Search

When heard about agarwood oil, it is very familiar with world community because of its beneficial. Unfortunately, there is no any standard grading model of agarwood oil was implemented. As a solution forms, it is very important to come out with a standard of quality classification model for agarwood oil grading's. By continuing of the research for the development of this standard, specific algorithm function has been used to make sure the ability of this model is totally not in doubt. Support vector machine (SVM) has been chosen as a main model and for the specific algorithm function that has been chosen was multiclass function. Then, in the function, the one versus one (OVO) strategy has been used to make multiclass work and can be applied on SVM. The analysis work has involving the data taken from the previous researcher that consists of four classes of agarwood oil quality's samples which are low, medium low, medium high and high quality. So, the output was the classification of quality between low, medium low, medium high or high quality while the input was the abundances (%) of compounds. The desk research has been conducted by using MATLAB software version r2020a for the simulation platform. The result showed that the model by using multiclass function has pass the performance criteria standard. The verdict in this research for sure will be valuable for the future research works of agarwood oil areas, especially quality classification part.

The advancement of computer algorithm and speed has brought about a revolution in modelling and simulation of especially complex systems. Various intelligent methods are available in an easy to use manner for examining and analyzing highly nonlinear systems. This paper demonstrates the elegant application of System Identification technique in processing complex sensor signals and turning them into easily understood and useable measurements. As example, System Identification technique was applied solve a unique issue of grading Agarwood oil which has been based on a manual technique for a long time. For this ground breaking research, data from a chemical analysis of the Agarwood oil, acquired via GCMS measurements, is employed for identification. A single sample of the oil can be represented by hundreds of chemical compounds; these were reduced via a data reduction method. After proper processing of the data thru the System Identification procedure, a valid model was obtained and used to produce a highly accurate grading for the Agarwood oil quality.

These days, agarwood oil becoming a high demand throughout the world and Malaysia is not excluded. It happens due to the variety of usages such as incense, traditional medicine, and perfumes. However, there has been a lack of research on the development of agarwood oil because there is no any standard grading method of agarwood oil was implemented. As a solution forms, it is very important to come out with a standard method of quality classification for agarwood oil grading’s. By continuing of the research for the development of this standard, the comparison of different type of kernel parameter on nonlinear data based on performance measure has been the main objective of this paper. Support Vector Machine (SVM) has been selected as intelligent technique to comparing the output of different type of kernel parameter used. The analysis work has involving the data taken from the previous researcher that consists of two classes of agarwood oil quality’s samples which is high and low quality. For the output of this research was the classification of two different quality while the input was the different percentage of the compounds added. The desk research has been conducted by using a software application named MATLAB with version R2016a. The research indicates that each of different kernel parameter used have pass the performance measures standard. The verdict in this research for sure will be valuable for the future research works of agarwood oil areas, especially quality classification part.

Essential oils are concentrated oils produced by different parts of plants via distillation. Essential oils need to grade into high and low quality due to high quality essential oil being popular and commonly used in variety industries such as in cosmetics, perfumes and traditional medicines. On the previous researcher, many techniques regarding on method of extraction, chemical compounds of essential oil and the quality of essential oils have been presented. Currently, artificial neural network (ANN) is a popular method but this method has many local minima to be chosen as the best for a task. In the future study, Cymbopogon Nardus oil has been chosen for the proposed model due to its applications in mosquito repellent, medicine, food flavor and so on while SVM technique is chosen as classifier in C.Nardus oil. This paper presents an overview of essential oils and its previous analysis technique. Besides, the review on Support Vector Machine (SVM) is done to prove the technique is suitable for future research studies.

