM), and this combined model structure is used to estimate the complex kinetics of the biofilm process as a consequence of the validation of its steady state solution. The performance of the proposed methodology is studied with the aid of the experimental data of an anaerobic fixed bed biofilm reactor. The statistical significance of the method is also analyzed by means of the coefficient of determination (R2) and model efficiency (ME). The results demonstrate the effectiveness of HMNNRFM for estimating the complex kinetics of the biofilm process involved in the treatment of industry wastewater.

A HIERARCHICAL ARTIFICIAL NEURAL SYSTEM FOR GENERA CLASSIFICATION AND SPECIES IDENTIFICATION IN MOSQUITOES

A hierarchical artificial neural system (HANS) is proposed for genera classification and species identification in mosquitoes, based on the internal transcribed spacer 2 (ITS2) data of ribosomal DNA sequence. The HANS is composed of two levels: the first level has a single network that serves as a genera classifier and the second level has multiple networks to perform as species identifiers. The proposed method is studied by considering a number of data sequences of various species of Aedes, Anopheles, and Culex genera. The significant feature of HANS is that when an unknown sequence is presented to the system, it first classifies the genera to which the sequence corresponds and then proceeds with the same sequence information to identify the species of the respective genera. Two more methodologies, a combined artificial neural system (CANS) and a separate artificial neural system (SANS), are also presented and compared with HANS. The results demonstrate the superior performance of HANS for genera classification and species identification in mosquitoes.

Two-level methods for incipient fault diagnosis in nonlinear chemical processes

Abstract Various two-level methods are formulated by using different forms of extended Kalman filter, recursive least squares and a reduced-order Luenberger observer for fault detection and diagnosis in nonlinear, time-varying and stochastic chemical processes. These methods are specified for state estimation in the first level and fault diagnosis via parameter identification in the second level. The performances of the proposed methods are evaluated by applying them to a nonlinear CSTR with a heat exchanger and a nonlinear batch beer fermentation with closed-loop control. Dynamic simulation results demonstrate that these two-level methods can be applied for process fault detection and diagnosis to transient time domain systems with closed-loop control in addition to the steady-state systems of normal operation. Among these methods, the two-level extended Kalman filter has exhibited better performance. Therefore, the method of the two-level extended Kalman filter is recommended for incipient fault diagnosis.

Dynamic Modeling and Metabolic Flux Analysis for Optimized Production of Rhamnolipids

A comprehensive metabolic network based on the fundamental pathways representing the central metabolism of rhamnolipid by Pseudomonas aeruginosa is proposed and a dynamic model compatible with the underlying metabolic network is developed involving the macro-reactions derived from the elementary flux modes of the reaction network. The experimentally validated mathematical model is then coupled with a global optimization technique called differential evolution (DE) to optimize the medium composition as well as the extracellular and intracellular fluxes of the metabolic network. The analysis of the results shows the usefulness of the integrated approach involving the development of a dynamic model based on the metabolic network structure and model-based optimization of the medium composition and metabolic fluxes by an efficient evolutionary optimization technique to enhance the productivity of rhamnolipid.

ADAPTIVE FUZZY MODEL BASED PREDICTIVE CONTROL OF AN EXOTHERMIC BATCH CHEMICAL REACTOR

An adaptive fuzzy model based predictive control (AFMBPC) approach is presented to track the desired temperature trajectories in an exothermic batch chemical reactor. The AFMBPC incorporates an adaptive fuzzy modeling framework into a model based predictive control scheme to derive analytical controller output. This approach has the flexibility to cope with different fuzzy model structures whose choice also lead to improve the controller performance. In this approach, adaptation of fuzzy models using dynamic process information is carried out to build a predictive controller, thus eliminating the determination of a predefined fixed fuzzy model based on various sets of known input-output relations. The performance of the AFMBPC is evaluated by comparing to a fixed fuzzy model based predictive controller (FFMBPC) and a conventional PID controller. The results show the better suitability of AFMBPC for the control of highly nonlinear and time varying batch chemical reactors.

DYNAMIC FUZZY ADAPTIVE CONTROLLER FOR pH

Control of pH processes is very difficult due to nonlinear dynamics, high sensitivity at the neutral point, and changes in the concentrations of known or unknown chemical species. In this study, a dynamic fuzzy adaptive controller (DFAC) with a new inference mechanism is proposed and applied for the control of pH processes. The DFAC consists of a low-level basic control phase with a minimum rule base and a high-level dynamic learining phase with an updating mechanism to interact and modify the control rule base. The DFAC can self-adjust its fuzzy control rules using information from the process during on-line control and create new fuzzy control rules or modify the present control rules using its learning capability from past control trends. The controller is evaluated by applying it to a weak acid-strong base pH process with input disturbances and to another pH process that involve that has changes in acidic/buffering streams. The results of the DFAC with the new inference mechanism are compared with the known inference mechanisms, the fuzzy controller, the conventional PI controller, and also with an adaptive PID controller. The proposed DFAC provides better performance for set point tracking of the pH and rejection of load disturbances and buffering affects.

Identification of Better Gene Expression Data for Mosquito Species Classification Using Radial Basis Function Network Methodology

Investigation in bioinformatics has developed promptly in latest years owing to improvements in sequence excavating techniques. Gene sequences in DNA are supplemented with great extent of information, but the intricacy and complexity of this information causes difficulty in analyzing it by using standard classical methods of classification. In this work, a Radial Basis Function Network (RBFN) methodology with self-network arrangement is presented for identification of mosquito species based on the genetic design content of ITS2 ribosomal DNA sequences.