S. Wilayat Husain

Image Analysis of Polymer-Dispersed Liquid Crystals

The quantitative analysis of the Polymer-Dispersed Liquid Crystals is generally difficult because the liquid crystal droplets do not appear as complete circles in the optical micrographs. In the present work we have devised a procedure to complete the circular images of the liquid crystals in an acquired image. In this method, the longest chord of an incomplete circle is found by analyzing the edge pixels of the droplet. This chord is then taken as the diameter, and a complete circle is drawn around this diameter. The procedure is repeated for all the droplets. Once all the circles have been completed, standard image analysis is carried out to compute the geometrical features of the image. The present method can also treat a difficult to handle problem, i.e., the appearance of broken circles, which is a common feature. For this purpose, some manual video editing has been suggested, whereby the broken pieces that appear to be the part of a given droplet are connected to each other.

Design of spiral grooves on a spherical bearing

Abstract In the present paper we formulate a set of equations for logarithmic spiral grooves on a spherical bearing. The grooves satisfy the condition that the angle between the velocity vector and the tangent to the groove remains constant. A procedure is briefly described to make grooves on a spherical ball. The results are then compared with those obtained using a simple equation for spiral grooves. It is found that the shape of the grooves and the axial load capacity of the bearings are not significantly different in the two cases.

The design of a flat negative pattern to make grooves on a spherical spiral groove bearing

Abstract In the present work we describe a procedure to design a Flat Negative Pattern which can be used to photo-etch the spiral grooves on a spherical thrust bearing. The design of the Flat Negative is accomplished in such a way that only one pattern is used for successive exposures resulting in perfect grooves. A groove is taken to be the space bounded by two logarithmic spiral curves separated by a distance equal to the desired width of the groove. The desired Flat pattern is obtained through normal projection of these curves on a plane parallel to the Negative. The projection is accomplished by using three-dimensional (3D) geometric transformations of a large number of points taken on the curves.

Characterization of dendritic structure using fractals

We have utilized fractals to characterize the dendritic structure in the cast alloys. In the present paper we apply this idea on the directional dendrites generated mathematically. For this, a computer program has been developed which can generate dendrites with varying arm spacings. Any number of levels such as primary, secondary, tertiary etc. can be generated. It is found that the ‘Fractal Dimension’ increases linearly on logarithmic scale as the dendrite arms spacing decreases. Traditionally, dendrite arm spacing has been correlated with cooling rate during solidification. Our work suggests that the cooling rate, in turn, may be related to the fractal dimension of the dendritic structure. The advantage of fractals over measurement of dendritic arm spacing is that fractal dimension is obtained as an average number representative of the given dendritic structure. On the other hand, dendritic arm spacing measurement, in a given case, is limited to one level only i.e. either primary, secondary or tertiary dendrite arms. In addition, the arm spacing may vary locally and the estimation of an average value may be difficult. For the real cases, the method can easily be implemented on an image analysis system.

Phase transformations in maraging steels using ultrasonic velocity measurements

Phase transformations in maraging steels grade 250, 300 and 350 aged at various temperatures have been studied using ultrasonic velocity measurements. It was found that the temperatures corresponding to various transformations can be easily monitored by observing the changes in ultrasonic longitudinal velocity. In the case of 350 maraging steel, for example, the precipitation starts at about 360° and the austenite begins to form around 560°. Maximum amount of austenite is observed in samples aged at 670°. At aging temperatures beyond 770°, the velocity approaches values corresponding to the annealed samples signifying that the process of precipitate dissolution has been completed. These results are consistent with the results obtained by resistivity measurements and x- ray diffraction studies.

A numerical approach to solve the inverse heat conduction problem

In the present work we have utilized a newly developed numerical approach to solve the inverse heat conduction problem. In our method we start with a guess of surface flux history and the direct conduction problem is solved numerically. This generates the temperature history at the location at which the experimental history is known. The calculated history is then compared with the experimental data. For this we find the area under the two curves. An iterative numerical procedure is utilized to improve the guessed surface heat flux. When the area difference is less than a given tolerance the problem has been solved. The method has been implemented both as a ‘whole domain method’ as well as a ‘sequential method’. In this paper we present results for ‘whole domain’ case only. We have used our method to solve various test cases. The results for one such case are presented and the performance of the method is compared with the method of least squares.

Use of Laplace transformation to determine corrosion parameters from electrochemical data

Activation controlled processes in electrochemical corrosion are described by an exponential type of relation between the applied over potential and the measured current density. This relation involves three unknown parameters i. e. corrosion current and the two Tafel constants. Due to the exponential nature of this relation the determination of these unknown parameters is difficult. We have used Laplace transformation for this purpose. Laplace transformation is used to separate the original equation into two parts; first having one unknown parameter and the second involving the remaining two parameters. Extensive numerical testing using exact hypothetical data has shown that the method converges rapidly providing correct values of the parameters. Keeping in mind that the real data generally involves experimental inaccuracies, the method has been tested for data having random error. The performance of our method has been compared with other existing methods and it is found that our method generally provides more accurate results. In fact our method can easily handle situations where other methods fail to provide any results.