Cymbopogon oil is derived from the Cymbopogon species which are lemongrass and citronella. The essential oil is primarily used as a mosquito repellent as a pesticide, and also has other insecticidal, acaricidal and herbicidal activity. It is not considered harmful to humans and pets but may cause skin irritation. The aim of this paper is to model the Cymbopogon oils species using k-Nearest Neighbours (k-NN). The work consists of Cymbopogon oils data extraction using GC and GC-MS machine for compound identification. After that, a quick observation on the major compound in the samples as significant compound identification was done. Then, pre-processing data was performed which consist of data randomization and data splitting into training and testing dataset with the ratio of 80%:20%, respectively. Next, the k-NN modelling was carried out using training dataset and also applying Euclidean as distance metric. It was followed by the tested of developed k-NN using testing dataset. The result showed that k=1 and k=2 achieved 100% accuracy for both testing and training dataset. This proven that k-NN modelling in this study was able to classify the Cymbopogon oil compounds according to its species which are citronella and lemongrass.

This paper presents an attemption to empirically assess statistical learning model Boolean Support Vector Machines (BSVM) to the problem of agarwood oil quality categorization. The modelling starts with data pre-processing of seven significant chemical compounds of agarwood oil, from high and low qualities. During this stage, the data was randomized, normalized and divided into training and testing parts. 80% of the training part was induced as examples and create the maximum margin hyperplane to separates high and low groups in a binary setting and build the model. Another 20% of testing part was used to validate the developed model. MATLAB software version R2016a was used to perform all the analysis. The result obtained a good model utilizing SVM in classifying agarwood oil significant volatile compound quality. The model achieved minimum of 80 % for precision, confusing matrix, accuracy, sensitivity and specificity. The finding in this study will benefit further work and application for agarwood oil research area especially its classification in quality of agarwood oil and many others.

Agarwood is a resinous heartwood and Aquilaria is one of many species that grows a lot in Asia. Traditionally, the quality of agarwood oil is based on color, odor, high fixative properties and consumer perception. This quality grading is performed by trained human graders using sensory panels. Human sensory panels has limitation such as fatigue. Therefore, this study focuses on chemical compounds of Agarwood oil. Using this compounds together with artificial intelligence technique, a new grading system will be proposed. This paper discusses only on the statistical analysis of the chemical compounds. 106 compounds were acquired using GC-MS analysis. To remove insignificant compounds, missing values ratio was computed and out of 109 only 19 compounds remained. These compounds were transformed using natural logarithm to improves the distribution of data.

Agarwood oil, famously known as costly oil, extracted from the resinous of fragrant heartwood. The oil is getting high demand in the market especially from China, Vietnam, India, Middle East countries, and Japan because of its unique odour. As one of the researches in grading the quality of agarwood oil, the evaluation of kernel tuned parameter using Radial Basics Function (RBF) and Multilayer Perceptron (MLP) are presented in this paper to classify the quality of agarwood oil by using support vector machine (SVM). The work involved of selected agarwood oil sample from high to low quality. The output was agarwood oil quality either low or high and the input was the abundances (%) of agarwood oil compoundS. The input and output data were pre-processed by following works; data processing (normalisation, randomisation and data splitting into two parts in which training and testing dataset (ratio of 80%:20%) and data analysis using SVM modelling. The training dataset was used to train in developing the SVM model and the testing dataset was used to test/validate the developed SVM model. All the analytical works were performed automatically via MATLAB software version R2013a. The result showed that SVM model with RBF tuning is better than SVM model with MLP tuning and passed all the performance measures; accuracy, precision, confusion matrix, sensitivity and specificity. The finding in this study is significant and benefits further work and application for agarwood oil research area especially its classification.

This paper presents the analysis of distance metric variations in KNN for agarwood oil compounds differentiation. The work involved of the development of k-Nearest Neighbor (KNN) by varying the distance metrics. The input is abundances (%) of agarwood oil compounds and the output is agarwood oil quality either high or low. The data is divided into two parts; training and testing dataset with ratio of 80% and 20% respectively. The training dataset is used to develop the KNN model from K equal to 1 until K equal to 5, and the testing dataset is used to test the developed model. During the training, distance metric parameters were varied using Euclidean, City-block, Cosine, and Correlation. The performance of each parameter was recorded and observed. All the analytical works are performed automatically via MATLAB software version R2014b. The results showed that, among four distance metric variations, Euclidean and City-block yield 100% accuracy for both training and testing dataset. After that, 89.5% of accuracy was obtained by Cosine and Correlation. In general, the accuracy yielded by all distance metrics is above 80.00% and indicating a good KNN model. This finding proved the capability of KNN in differentiating the agarwood oil compounds to high or low qualities. The results in this study are important and contributed to further research work in agarwood oil grading system.

