AMIRKABIR
Year:2002 | Volume:13 | Issue:52|Papers Count:24

this paper, genetic algorithms are used in the design and robostification of various moel-based/non-model-based fuzzy-logic controllers for robotic manipulators. It is demonstrated that genetic algorithms provide effective means of designing the optimal set of fuzzy rules as well as the optimal domains of associated fuzzy sets in a new class of model-based-fuzzy-logic controllers. Furthermore, it is shown that genetic algorithms are very effective in the optimal design and robustification of non-model-based multivariable fuzzy-logic controllers for robotic manipulators.

This paper present a joint numerical-experimental study of the flow behaviour in a vertically downward intake in absence of any rotational effects. A numerical method based on finite volume is applied on governing equations in a cylindrical coordinate system to simulate the gravitational flow with fixed free surface grid. Results introduce critical condition based on Froude number. In this study, it is assumed that the critical condition is due to existing of continuous zero pressure line from the free surface to the intake pipe. In the region Fr<1.0, a fairly good agreement between numerical results and experimental data are observed. Predicted submergence depth from numerical approach is less than values from the experimental data for Fr>1.0. It expresses that the critical condition in high flow rate highly depends on random circulation due to non-uniformity in the flow field rather than flow regime and hydraulic parameters.

A servo system moves a mass from one position to another as quickly as possible. A servomotor: requires very high short term torque and should be able to accelerate and decelerate very quickly. In other words, a servomotor should be able to produce high torque and its inertia must be low. This paper concentrates on the mechanical design optimisation of the motor shaft and rotor assembly, given the existing physical constraints of the prototype motor. The existing physical constraints include the stator and rotor dimensions. To reduce the cogging torque, the different magnet shape and skew were manufactured and tested. Optimum shaft and rotor configuration were developed using theoretical calculations and FEA techniques. Using the above design the ratio of torque to the moment of inertia of the rotor goes to its minimum amount. As a result, the acceleration and deceleration of the system is improved.

In this paper we propose a new guidance law based on fuzzy logic can de successfully used for modelling and generating complicated offensive maneuver in a planar air-to-air combat encounter between two aircrafts. Based on fighter's pilot decision-making process, a two-phase pursuit law represented as fuzzy "if... then..." rules is introduced. The rules are obtained from expert's knowledge Each rule relates the desired moving direction of the pursuer to combat parameters such as distances and heading angles. Massive simulations are used to ensure the satisfactory performance of the model.

Multiple access interference and near-far effect cause the performance of the conventional single user detector in DS/CDMA systems to degrade. Due to high complexity of the optimum multiuser detector, suboptimal multiuser detectors with less complexity and reasonable performance have received considerable attention. In this paper, we analyze and examine the performance of multilayer perceptron neural networks using back-propagation algorithm as multiuser detectors of CDMA signals in A WGN and multipath fading channels. Our results show significant improvement over the previous research. We compare the performance of neural network with the other detectors used in CDMA system. We also apply different neural networks and criterions such as the decision based, fuzzy decision, discriminative learning, minimum classification, and cross entropy neural nets and compare their performance. Further, we propose modified decision based network which improves the performance of decision based network

AVR system which are used for generator voltage regulation, may have negative influence on damping of power swings. To eliminate this negative influence a supplementary control loop known as the power system stabilzer (PSS) is used. Different methods have been developed for PSS design problem which among them linear quadratic regulators (LQR) can be pointed. Recent researchs in control engineering show that using nonlinear controllers may increase the stability of system with controller. In this paper concepts of nonlinear optimal control are used for design of nonlinear PSS. For this purpose HJB equations for generator & AVR are introduced and solved with an approximation method. It has been shown that the stability of the system with nonlinear PSS is more than linear PSS whose design is based on LQR method. This conclusion has been confirmed with simulation results on a typical system.

Winding function model is widely used for analyzing the behaviour of induction motor, but there is no systematic method for accurate determination of its parameters. To compute, rotor bar leakage inductance, stator leakage inductance, average air gap and rotor bar resistance, approximate formulas are used. In this paper for the first time a systematic method for accurate parameter identification of winding function model is presented. The proposed scheme uses a genetic algorithm with data collected from a simple experiment on the selected motor. The genetic algorithm is real coded, and in order to be able to use it, the model should be modified to take into account the actual winding distribution in a slot. The validity of the scheme, correctness and accuracy of the parameters obtained are shown by comparing and finding the matching of experimental and simulation results, for a machine under different operating point conditions, including stator turn faults.

An implementation of a speaker verification system is presented in which the speaker claims his/her identity by a seven digits PIN over telephone. PIN digits are recognized by a hybrid predictive neural network system using multi-layer perceptron and dynamic programming algorithm. The speaker verification was done by comparing a speaker's model to his/her speech when uttering the seven digits PIN. The speaker's reference model was trained using Bumb tree and Genetic algorithm. The total number of epochs used to train the neural network, feature types and the dimension of feature vectors, and the effect of different neural net parameters were studied on Farsi digit recognition performance. Some other experiments were conducted to assess the performance of the proposed speaker verification method and to evaluate the effect of different types of decision threshold calculation on the speaker verification error rate.

