AMIRKABIR
Year:2001 | Volume:12 | Issue:46|Papers Count:14

An instability analysis in a horizontal porous layer is made for a fluid with an inverse density gradient. The governing equations resulted when instability analysis is applied to this problem, are non-linear and therefore mathematically complex. In published literature, the problem is solved with some simplifications such as ignoring certain terms in the governing equations, or finding algebraic approximations that are valid in some specific range of physical parameters. In this study, the problem is solved in its general form for two rigid, and one rigid and one free boundary conditions using variational methods. The resulting critical Rayleigh numbers and critical wave numbers vs. l2/K curves are compared to previous works.        

Compilation of optimum model for determining the extent of economic activities and the allocation of limited resource to them, regarding to reach the different goals in economic planning, is the principle research subject in this paper. In this context, it seems relevant to utilize linear programming based on Leontief's input-output system. In this model compatibility and optimality of economic activities have been taken into account simultaneously. So, compared to Leontiefs input-output model and other models like statistical model, the said optimum model, due to its more focus on planning principles would bring about more useful results. However, it would be desirable and indeed more realistic to take into account all of the goals influencing the system. Regarding these purposes, in this paper, a multiple objective optimum model based on Leontiefs input-output system has been prepared and then it has been used to economic planning in Iran for the year 1378 (1999). In the end, to evaluate the model validity, the results of the said optimum model have been compared with the results of current model being used for second economic plan for the year 1378 (1999).        

One of the unsolved problems in multi-layer neural networks is the problem of determination of optimal topology a topology with minimum number of hid den units. There is no method to determine the optimal topology for multi-layer neural networks for a given problem. Usually the designer selects a topology for the network and then trains it. The selected topology remains fixed during the training period. Since the problem of determination of optimal topology for neural networks belongs to class of NP-Hard problems, most of the existing algorithms for determination of the topology are approximate algorithms. These algorithms could be classified into four main groups pruning algorithms, constructive algorithms, hybrid algorithms, and evolutionary algorithms. These algorithms can produce near optimal solutions. Most of these algorithms use hill-climbing method and may stick at local minima. In this paper a learning automata based algorithm called survival algorithm, for determination of the number of hidden units of three layers neural networks is proposed. The proposed algorithm uses learning automaton as a global search method in order to increase the probability of escaping from local minima and hence to increase the probability of obtaining the optimal topology. In survival algorithm, the training begins with a large network, and then by adding and deleting hidden units the optimal topology is obtained.        

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 con servation 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. Further, with the proposed model, the performance and comparison of fluidized and packed bed dryer can be shown.      

The trip demand information in the form of the origin-destination (OD) matrix is an essential input to the transportation studies. Obtaining such a matrix by conventional methods needs considerable amount of time, cost, and manpower. Since obtaining traffic volume data for the links of the transportation net work is easy, recently many researchers have tried to obtain trip demand matrix by using this information. Among them, Spiess introduced the gradient method which is applicable for real large scale networks. The gradient method is an iterative procedure which starts from an initial O-D matrix and at each iteration tries to reach a matrix which generates flows closer to the observed flows and while staying close to the initial matrix as much as possible. In this paper, the gradient method has been implemented based on tow type of traffic assignment procedures, namely Frank-wolfe and Aashtiani algorithms (which both of them are based on traffic equilibrium), in EMME /2 software and C programming environments respectively. The resulting algorithms have applied to real network (city of Mashhad). In the implementation of the gradient method using complementarity assignment, a method has been introduced for increasing the convergence speed. The computational results for the case under study show that using the complementarity algorithm in the gradient method makes it about 34 times faster as compared with the Frank-wolfe assignment in EMME /2 environment. Also for keeping predicted (corrected) O-D matrix as close as possible to the initial matrix, an improved gradient method has been introduced. For the network of the city of Mashhad, the improved method predicted a matrix which was very close to the initial matrix while generating flows in the links very close to the respective observed ones.    

Citric acid was produced from cane molasses treated chemically with tricalcium phosphate, by Aspergillus niger which was entrapped in calcium alginate beads. The citric acid fermentation was carried out in shake flasks and in order to get comparable results some other treatments were also conducted treated and untreated-canem olasses as the substrate for the fermentation and using A.niger in the form of free cells, untreated cane molasses and using A. niger in the form of immobilized cells. Time coursers) of three main operational variables in the citric acid fermentation was determined: citric acid produced, sugar utilized, sugar available (residual sugar) (both of these sugars were calculated as reducing sugar). Moreover, at the end of fermentation period (14 days) biomass yield was also determined (amount of biomass produced dry weigh based on amount of sugar utilized). With approach taken for comparing the trend of changes in concentration of the main fermentation parameters over time, results were discussed in detail and it was concluded that production of citric acid from treated cane molasses with using the immobilized cells was the most suitable conditions for the fermentation among other treatments. Citric acid concentration was 7.76 g / L (average value) and its yield based on amount of sugar utilized was 10.3%.  In studying fermentation processes, it is desirable to use a mathematical model to present the kinetic behavior of product formation. Using such a mathematical formula to present the results could make some simplicity in studying the mechanism of the fermentation. In order to provide a rate expression for fermentation in most cases, however, one uses data from the plot of changes in substrate concentration versus time. In this research trend of decrease in substrate concentration over time (residual sugar in fermentation medium) showed to have a typical pattern, on the other hand changes in citric acid concentration  during fermentation time had some fluctuation. The trend of increasing in sugar utilized during fermentation showed an exponential form and could be treated as first-order reaction. Rate of increasing in sugar utilization was taken as index for citric acid production along the biomass formation. First order rate constant was calculated as 3.452 × 10-3 hr-1.    