This study investigates the performance of the Multilayer Perceptron (MLP) classifier in discriminating the qualities of agarwood oil significant compounds by different qualities based on three training algorithms namely Scaled Conjugate Gradient (SCG), Levernbergh-Marquardt (LM) and Resilient Backpropagation (RP) Neural Network by using Matlab version 2013a. The dataset used in this study were obtained at Forest Research Institute Malaysia (FRIM) and University Malaysia Pahang (UMP). Further, the areas (abundances, %) of chemical compounds is set as an input and the quality represented (high or low) as an output. The MLP performance was examined with different number of hidden neurons which is in the ranged of 1 to 10. Their performances were observed to accurately found the best technique of optimization to apply to the model. It was found that the LM is effective in reducing the error by enhancing the number of hidden neurons during the network development. The MSE of LM is the smallest among SCG and RP. Besides that, the accuracy of training, validation and testing of LM performed the best accuracy (100%).

This paper presents the modelling of agarwood oil (AO) significant compounds by different qualities using Scaled Conjugate Gradient (SCG) algorithm. This technique involved of data collection from Gas Chromatography-Mass Spectrometry (GC-MS) for compound extraction. The development of Multilayer perceptron (MLP) is used to discriminate the qualities of AO chemical compounds to the high and low quality. The input and output data was transferred to the MATLAB version R2013a for extended analysis. The input is the abundances of significant compounds (%) and the output is the oil quality either high or low. This involved of identification, selection and optimization of a MLP as classifiers to identify and classify the agarwood oil quality. The result showed that MLP as pattern classifier is successful classify agarwood oil quality using SCG algorithm with 100% accuracy. This finding is important in agarwood oil area especially in grading system.

Agarwood is well known as one of the expensive woods in the world. It has a unique scent which brings it to have wide usages especially in perfumery ingredient, as incense, in traditional medical preparation, and as symbol of wealth. Due to that, this paper presents the analysis on chemical profiles of agarwood chipwood, as a part of agarwood grading system. The work involved of Solid Phase Microextraction (SPME) coupled with Gas Chromatography — Mass Spectrometry (GC-MS) GC-MS in extracting high quality. Three headspace fibers; PDMS-DVB, CAR-PDMS and DVB-CAR-PDMS were used during the extraction to identify the compounds with the sampling time of 60 minutes. The result showed that high quality agarwood chipwood is made up of terpene group which are monoterpene hydrocarbon, sesquiterpene hydrocarbon and oxygenated sesquiterpene. The relative peak areas (%) for compounds are tabulated and plotted. The finding in this study confirmed that the difference in compounds extracted and their relative peak area (%) are due to different fiber's polarity and absorbent, Thus, it is significant and benefit especially in agarwood oil quality grading and its related area.

Recently, the flood modelling based on neural network has become popular among the researches due to its treat not only to human but also animals and plants. This paper presents the application of Z-Score and Feedforward Neural Network (FFNN) in identifying significant rainfall stations in Kelantan River for flood modeling especially at Kuala Krai Station. The daily rainfall and river water level are collected in time series with hourly time interval in year 2014. Seven rainfall stations along the Kelantan River are computed their Z-score values and it is highlighted that two rainfall stations, Tualang (S3) and Kuala Krai (S4) are significant among others. The river water level and rainfall from these stations are used as inputs to FFNN and water level at S4 as output for network employment. Regression for training network is obtained and Mean Square Error (MSE) is varied from one to five neurons for network performance purposes. The result showed that the MSE at two neurons in hidden layer is the lowest among other with the value of 0.0016 and the regression for training network is 0.99987. The findings obtained in this study revealed the capability of Z-Score and the employed FFNN model is success to predict water level at S4 station. Thus, it is significant and beneficial for further work especially for flood modeling and analysis.