This parer presents a new method to solve the problem of finding the best configuration of fuel assemblies in a PWR (Pressurized Water Reactor) core, Finding the optimum solution requires a huge amount of calculations in classical methods. It has been shown that the application of continuous Hopfield neural network accompanied by Simulated Annealing method to this problem not only reduces the amount of calculations, but also guarantees finding the best solution. In this study flattening of neutron flux inside the reactor core of Bushel NPP is considered as the objective function. The result is the optimum configuration which is in agreement with the pattern proposed by the designer.

Rotary heart pre-heat furnace in Ahwaz pipe Mill Company is numericall simulated and analyzed using a computercode Temperatures of different zonesof furnace which are necessary as the boundary condition for the numerical simulation are measured using themmocouples. Results of the numerical simulation reveal that the velicity of motion of steel billet and the lieatflux through the burneers are the main cotrolling parameters in regulating the temperature distribution ot billet Steel billet enter thefurnace at the ambient temperature, but the temperature distribution at the exit is very furnace at the ambient temperature, but the temperature distribution art the exit is very important The temperature of the center-line of billet and pipe should be higl enough to let the easy rolling process and prevent breakage of the rollers, and on the other hand the urface temperature of the billet should also be kept low enough to prevent the oxidation of the surface. Numerous numerical experiments are conducted to choose the best set of the effective parameters. In the otlrer ca.Yes although the resident time of the billet could be reduced, but it may lead to either low temperature of the centerlineor iilcreasing precentage of oxidation of Under the best condition, 1.2 percent of oxidation is result. By reduction of the residencetime of billet by 15 percent which is dueto the increaseof the temperature pof the different zones of thefurnace, the ampunt of oxidation increaseto 4 percent The use of the selcted parameters in therear production proces in tire factory shows that a good research may lead to establishment of the furnace design technolagy.

In this paper an analytical and experimental investigation of wrinkling of cone in the, explosive forming is presented. An analytical method based on energy analysis is presented which gives governing equation of wrinkling. Solution of this number and initiation point of wrinkles. Effects of mechanical property and plate dimensions on wrinkling may be analyzed by this equation. In this paper governing equation of wrinkling which was presented in Ref. [22] is modified and explosive cone forming is developed by using blank holder to control the cone wrinkling. The analytical results of this paper agree, with the experimental results.

The purpose of this paper is to analyze the interacting forces acting upon reinforced cylinderical composite shells. The shells are reinforced by stringer and circumferential stiffener (ring or frames) each locatedat equal intervals along the length and circumference. The stiffened shell is subjected to combined axial loads and internal pressure.The effects of the intracting forces on connections, and internal momemts are considered,and the displacement distributions and strains on the shell, ring, and stringer are also evaluated.The article also highlights the calculations of the critical interacting force which is a function of radius and thickness with the aim of designing the connections between the stiffeners and the shell. Furtheremore, the effects of interval between rings and stringers on interacting forces are calculated and the probable seperation mechanism of stiffeners from the shell is dicussed.The article concludes with the comparison of obtained results and the corresponding available data.

In this study a new method is presented in which the number of neps of a carded webs, their area, distribution and also an estimation of their shape are determined by means of image processing. In comparison to most other methods, this method is quick, cheap and accurate.

Polypropylene yarn has poor textureability, wich affects its textile uses. Textureability of pp filament ytirns have been Improved by adding liquid crystal polymer (LCP) and polyamide 6 (PA6). The best condition was achieved by using 5% PA6 together with compatibilizers in tR The blends with liquid crystal polymer decreased the physical properties of the yarn, but Iblends of 3% LCP improved the crimp properties of the yarn. fibrilar formation and its adhesion to the matrix improved the textureability of the polyblends.

In this research new method for measuring of cloth drape was obtained. With this new method we can measure horizontal drape of cloth. At the end horizontal and vertical drape were compared.

Corrosion behaviour of sintered Nd-Fe-B permanent magnets was investigated using potentiodynamic polarization measurements in 0.001, 0.01 and 0.1% H2SO4 and HCl solutions. The results show that the corrosion rate of the magnets in H2SO4 solution is I higher than that of HCI solution. By increasing the amount of Co and Ni and/or reduction of inter granular phases, a significant improvement in the corrosion resistance of the magnets was observed. In H2SO4 solution corrosion is attributed to the selective oxidation I boundaries, but in HCI solution the corrosion occurs by intergranular attack leading to formation of pits.

