Chang Kyoo Yoo

Nonlinear PLS modeling with fuzzy inference system

Abstract We propose a new nonlinear partial least squares (NLPLS) algorithm that embeds the Takagi–Sugeno–Kang (TSK) fuzzy model into the regression framework of the partial least squares (PLS) method. We call the new algorithm fuzzy partial least squares (FPLS). Several NLPLS algorithms have been proposed. However, they can lead to overfitting and contain ambiguities in the meaning of regression parameters. The proposed FPLS algorithm applies the TSK fuzzy model to the PLS inner regression. Using this approach, the interpretability of the TSK fuzzy model overcomes some of the handicaps of previous NLPLS algorithms. The proposed method uses the PLS method to solve the problems of high dimensionality and collinearity and the TSK fuzzy model is used to capture the nonlinearity and to increase the use of experts knowledge. As a result, the FPLS model gives a more favorable modeling environment in which the knowledge of experts can be easily applied. In addition, we propose a new input and output weight update algorithm to enhance the regression performance of FPLS. The power of the proposed method is illustrated by application to a simple mathematical simulation data set and a real near infrared spectral data set.

Process system engineering in wastewater treatment process

This paper reviews the research and development of process system engineering (PSE) in the wastewater treatment process (WWTP). A diverse range of PSE applications have evolved in the wastewater treatment process, such as modeling, control, estimation, expert system, fault detection and monitoring system. This article describes several types of PSE that have proven to be effective in WWTP. The merits and shortcoming of PSE and its detailed applications are presented. Since its development is the forefront in WWTP, a reasonable review of the research progress in this field is addressed.

Generalized damped least squares algorithm

We propose a new algorithm for adaptive control and self tuning control, referred to as the generalized damped least squares (GDLS) algorithm. This algorithm is constructed by adding a multi-step penalty for parameter variations to the objective function of the normal least squares algorithm to prevent the singularity problem that leads to estimation windup. We show that the proposed method has properties almost equivalent to those of the normal least squares method, which guarantees that the proposed algorithm is suitable for poorly excited situations. Simulation results show that the proposed method gives better estimation performance than previous methods in spite of its simplicity. The proposed method also shows good parameter tracking performance and no estimation windup.

COMPARISONS OF PROCESS IDENTIFICATION METHODS AND SUPERVISORY DO CONTROL IN THE FULLSCALE WASTEWATER TREATMENT PLANT

Abstract Dissolved oxygen (DO) control in the wastewater treatment plant (WWTP) has been the most critical factor in terms of improving the wastewater treatment efficiency and energy saving. However, the result of DO control with simple PID controller is not satisfactory since the dynamics of DO show some time-varying and nonlinear characteristics. The objectives of this study are as follows. First aim is to apply and compare some process identification methods of PID autotuning for stable DO control in a real coke WWTP. Second, a simple algorithm for the supervisory control of set point decision is proposed to decide a proper DO set point for the current operation condition of the aeration basin by estimating the respiration rate simultaneously during process identification step. The key idea in this method is that the DO set point is proportional to the respiration rate, which is the indicator of the biologically degradable load. In the experimental results, process identification methods have been practiced and compared at the modeling and control performance and supervisory control reduced the aeration cost in the full-scale WWTP.

Adaptive Modeling and Classification of the Secondary Settling Tank

In biological wastewater treatment plants the biomass is separated from the treated wastewater in the secondary settler; thus, efficient operation of the secondary settler is crucial to achieving satisfactory effluent quality in the wastewater treatment process (WWTP). In the present work, system identification and soft-computing techniques were used to formulate a model for predicting the solid volume index (SVI) and classification of the sludge bulking phenomenon in the settler. An adaptive time series model was applied to predict the SVI of the secondary settler; this model uses the recursive least square (RLS) method to update the model parameters. The method for classifying the current state of the secondary settler is based on the strong correlation that was observed between the settler state and the values of the time series model parameters, which enabled the time series model parameters to be used as effective features for monitoring the secondary settler. To classify the current state of the secondary settler, a neural network (NN) was used to classify the adaptive time series model parameters, where a hybrid Genetic Algorithm (GA) was used to decide the number of hidden nodes of the NN classifier. Application of the proposed method to a full-scale WWTP demonstrated the utility of the method for simultaneously predicting the SVI value of the secondary settler and classifying the current state of the settler.