Due to its popularity and high market demand, critical analysis on agarwood vapour chemical compounds may provide an alternative quality discrimination of agarwood oil. The proposed work involves the extraction of high quality agarwood using headspace volatile divinylbenzene-carboxen-polydimethysiloxane (DVB-CAR-PDMS) solid phase microextraction (SPME) with different sampling time at 15, 30 and 60 minutes. Then, Gas chromatography — Mass Spectroscopy (GC-MS) is performed to identify the chemical compounds. Generally, agarwood vapour is rich in terpene group especially monoterpene, sesquiterpene and oxygenated sesquiterpene. Analysis showed that at least 52, 50 and 54 compounds are extracted at 15, 30 and 60 minutes, respectively. Among all, duration of 60 minutes produced the highest abundance (%) for caryophellene oxide. The finding proves that caryophellene oxide as one of the important compounds in high agarwood and different sampling time plays a major role that effects the extraction. Thus, the analysis in this study is significant and brings benefit especially to the agarwood and its essential oil research area.

In Malaysia, east coast of peninsular is experiencing the rainy season between mid - October until March every year. Heavy seasonal rains cause the Kelantan River to overflow and flood the surroundings area. In this paper, the application of feed forward multi-layer perceptron (FFMLP) in neural networks for flood water level prediction is presented. The method focused on the neuron variation in hidden layer. By using measured data of three stations; Tualang, Kuala Krai and Kusial, FFMLP neural networks was developed. The inputs are water level river at three stations and output is water level river at Kusial station. The numbers of neuron in hidden layer were varied from one to ten and Levenberg Marquadt algorithm is used to train the network. The performance of network was evaluated using Mean Square Error (MSE). It is shown that three neurons in hidden layer afforded the lowest MSE, 0.043. The Regression, R for training network is closed to 1 (0.991), supports that the model is acceptable and able in predicting water level at Kusial station.

This paper presents the comparison between ARX model and ARMAX model using the data from steam distillation essential oil extraction (SDEOE). The work is implementing system identification approach. The aim of this research is to identify the performance of ARX and ARMAX model toward the system. The input for every model is either Pseudo Random Binary Sequence (PRBS) or Pseudo Random Sequence (PRS) with temperature as output. Each dataset consists of 3000 sample, and being separated into estimation and validation for model estimation and validation with the ratio of 2000:1000, respectively. All the analysis work is done via Matlab R2013a. The result generally showed that ARX is slightly better than ARMAX model especially for model best fit. However, generally both models ARX and ARMAX can be implemented for PRBS and PRS perturbation since it passes all the validation's criteria in this study; model fits, correlation test and distribution of the residual. This finding is important and it will benefits for further work in steam distillation extraction especially for controller design and real time application.

This paper presents a review on agarwood and its quality determination. It was found that Aquilaria Malaccensis is the main species in Malaysia and Indonesia. The agarwood has different quality or grades; similarly its essential oil. They are traded at different price according to their quality. Scientific reason showed that chemical compositions of agarwood make them different from each other, thus affect their quality. Physical appearances such as odour and colour have been used to grade their qualities. However the method has drawbacks in terms of subjectivity, poor reproducibility, time consumption and large labour expense. Therefore there is a requirement for agarwood and its oil to be graded according to their chemical profiles. This interest began to attract the researchers and also is mainly to ensure the quality of agarwood oil.

This paper presents a review on significant parameters and techniques in water quality assessments used for drinking, domestic purposes and recreational purposes. To ensure water in good quality there are several parameters that have to be considered. Water quality index (WQI) is one of the most effective tools to an attempt to ascertain the water quality. There are many techniques on water quality assessments and some of them are regression analysis, fuzzy reasoning, support vector machine (SVM) and Artificial Neural Network (ANN). The objectives of this review paper are to determine the parameters and ideal technique for water quality assessment.