In this paper, enzymatic hydrolysis treatment, which cause the biological degradation of cell wall of oil seeds, such as soybeam and sunflower before oil extractionfor increase in the yield of oil extraction were investigated. Pectinase at pH=5 and 50C in 10 and 12% showed the best results in soybean and sunflower treatment. Alpha amylase at 1% 55C and pH=5 resulted 0.2% increase in oil recovery. Mixture of pectinase and cellulase at equal volume, 50C, pH=4.8 in 14% (v/w) with treament duration of 10 and 12 hrs for soybean and sunflower respectfully showed high effect and increase in oil yield of 1.43% for soybean and 1.56% for sunflower. Oil extraction after enzymatic treatment had lower iron, peroxide and iodine value and higher free fatty acids compared to direct solvent extraction of oil from seed oils.

In Compressed domain image retrieval, the image compression methods must be selected which include image content accessibility. Image correlation is a significant feature that describes image content. With this approach, in this paper a new image-indexing scheme in compressed domain will be proposed. The new scheme exploits the genealogical relation of indices obtained from TSVQ-based compression method and generates a feature vector as index of the image. This feature vector is capable to describe image content at different resolutions and is robust against the scaling, rotation and translation variations. This paper compares the effectiveness of the new scheme with two similar methods; the VQ and the color-based schemes in the compressed and pixel domain, respectively. VQ is a compression scheme that exploits only intra-block correlation and color-based scheme avoids any image correlation for generating feature vector whereas the TSVQ scheme exploits inter-block as well as intra-block correlation of image. In the VQ and color based methods, the role of image correlation is less than the TSVQ-based scheme, consequently TSVQ-based scheme provides better retrieval results. The experimental results show as exploiting of image correlation in generating of feature vector becomes more, the effectiveness of image retrieval scheme increases. In this regard, the experimental results show that the precision rate of the TSVQ-based scheme, is 1.56 and 4.22 times better than the VQ and color based schemes, respectively. These ratios for recall rate are 1.51 and 3.36, respectively.

One of the effective drying mechanisms is fluidized bed dryer where particles are fluidized by hot air. Mixing and heat transfer are very rapid, therefore the drying time is short. In order to study the mechanism of drying under different conditions a two-fluid model is imposed. The model uses one-dimensional conservation of mass, heat and momentum equations for each phase of gas and solids. Numerical method is carried by finite-volume scheme. The verification of the theoretical results was done by the use of an experimental setup. The results indicate a good qualitative agreement. By eliminating the convective terms in a solid phase and drags model; packed bed condition is obtained. The result verified with the performed packed bed experimental data. Futher, with the proposed model, the performance and comparison of fluidzed and packed bed dryer can be shown.

The Sungun porphyry copper deposit that intruded Eocene volcanosedimentary and Cretaceous carbonate rocks is hosted in a diorite/granodioritic to quartz-monzonitic stock. Copper mineralization is esentialy associated with potassic and less sericitic alteration. Based on the fluid inclusion and isotopic data from oxygen and sulfer, show that main copper mineralization at Sungun occurred in the temperature range of 300-400oC (with the mode of 360oC ).These data also indicate that the hydrothermal system involved both magmatic and meteoric waters, and boiled extensively. In this research, thermodynamic data are used to delineate the stability fields of alteration and ore assemblages as afunction of fS2 fO2 aCl- and pH (in the fixed temperature of 3600C and the pressure of 500 bars) for the chalcopyrite dissolution and deposition. During potassic alteration (>450oC)copper solubility is calculated to have been >50,000ppm, whereas the copper content of the initial fluid responsible for ore deposition is estimated from fluid inclusion data to have been 1200-3800ppm. This indicates that the fluid was initially undersaturated with respect to chalcopyrite, which agrees with the observation that veins formed at T >400oC and contain molybdenite but rarely chalcopyrite. Copper solubility drops rapidly with :decreasing temperature, and at 400oC is approximately 1000 ppm, within the range estimated from fluid inclusion data. At temperatures less than 350 or 360oC ,the solubility drops to < 25 ppm. These calculations are consistent with observations that the bulk of the chalcopyrite deposited at Sungun is hosted by veins formed at approximate temperatures of 360±60oC. Other factors that may reduce chalcopyrite solubility are increases in pH, and decreases in fO2 and aCl-. Based on the calculations done in this paper, it is proposed that the Sungun deposit formed partly in response to the sharp temperature decrease and pH increase that accompanied boiling, and partly as a result of the additional heat loss, increase in pH and decrease in aCl which occurred as a result of mixing of acidic Cu bearing magmatic waters with cooler meteoric waters of lower salinity and near-neutral pH.

This paper introduces a new algorithm for shape-based image indexing and retrieval. In order to generating index of images, this algorithm considers a window around of each kernel edge and uses correlation information between edges in window. Hence it includes information of continuous edges and lines of images and describes major shape properties of images. This scheme is effective and robustly tolerates translation, scaling, color, illumination, and viewing position variations. Experimental results show superiority of proposed scheme over several other indexing methods. Averages of precision and recall rates of this new indexing scheme for retrieval as compared with traditional color histogram are 1.99 and 1.59 times, respectively. These ratios are 1.26 and 1.04 compared to edge direction histogram. Whereas this algorithm is a one-pass process algorithm it is also efficient.