Direct Identification Method of Second Order Plus Time Delay Model Parameters

In this paper, we propose an identification method to estimate the second order plus time delay (SOPTD) model parameters. The proposed method directly obtains these parameters using a frequency-weighted integral transform and a least-squares method, without iterative calculation step. This calculation procedure can be applied regardless of initial states, in contrast to almost all previous identification methods, which require the assumption of zero initial states. In addition, it does not require any additional model reduction steps to tune the PID controller. Using simulations, it is demonstrated that the proposed method provides better modeling performance than previous methods, in spite of its simplicity. The new method shows an acceptable robustness to disturbance and measurement noise.

Sensitivity Analysis and Parameter Estimation of Activated Sludge Model Using Weighted Effluent Quality Index

Many modeling and calibration methods have been developed to analyze and design the biological wastewater treatment process. For the systematic use of activated sludge model (ASM) in a real treatment process, a most important step in this usage is a calibration which can find a key parameter set of ASM, which depends on the microorganism communities and the process conditions of the plants. In this paper, a standardized calibration protocol of the ASM model is developed. First, a weighted effluent quality index(WEQI) is suggested far a calibration protocol. Second, the most sensitive parameter set is determined by a sensitive analysis based on WEQI and then a parameter optimization method are used for a systematic calibration of key parameters. The proposed method is applied to a calibration problems of the single carbon removal process. The results of the sensitivity analysis and parameter estimation based on a WEQI shows a quite reasonable parameter set and precisely estimated parameters, which can improve the quality and the efficiency of the modeling and the prediction of ASM model. Moreover, it can be used for a calibration scheme of other biological processes, such as sequence batch reactor, anaerobic digestion process with a dedicated methodology.

Optimization of Water-Reusing Network among the Industries in an Eco-Industrial Park Complex Using Water Pinch Technology

An water-reuse network design has drawn attention as a systematic method of reducing fresh water usage and increasing water-using efficiency. The final goal of an eco-industrial park(EIP) is accomplishing industrial sustainability and constructing water-reuse network can be a solution. When designing water-reuse network connecting various processes which consume water, the water pinch technology can be used frequently, since it simultaneously minimize freshwater usage and wastewater discharge. In this research water pinch technology is applied to develop an effective water-reuse network in an EIP. Three scenarios based on different reusing strategies were developed. The results show that the final water-reuse network can reduce the total fresh water usage more than 30%, while the water expenses decrease by 20%. It can be concluded that water pinch technology is an effective tool to optimize water-reuse network among different industrial facilities.

Analysis and Subclass Classification of Microarray Gene Expression Data Using Computational Biology

Application of microarray technologies which monitor simultaneously the expression pattern of thousands of individual genes in different biological systems results in a tremendous increase of the amount of available gene expression data and have provided new insights into gene expression during drug development, within disease processes, and across species. There is a great need of data mining methods allowing straightforward interpretation, visualization and analysis of the relevant information contained in gene expression profiles. Specially, classifying biological samples into known classes or phenotypes is an important practical application for microarray gene expression profiles. Gene expression profiles obtained from tissue samples of patients thus allowcancer classification. In this research, molecular classification of microarray gene expression data is applied for multi-class cancer using computational biology such gene selection, principal component analysis and fuzzy clustering. The proposed method was applied to microarray data from leukemia patients; specifically, it was used to interpret the gene expression pattern and analyze the leukemia subtype whose expression profiles correlated with four cases of acute leukemia gene expression. A basic understanding of the microarray data analysis is also introduced.

Generic Detection and Isolation of Process Events in WWTP

Abstracts In this paper, a new monitoring approach utilizing a change in distribution of process data is presented since the distribution reflects the corresponding operating condition. In order to quantitatively evaluate the difference between two data sets, a generic dissimilarity measure (GDM) is used. It considers the importance of each transformed variable. As an application, it is used to monitor the real biological wastewater treatment process (WWTP) of the iron and steel making plant in Korea. The application results showed that the proposed algorithm has an efficiency of detecting process disturbance. Additionally, it approximately discriminated between serious and minor process change. That is, it can not only detect various disturbances, but isolate them to a degree.