This paper presents the performance of Non-Linear Auto Regressive with Exogenous input (NLARX) model structure that is applied in modeling of induction based steam distillation system. The input is Pseudo-Random Binary Sequence (PRBS) and the output is temperature. The input-output data was split into two equal set for model estimation and model validation. All the data are transferred to MATLAB R2013a software for analysis. Wavelet Network, Sigmoid Network, Tree partition Network and Feedforward Neural Network are the nonlinearity estimators used to build the NLARX model structure and their performances have been compared. The validation of estimated model will be based on best fit (R2), final prediction error (FPE), loss function, auto-correlation function (ACF) and cross correlation function (CCF). The result showed that NLARX with Feedforward neural network is the most suitable estimator among others due to it yields the highest percent of best fit (R2), lowest final prediction and loss function, and all the coefficients are within the confidence limit for CCF test.

This paper presents the proposed Z-Score in identifying the significant chemical compounds of agarwood incense smoke. The technique is applied to commercial, low and high quality of agarwood samples and the extraction of the compounds were performed by gas chromatography — flame ion detector (GC-FID) and gas chromatography — mass spectrometry (GC-MS) coupled with solid phase microextraction (SPME) analysis. The Z-Score has successfully identified eight significant compounds in those qualities of agarwood smoke. They are β-maaliene, nor-ketoagarofuran, epoxybulnesene, 10-epi-γ-eudesmol, agarospirol, α-eudesmol, epi-α-bisabolol and rotundone. Among all, epoxybulnesene gave the highest abundant of 14.81%. The finding from this study showed that the chemical compounds for agarwood smoke is varied depending on their qualities. The identified compounds are useful for further study in agarwood smoke quality grading analysis

The clutter model for ultra high frequency (UHF) band for forward scatter radar (FSR) is developed in this paper using Log-Logistic distribution. The model follows maximum likelihood estimation (MLE) approach in finding the parameters for the distribution as well as root mean square error (RMSE) of the data between the clutter signal and statistical model as the goodness of fit (GOF) test. The model is then compared to four different distribution models, which are Log-Normal, Gamma, Weibull and Nakagami. The results of the comparisons show that Log-Logistic model best fits the clutter signal.

Research on gait is increasing among researchers and got worldwide attention. In order to explore and inspect the nature of variables and as part ongoing research in gait studies among children, this paper presents preliminary study of gait which is involved of analyses of several factors i.e. speed, gait cycle and leg length of male and female children. The analysis is performed using statistical techniques; boxplot, correlation and several plots which are done via SPSS software. The results show that there is significant differences between male and female for the variable walking speed, gait cycle duration and leg length. Therefore, in order to characterize human behavior, the walking speed, gait cycle duration and leg length are important parameters.

This paper presents the application of Artificial Neural Network (ANN) in agarwood oil grade classification. The work involved of the extraction of chemical compounds by GC-MS, identification the significant chemical compounds using Z-score, generating the synthetic data using a dedicated formulae and application of ANN classification. The ANN classification is performed and its performance is measured using accuracy, sensitivity and specificity. The result showed that the performance of ANN classification for original GC-MS data is increasing when the data is added with synthetic data. This study showed that the ANN application in this study required a large number of sample size for it to have high accuracy in classification.

This paper presents the application of k-Nearest Neighbor (k-NN) in grading the quality agarwood oil. Six agarwood oil samples obtained at Forest Research Institute Malaysia (FRIM) were extracted and their chemical compounds were examined by GC-MS. The work is followed by the grading system using the proposed k-NN. The study shows that there are 10 significant chemical compounds of agarwood oils. They are β-agarofuran, α-agarofuran, 10-epi-□-eudesmol, □-eudesmol, longifolol, oxo-agarospirol, hexadecanol and eudesmol. These compounds are used as inputs to the k-NN algorithm for grading them. The performance of the k-NN is measured and the highest accuracy obtained by k-NN which is above 83.3% shows that k-NN is a reliable classifier in grading the agarwood oil quality.

