Shafishuhaza Sahlan

A generalised partial-fraction-expansion based frequency weighted balanced truncation technique

In this paper, we present some new results on a frequency weighted balanced truncation technique based on well-known partial-fraction-expansion idea. The reduced order models which are guaranteed to be stable in case of double-sided weighting are obtained by direct truncation. Two sets of simple, elegant and easily computable a priori error bounds are also derived. Relationships between the proposed method and the previous methods based on partial-fraction idea are also derived. The technique is illustrated using a numerical example of a practical application and then compared with other well-known techniques, to show the effectiveness of the method.

Micromachined Shape-Memory-Alloy Microactuators and Their Application in Biomedical Devices

Shape memory alloys (SMAs) are a class of smart materials characterized by shape memory effect and pseudo-elastic behavior. They have the capability to retain their original form when subjected to certain stimuli, such as heat or a magnetic field. These unique properties have attracted many researchers to seek their application in various fields including transportation, aerospace, and biomedical. The ease process adaption from semiconductor manufacturing technology provides many opportunities for designing micro-scale devices using this material. This paper gives an overview of the fabrication and manufacturing technique of thin-film and bulk micromachined SMAs. Key features such as material properties, transformation temperature, material composition, and actuation method are also presented. The application and micromechanism for both thin-film and bulk SMA are described. Finally, the microactuator devices emphasized for biomedical applications such as microgrippers and micropumps are highlighted. The presented review will provide information for researchers who are actively working on the development of SMA-based microscale biomedical devices.

Fouling control strategy for submerged membrane bioreactor filtration processes using aeration airflow, backwash, and relaxation: a review

AbstractAs of today, the application of membrane bioreactors (MBRs) in wastewater treatment plants has become significantly important in various industries. The success of MBR is mainly because of the effective membrane technology used in this kind of reactor. However, membrane technology does not escape from several drawbacks, which includes high maintenance costs and fouling problems. In order to run a plant with both an optimum capability and cost, an in-depth understanding of the behavior of filtration systems is very important to the plant operator. In this paper, we review and have considered the fouling control techniques applied to submerged membrane filtration processes including hollow fiber membranes and flat sheet types of membrane filtration. The review covers the techniques involving operational parameters, such as aeration, backwashing, and relaxation. In addition, in this paper, the modeling techniques and available automatic control techniques, using stated parameters to control fouling i...

A new method for the model reduction technique via a limited frequency interval impulse response Gramian

A newly proposed method of frequency-weighted model reduction for single-input–single-output (SISO) linear continuous-time systems is proposed. It is based on matching the elements of the limited frequency interval impulse response Gramian (IRG). A numerical example is given to illustrate the method, which is then compared with other well-known model reduction methods.

Particle Swarm Optimization for multivariable PID controller tuning

In this paper, the tuning of multivariable PID (MPID) controller using particle swarm optimisation (PSO) is presented. To provide better performance in the industry, MPID implementation often uses heuristic approach, which is based on experience. However, this method of tuning is time consuming and inadequate for complex systems. To obtain a better performance in decoupling capabilities and closed loop performance, in this study, the MPID parameters are fine-tuned using Multi-objective PSO. Simulation results show tuning parameters of MPID controllers obtained via PSO yield better performance when applied to nonlinear system for different control strategies.

N4SID and MOESP subspace identification methods

Multivariable Output Error State Space (MOESP) and Numerical algorithms for Subspace State Space System Identification (N4SID) algorithms are two well known subspace identification techniques discussed in this paper. Due to the use of robust numerical tools such as QR decomposition and singular value decomposition (SVD), these identification techniques are often implemented for multivariable systems. Subspace identification algorithms are attractive since the state space form is highly suitable to estimate, predict, filters as well as for control design. In literature, there are several simulation studies for MOESP and N4SID algorithms performed in offline and online mode. In this paper, order selection, validity and the stability for both algorithms for model identification of a glass tube manufacturing process system is considered. The weighting factor α, used in online identification is obtained from trial and error and particle swarm optimization (PSO). Utilizing PSO, the value of α is determined in the online identification and a more accurate result with lower computation time is obtained.

A frequency limited model reduction technique for linear discrete systems

This paper describes the model reduction framework for single-input single-output (SISO) discrete-time systems based on the preservation parameters such as Markov properties of the original system by applying a Frequency-Limited Impulse Response Gramian based Balanced Truncation method. This proposed method extends the Frequency-Limited Impulse Response Gramians model reduction method for continuous systems described in the recent paper in [20] to be applicable for discrete time systems. A numerical example is provided to compare the performances between various frequency limited model reduction methods at an arbitrarily selected frequency range within the passband of a digital filter. The stability of the reduced order models are also checked for each scenario.

Modeling of Waste Water Treatment Plant via system ID & model reduction technique

This paper investigates the application of Model Order Reduction (MOR) technique to Waste Water Treatment Plant (WWTP) system. The mathematical model of WWTP is obtained by using System Identification. In this paper, Prediction Error Estimate of Linear or Nonlinear Model (PEM) is proposed as the System Identification method which is used to find the parameter of linear or nonlinear system in state-space model from an experimental input-output data WWTP. The result shows that the estimated model of WWTP is a high order system with good best fit with 91.56% and 80.19% compared to the original experimental model. To simplify the obtained model,the MOR technique is proposed to reduce the high order system to lower order system while still retaining the characteristics of the original system. In this paper, the balanced truncation and Frequency Weighted Model Reduction (FWMR) are proposed to obtain a lower order WWTP model. The result shows that by MOR techniques, the higher WWTP system can be simplified to lower order system with a low error of the reduced system. The result of reduced model will be represented in sigma graph and numerical value.

System identification of activated sludge process

In this paper, an activated sludge process model is obtained using two nonlinear system identification techniques. These techniques are Nonlinear ARX Modeling and Hammerstein-Wiener Modeling. A set of raw data from an existing activated sludge process, considered as black box, two different models are obtained and compared in terms of its best fit percentage. From the result, it is concluded that the Hammerstein-Wiener Modeling technique yields better result with lower order and best fit of 91.6%.

Modeling of submerged membrane bioreactor filtration process using NARX-ANFIS model

This paper presents modeling techniques for submerged membrane bioreactor (SMBR) filtration process using. The Nonlinear Auto Regressive with Exogenous Input (NARX) structure was used with adaptive neuro-fuzzy interface system (ANFIS) and feed forward neural network (FFNN) are employed to model the filtration system. The transmembrane pressure and the permeate flux were model during the relaxation and permeate cycle. In this work diluted palm oil mill effluent (POME) will be used as an influent of the treatment process. The performance of the models was measured using the R2, mean square error (MSE) and mean absolute deviation (MAD). The result showed that the ANFIS with NARX structure perform slightly better compare with ANN with NARX structure.