Ahmad Banakar

Optimization of ultrasonic assisted continuous production of biodiesel using response surface methodology.

This paper evaluates and optimizes the continuous production of biodiesel from waste cooking oil. In this research work, methanol and potassium hydroxide were used as catalyst engaging response surface methodology. For this purpose, the central composite experimental design (CCED), the effects of various factors such as irradiation distance, probe diameter, ultrasonic amplitude, vibration pulse and material flow into the reactor on reaction yield were studied to optimize the process. The results showed that all of the considered parameters affect the reaction efficiency significantly. The optimum combination of the findings include: irradiation distance which was 75 mm, probe diameter of 28 mm, ultrasonic amplitude of 56%, vibration pulse of 62% and flow rate of 50 ml/min that caused the reaction yield of 91.6% and energy consumption of 102.8 W. To verify this optimized combination, three tests were carried out. The results showed an average efficiency of 91.12% and 102.4 W power consumption which is well matched with the models predictions.

Potential Applications of Computer Vision in Quality Inspection of Rice: A Review

Among the cereals, rice is the major foodstuff for a large part of the world’s population. Due to its tremendous importance in the global market, its qualitative economic aspects during processing have always been attended by producers. As the most delicate of the cereals, rice needs the utmost care during post-harvest handling and processing, because in most cases, it is consumed as whole kernel. The growing demand for production of rice with high-quality and safety standards has increased the need for its accurate, fast and objective quality monitoring. Computer vision techniques, as novel technologies, can provide an automated, nondestructive and cost-effective way to achieve these requirements. In recent years, various studies have been conducted to evaluate rice qualitative features based on computer vision techniques. This paper presents the theoretical and technical principles of computer vision for nondestructive quality assessment of rice combined with a review of the recent achievements and applications for quality inspection and monitoring of the product.

Parameter identification of TSK neuro-fuzzy models

Abstract In the framework of the TSK neuro-fuzzy model a combination of the two well-known identification methods are employed for parameter estimation of the neuro-fuzzy inference system, namely the series–parallel and the parallel configurations. The presented paper proposes two new possible configurations for identifying the parameters of the TSK neuro-fuzzy model using the combinations of these two existing configurations. One of the proposed configurations constitutes the series–parallel configuration to the premise part and the parallel configuration to the consequent part of the neuro-fuzzy model, termed as PS-P configuration. The second one is composed of the series–parallel configuration to the consequent part and the parallel configuration to the premise part of the neuro-fuzzy model, termed as CS-P configuration. The presented work mainly deals with a comparative study of the proposed configurations and the existing configurations in the context of parameter identification of the TSK neuro-fuzzy model on three different benchmark examples. Moreover, it investigates upper bound of the learning rates, using the Lyapunov stability theorem, to assure the stability and the convergence of the model learning process. Implementation of the modified mountain clustering (MMC) and the cluster validity function yields initial models. To restrict the upper bound during the learning process it also presents a two-phase adaptive learning rate.

Local recurrent sigmoidal-wavelet neurons in feed-forward neural network for forecasting of dynamic systems: Theory

In this paper different structure of the neurons in the hidden layer of a feed-forward network, for forecasting of the dynamic systems, are proposed. Each neuron in the network is a combination of the sigmoidal activation function (SAF) and wavelet activation function (WAF). The output of the hidden neuron is the product of the output from these two activation functions. A delay element is used to feedback the output of the sigmoidal and the wavelet activation function to each other. This arrangement leads to proposed five possible configurations of recurrent neurons. Besides proposing these neuron models, the presented paper tries to compare the performance of wavelet function with sigmoid function. To guarantee the stability and the convergence of the learning process, upper bound for the learning rates has been investigated using the Lyapunov stability theorem. A two-phase adaptive learning rate ensures this upper bound. Universal approximation property of the feed-forward network with the proposed neurons has also been investigated. Finally, the applicability and comparison of the proposed recurrent networks has been weathered on two benchmark problem catering different types of dynamical systems.

Machine vision system for grading of dried figs

Display Omitted In this study, a grading system based on machine vision was developed for grading figs.An image processing algorithm was introduced to extract fig marketing features such as color, size and split size.An algorithm for detecting the whole of fig in every trigger of camera was developed.This systems fabricated feeding unit, simulation unit, machine vision and separation unit.This system showed which has good ability for grading figs into the five classes. Fig is a horticultural product which requires sorting at the postharvest stage before being marketed. In this study, a grading system based on machine vision was developed for grading figs. The system hardware was composed of a feeder, a belt conveyor, a CCD camera, a lighting system, and a separation unit. Three quality indices, namely color, size, and split size, were first classified by fig-processing experts into the five classes. Then, the images of the fig samples were captured using a machine vision system. First, the length of pixels in each image and longitudinal coordinates of the center of gravity of fig pixels were extracted for calculating the nozzle eject time. For extracting the three quality indices of each class, a machine vision algorithm was developed. This algorithm determined color intensity and diameter of each fig as the indicators of its color and size, respectively. For calculating the split area, the images were first binarized by using the color intensity difference between the split and other parts of the fruit in order to determine the area of the split section. A grading algorithm was also coded in Lab-VIEW for sorting figs based on their quality indices extracted by the image processing algorithm into five qualitative grades. In the grading algorithm, the values of these features were compared with the threshold value that was predetermined by an expert. Results showed that the developed system improved the sorting accuracy for all the classes up to 95.2%. The systems mean rate was 90kg/h for processing and grading figs.