In this paper, a method for analyzing the human knowledge is introduced. Using this method, the knowledge engineer can determine the appropriate basic knowledge representation model (the Semantic Net, the Logic, the Frame, and the If-Then rules) to be used in the intelligent system. The analysis begins with representing the domain knowledge with the Intermediate Knowledge Representation Model (IKRM) which is described here in detail. The result is a graph with different edges. It has been shown that there is an algorithm that could make man subgraphs from the original one. And each subgraph best fits to one of the basic knowledge representation models. The way each subgraph should be transformed to basic knowledge representation (KR) model is shown by definition of another algorithm and examples. This helps in determining the omitted knowledge elements when using one of the basic KR models. And defining the required KR models in the hybrid system.      

A method for the analysis of inverse nonlinear heat conduction having point or line distributed heat sources is presented. The contribution due to the loading vector coming from either point or line heat sources integrals are evaluated analytically and in an exact manner. The inverse solution may be categorized into finding either the solution of the problem in the form of the intensity of the loading at a specific location or finding the location or orientation of the generators assuming the intensity is known. The first category is an ill-posed type problem, whereas the solution to the second category type problems may be found by an optimization procedure. To regularize the solution to the first category type, least squares method is employed in conjunction with an addition of regularization term. To find the solution to the second category a new algorithm to be called "A Good Neighboring" is devised which make use of derivatives of order zero. For the problems with thermal conductivity temperature dependent a nonlinear behaviour is expected which would be dealt with Kirchhoff's transformation. The efficiency and accuracy of the methods are explored through several examples.    

In this paper, stress concentration around a hole is examined in a hybrid composite lamina. Using shear-lag theory, stress concentration factor is calculated and it is shown that its magnitude decreases considerably with an in crease in holes larger to smaller diameters ratio. Compared to a crack, the amount of decrease for an elliptic hole (a / b=2), with eleven broken fibers is 42.5%, if all the fibers are assumed to be of the same type. A decrease in diameters ratio, causes the magnitude of stress concentrations to approach that of a crack. The same behavior is observed for maximum shear stress. According to the results, hybridization has a unique effect on a decrease in both HM fiber stresses and maximum shear stress in the lamina. For N = 31, the decrease in maximum shear stress is 94% at r =11 and a / b = 1.      

In this research, we have studied the static features of three main Farsi vowels (/a/, i /, / u /), uttered by four prelingual Farsi-speaking cochlear implant (CI) children (7-13 years old). For this reason, we have recorded five meaningful sentences uttered by the patients before the operation and after three, six and nine months post-operation. At each session, patients read the sentences once in device-on (PN) condition and then after a 30 minutes stay, in a device-off (PF)  condition. To be able to conduct a more objective study (rather than subjective evaluations conducted by other researchers), we extracted features such as fundamental frequency, the first two formants frequencies and the vowel location in the vowel triangle. In addition, we have introduced two new features called Relative Energy (RD, the ratio of the vowel energy to the energy of its preceding consonant), and Relative Duration (RD, the ratio of the vowel duration to the word, duration). We have shown that the two new features are consistent and valid for normal speakers, and thus can be employed as measures of speaker's control on vowel production, voicing and duration. Quantitative results show that :1) Almost all of the static features considered have consistently improved in time after the operation. It is particularly interesting to notice that due to the adverse quality of the uttered speech, subjective evaluations could hardly reach to a conclusion. 2) At least for the static features under study, the patient's reliance on the AF "decreased" consistently by time. While in the few first months after the operation loss of the AF could degrade the speech quality to a noticeable extent, after nine months, their dependency on the AF decreased considerably. It implies that the speech production motor patterns of the patients have been trained in time.  

Carotenoids are the most important and the largest pigments in nature, but only the plants and microorganisms are able to produce them in the nature. These components, due to having suitable coloring property, relative stability and also antioxidant activity, have different uses in Food Industries. These components are the most important precursors of vitamin A in the body, which their presence in the consuming foods is necessary. One of the most important microorganisms which can produce carotenoids especially beta-carotene is the yeast named Rhodotorula. Most species of this yeast have the ability of producing carotenoids pigments in the considerable amounts. For isolaion of Rhodotorula, a number of foods and media were studied. At the end, it was isolated from decayed lime, then after microbial evaluations, it was identified as Rhodotorula muciligenous (rubra). In continuing, the amount of carotenoids especially beta- arotene which were obtained from the indigenous species was compared with other species purchased from collection centers in U.S. and China. The results indicated that produced beta-carotene on dry weight biomass in the yeasts, Rh. glutinis (C499) and Rh. muciliginous  indigenous spp.) were 447 and 285 micg /g respectively.        

This paper presents a new method to estimate vehicle velocity and tire-road friction forces using an extended Kalman filter (EKF). A stochastic dynamic model is considered for tire-road friction forces and the vehicle motion equations are augmented to include friction force models. The state vector of this system consists of the vehicle longitudinal and lateral velocities, yaw rate, wheel angular velocities and friction forces. A square root algorithm is used for EKF implementation. The main features of this method are numerical stability and computational efficiency. The effectiveness of the proposed method is demonstrated by simulation.