This paper presents classification of the quality of agarwood oils from Malaysia using Z-score technique. Six agarwood oil samples named as MU, MUS, MN, MNS, R5 and HD are analyzed by GC-MS to examine their chemical profiles. The extraction showed that at least forty three volatile compounds are found. The Z-score technique is proposed to identify the significant chemical compounds of the agarwood oils. It is found that six chemical compounds are recognized. They are β-agarofuran, α-agarofuran, 10-epi-□-eudesmol, α-eudesmol, dihydrocollumellarin and □-eudesmol. These volatile compounds have different abundances pattern responsible to the different qualities of agarwood oil such as high and low. The Z-score applied in this study give a promising result in discriminating agarwood oil to high and low quality. It is important and useful in solving the grading agarwood oil system which is currently done manually.

A Forward Scatter Radar topology based on two-ray path propagation configuration is discussed. Based on this configuration, a study on the influence of different target's crossing baseline lengths to the target's spectra in Forward Scatter Radar (FSR) is conducted. By using simulated and measured signals, it was observed that, the stability and similarity of target's power spectra depends on the baseline length. The employment of baseline normalization procedure is shown to improve significantly the differences in the spectra.

The quality of the agarwood oils plays an important role as it reflects the price of the oil in the market. High quality agarwood oil is traded extensively. In order to obtain good quality agarwood oil, temperature condition during the extraction is one of the important parameter instead of type and time of extraction. An analysis of chemical compounds of agarwood oil involves of three different temperature conditions such as 40°C, 60°C and 80°C has been carried out in this study. The analytical work involves HS-SPME combined with GC-MS analysis to extract and identify the chemical compounds composition in agarwood oil. The zscore technique was introduced to identify the significant compounds that contribute to the aroma of agarwood oil. The result shows that five compounds were identified and they were aromadendrane, β-agarofuran, α-agarofuran, 10-epi-□-eudesmol and □-Eudesmol. These compounds were found to have similar pattern of plot but different in chemical composition due to the different chemical properties of the oils.

This paper presents a correlation analysis between high grade and low grade of Aquilaria Malaccensis (AM). Seven selected samples of AM were identified using Gas Chromatograph - Mass Spectra (GC-MS) and several statistical techniques have been employed. The samples were separated into two grades such as high grade and low grade. The analysis involves of normality test such as Shapiro Wilk, standard deviation, skewness, kurtosis and correlation analysis. All the analysis was done using SPSS ver. 15.0. The result shows that there are no significant correlation between high grade and low grade and it is a high correlation among compounds in similar grade. Three common compounds and highly correlated found in both grade and they are hexadecanoic acid, valencene and α-guaiene.

Agarwood oil has been widely used especially in fragrance, incense, prayers and traditional medicinal. In the Middle East, the market demand for Agarwood oil is very high. Agarwood oil is traded based on high grade and low grade, corresponding to expensive price and cheap price, respectively. Currently, the grading of Agarwood oil, specifically Aquilaria Malaccensis, depends on its physical appearance such as color and odour. This paper presents the analysis of Aquilaria Malaccensis based on GC-MS data. The work involves of statistical technique such as boxplot and PCA. The analysis part was done on 64 chemical compounds on 7 samples of agarwood oil obtained by Forest Research Institute Malaysia (FRIM). It was done via MATLAB ver. R2010a. The result shows that the distribution of chemical compounds in Agarwood oil is not normal and five componets is identified from 64 variables Agarwood oil samples, gathered by boxplot and PCA, individually.

This paper presents an investigation of ARX model on partial input-output data from heating process in a steam distillation essential oil extraction (SDEOE) system. The study was performed by applying system identification and it was carried out by using four general stages; experimental work to collect input-output data, model structure selection where ARX model is chosen, estimation to obtain the ARX model, and validation. The result from this investigation showed that the ARX model obtained by the partial input output data is sufficient and adequate, i.e. the model behaviour of heating process is closed to the real system performance.