Design, development and performance evaluation of an automatic control system for rice whitening machine based on computer vision and fuzzy logic

An automatic control system was developed for controlling rice whitening machine.The system was designed based on machine vision and fuzzy logic techniques.The developed system had a high accuracy in controlling of the whitening machine.Performance speed of the system was satisfactorily higher than human operator.Decisions made by the system resulted in a good improvement in the quality of product. The objective of this study was developing an intelligent automatic control system (ACS) based on machine vision and fuzzy logic techniques to control the performance of rice whitening machines. The developed ACS consisted of two main parts, namely hardware (including sampling unit, kernel singulation unit, image capturing unit, processor (computer), discharge pressure control unit and data acquisition unit), and software (including image processing, fuzzy inference and central control units). Two important qualitative indices, degree of milling and percentage of broken kernels, were considered as input variables and the level of pressure on the discharge section of the whitening machine was selected as the output variable in the fuzzy inference unit. Results of the evaluations indicated that the developed ACS had 89.2% accuracy in determining the desired working conditions for the whitening machine. The total time of each monitoring round was, on the average, equal to 14.73s, of which 6.4s was devoted to kernel sampling and transporting the samples into the imaging chamber, 7.33s for taking three images from the kernels, processing the captured images and executing the fuzzy inference process, and the remaining 1.5s for making the adjustments in the level of pressure in the mechanism. Based on this information and in contrast to the corresponding time spent by the human operator to perform a similar process, it was revealed that the performance speed of the ACS was, on average, 31.3% higher than that of the human operator. Evaluation of the samples obtained from the discharge section of the rice whitening machine at different stages of the control process showed that the decisions made by the developed ACS during the control process resulted in a satisfactory improvement in the quality of the output product.

COMPARATIVE STUDY OF WAVELET BASED NEURAL NETWORK AND NEURO-FUZZY SYSTEMS

Based on the wavelet transform theory and its well emerging properties of universal approximation and multiresolution analysis, the new notion of the wavelet network is proposed as an alternative to feed forward neural networks and neuro-fuzzy for approximating arbitrary nonlinear functions. Earlier, two types of neuron models, namely, Wavelet Synapse (WS) neuron and Wavelet Activation (WA) functions neuron have been introduced. Derived from these two neuron models with different non-orthogonal wavelet functions, neural network and neuro-fuzzy systems are presented. Comparative study of wavelets with NN and NF are also presented in this paper.

An Intelligent Procedure for the Detection and Classification of Chickens Infected by Clostridium Perfringens Based on their Vocalization

In this study, an intelligent method was implemented for the detection and classification of chickens by infected Clostridium perfringens type A based on their vocalization. To this aim, the birds were first divided into two groups that were placed in separate cages with 15 chickens each. Chickens were inoculated with Clostridium perfringens type A on day 14. In order to ensure the absence of secondary diseases and their probable effect on bird vocalization, vaccines for common diseases were administered. During 30 days of the experiment, chicken vocalization was recorded every morning at 8 a.m. using a microphone and a data collection card under equal and controlled conditions. Sound signals were investigated in time domains, and 23 features were selected. Using Fisher Discriminate Analysis (FDA), five of the most important and effective features were chosen. Neural Network Pattern Recognition (NNPR) structure with one hidden layer was applied to detect signals and classifying healthy and unhealthy chickens. Firstly, this neural network was trained with 34 samples, after which eight samples were tested for accuracy. Classification accuracy was 66.6 and 100% for days 16 and 22; i.e., two and eight days after the disease, respectively. The results of this study demonstrated the usefulness and effectiveness of intelligent methods for diagnosing diseases in chickens.

WAVELET NEURO-FUZZY MODEL WITH HYBRID LEARNING ALGORITHM OF GRADIENT DESCENT AND GENETIC ALGORITHM

In this paper, a Wavelet Neuro-Fuzzy model has been proposed. The proposed work caters an application of wavelet network used in fuzzy systems for forecasting of dynamic systems. A wavelet network approximates the consequent part of each fuzzy rule. The wavelet network is a feed-forward neural network with one hidden layer that uses a combination of Wavelet and Sigmoid Activation Function. A hybrid learning method composed of genetic algorithm and gradient descent is proposed to tune the learning parameters of the proposed Wavelet Neuro-Fuzzy model. Further, an analysis regarding the convergence and stability of gradient descent learning is presented for the proposed Wavelet Neuro-Fuzzy model. To evaluate the effectiveness of proposed model and learning strategy, three different classes of benchmark problems have been considered.

Fault detection of engine timing belt based on vibration signals using data-mining techniques and a novel data fusion procedure

In this research, an intelligent procedure was designed and implemented based on vibration signals for detecting and classifying prevalent faults of an internal combustion engine timing belt. The vibration signals of the timing belt were captured during operation in six different states: healthy, tooth crack, back crack, wear, separated tooth, and oil pollution. These signals were processed at three domains, namely, time, frequency, and time–frequency domains. Time-domain signals were transformed into the frequency and time–frequency domains using fast Fourier transform and wavelet transform, respectively. Then, six statistical features were extracted from vibration signals at all three domains. The extracted features were used as inputs to an artificial neural network for the primary classification of timing belt defects. Classification accuracy of artificial neural network in detecting and classifying timing belt faults in the time, frequency, and time–frequency domains have obtained 71%, 78%, and 84%, respectively. Combining separate classification accuracies from time, frequency, and time–frequency domains has been implemented using Dempster–Shafer theory of evidence. Classification accuracy based on the fusion of time- and frequency-domain classifiers was 97%, from time and time–frequency results was 98%, and from frequency and time–frequency results was also 98%, whereas the combination of results for all domains led to a >99% accuracy. Results show that the proposed methodology can detect and classify timing belt defects with high precision and reliability before failure occurrence.