This paper presents a review on models for heating process in food, manufacturing and agricultural industries. The review come out with the table to summary on different modeling technique and their developed models for heating process in various applications. There are many techniques to develop the heating process and some of them are numeric, first principle, system identification and a combination of first principle and system identification. During the review, it was found that heating is a linear process and more review for linear model on heating process are done. They are Finite Impulse Response (FIR), Auto-Regressive with Exogenous Input (ARX) model, Auto-Regressive Moving Average (ARMA) model, Auto-Regressive Moving Average with Exogenous Input (ARMAX) model, Output-Error (OE) model and Box-Jenkins (BJ) model. Among all, the review has highlighted that ARX model is the simplest structure, easy to find analytical solutions and provide excellent performance.

In this paper, recurrent adaptive neuro-fuzzy inference system (RANFIS) structure has been proposed to solve approximation problem in identifying a global model of steam temperature of packed distillation column in steam distillation essential oil extraction process. The input-output data is acquired from field experimentation via MATLAB Real-time Workshop (RTW) integrated to the plant. The derived RANFIS model is optimized in order to get the optimum ANFIS structure that includes the optimal number of membership function, fuzzy rules, data selection, epoch which gives low computation time and root means squared error (RMSE). Several experiments were carried out using both pseudo random binary sequence (PRBS) and noise as perturbation signals. Performance comparison of RANFIS with ARX model shows that RANFIS identification gives an excellent global modeling method with RMSE of 0.1778 and consumed less computation or training time.

This paper presents the development and implementation of an ARX (Auto-Regressive eXogenous) model for Model Predictive Control (MPC) in steam distillation extraction system. The mathematical model was developed using the system identification technique. The MPC is proposed as a controller in a way to regulate the system to maintain the optimum operation temperature besides minimizing the energy that is used to power up the plant. Tuning of the MPC was examined with several values of prediction horizon with the similar default parameter to achieve the optimal setting for the better controller's performance. The performance of MPC was evaluated at optimal temperature setting. Simulation result indicates that MPC is able to control the steam temperature in more efficient way using a first order ARX model.

This paper studied the robustness of ARX model by varying the noise sequence adapted to the output of a heating process. The ARX model is fitted from output data which is obtained from a pilot plant of essential oil extraction by applying steam distillation technique. The plant is supplied by 1.5kW power delivered to the heating element which then is converted into PRBS signal as input and produced output which is steam temperature from ambient (23°C) to steady state condition (93°C). The noises with different variations have been used to corrupt the validation data, which ranged from ±0.2°C to ±2.0°C. Noises and PRBS signal both are generated using Matlab software version R2008a. The study has shown that as noise is increased, there are small decrement of ARX model fit. Therefore the fitted ARX model is highly suitable for any extended contribution in controller design, prediction and simulation.

This paper presents a model of dynamic response of essential oil extraction process using system identification approach. A collection of samples was collected from a pilot plant of essential oil extraction process using steam distillation technique. Input signal for the process is pseudo-random binary sequence (PRBS) and the output is temperature. The sample was separated into training and testing data by using interlacing technique. Based on Auto Regressive Exogenous Input (ARX) model validation, the results showed that partial data will produce adequate model to describe the full dynamic of essential oil extraction process.

This paper presents estimation on nonlinear ARX model for steam distillation process by wavenet estimator. A set of input-output data samples was collected at a sampling rate of 0.1 second from an essential oil extraction pilot plant by using steam distillation technique. PRBS input signal delivered to the heating element of the plant while the output was the steam temperature. By applying interlacing technique the data was split into 2 equal size subsets of data. One set of data was used for model estimation and the other one was used for model validation. During the estimation, we applied Nonlinear ARX based on wavenet estimator with different unit i.e. unit 1, unit 2, unit 3 and unit 8 in order to see their performances. We also modeled linear ARX for performance comparison. The comparison was done in term of the agreement between measured data and those models based on wavenet with different unit and the result has shown that the nonlinear ARX based on wavenet estimator with 3 units has outperformed the other models. For future work, the result obtained from estimation is contributed for extended application such as Hammerstein-Wiener model.