Minimal Material Waste Research Articles

An Approach to Selection of Material and Manufacturing Processes for Rocket Motor Cases Using Weighted Performance Index

Material selection is a very critical design decision, which has a profound influence on the entire development program for rocket motor cases. In the selection process, the main performance parameters and the most appropriate fabrication technology with proven processes must be considered. Many years of practical experience in material selection process with a thorough understanding of materials behavior under various loading environments and hands-on experiences with various available manufacturing processes are of immense help to the design and development engineer for successful completion of the development program. In this paper, an attempt has been made to present an approach for selecting appropriate material and manufacturing process for rocket motor case based on method of Weighted Performance Index (WPI) with the hope that this approach will also provide additional aid to the design engineer for the selection of material and manufacturing process for rocket motor cases. In this method, different properties are assigned a certain weight depending upon their importance to the service requirements. Different properties are normalized using a scaling factor, and finally a weighted property index is computed. The material that scored the maximum numerical value is chosen as the material for fabrication. This approach closely matches with the actual performance. Maraging steel and D6AC are found to be the preferred materials for rocket motor cases for critical missions. HSLA steels are appropriate for less-critical applications, in which rocket motor cases are required in very large numbers (e.g., flow-formed AISI 4130 motor cases[8]). For the selection of an appropriate manufacturing method, the major parameters considered are dimensional accuracy, cost of production, minimum material waste, and flexibility in design. Again, these properties are given a relative grading, which is then converted into a scaled property. Finally, the weighted performance indices are estimated. The flow-forming method has emerged as the manufacturing method of choice for motor tubes.

An automated meniscus coating system for polymer deposition on large-area MCM-D and MCM-L substrates

Material deposition techniques can be a significant contributor to the overall electronics manufacturing costs. The present study evaluates meniscus coating as a low-cost tool for large-area polymer deposition. Meniscus coating has the following advantages over conventional spin coating: (1) minimal material waste; (2) higher throughput; (3) higher planarity; (4) minimal defect density; (5) thickness uniformity over a large area. Each of these factors help reduce the overall manufacturing costs. The objective of this effort is to develop, install, and qualify an automated, low-cost, high-throughput polymer deposition method for large area MCM-D and MCM-L substrates. A variety of polymer photoresists, dielectrics, and composites are used for thick film coating on 12 in PWB and glass substrates. For each material, the deposition process is optimized to the desired film thickness. In this paper, we report materials and process optimization for 12 in/spl times/12 in substrates that is scalable to 24 in/spl times/24 in substrates. For higher throughput and lower manufacturing cost, an automated coating workcell is designed. The proposed workcell includes a robotic material handling and convection drying of coated films with automated loading/unloading capability. Design and fabrication of the integrated workcell are also addressed.

Double-Drift Thin-Window Planar Ge(Li) Detectors

An improved method for fabricating a thin-window planar detector was developed at Lawrence Radiation Laboratory, Livermore. This detector is designed to analyze spectra of low-energy as well as high-energy gamma and x rays. The detector has a planar configuration with a relatively large gold surface barrier entry window. The main advantage of this detector is its faster drifting time coupled with minimum waste of material. This method of fabrication has significant advantages over earlier techniques.

The study of an algorithm for the recognition of typewritten characters

In this article we describe the work on an algorithm to recognize standard characters with a very high degree of reliability [1]. The algorithm is based on the assumption that all symbols of one definite type-face, that of a typewriter, for example, are certain idealised, standard characters subject to distortion: the picture of a character can differ from its standard by a shift, uniform darkening, or change in contrast. In addition, noise is superimposed on the picture, i.e. individual dots which are independent of the others can be darker or lighter than the corresponding dots of the standard. It is assumed that all these distortions can be described by simple statistical laws. The recognition algorithm is as follows. The coefficients of correlation between each of the standard characters and the unknown character are calculated for all possible shifts within a limited region. The unknown character is identified with the standard character to which the highest correlation coefficient corresponds. For simplicity, we do not calculate the correlation coefficient itself, but the scalar product of the non-normalised vector describing the unknown character and the normalised vector of the standard character, to which it is proportional. The vectors of the standard characters and their normalisation are found beforehand. The algorithm thus consists of two parts: the “preparation”, during which the normalised vectors of the standard characters are calculated, and the recognition itself. Since the systems for automatic pattern recognition are highly complex, particularly for letters and numbers, it is convenient to carry out the examination by simulating these systems on a universal computer. The simulation method enables the efficiency and reliability of any recognition algorithm to be checked with the minimum waste of material and the maximum economy in time and labour. A special attachment has been developed at the Cybernetics Institute of the UkSSR Academy of Sciences for the universal computer Kiev. This is a universal picture transformer [2]. Its purpose is to simulate any algorithm concerned with the processing of information contained in picture form, and it works according to the following scheme. The coordinates of the centre of the small element of picture under examination at a given moment arrive in binary form from the computer at a transformer of numbers into voltage. The resulting voltages act on the deflection plates of a cathode-ray tube, the screen of which is projected by an objective on to the given picture. The light flow reflected from the element falls on to a photomultiplier. The brightness of the light spot is not homogeneous, but varies in proportion to the distance from the centre of the spot. Therefore, in obtaining information about the reflection coefficient of the element a weighted average of the local reflection coefficients is found. The current in the photomultiplier which is proportional to the mean reflection coefficient is transformed into binary code and returns to the computer. The diameter of the light spot, and therefore also the size of the scanning elements, can vary within certain limits. The scanning sequence of the picture is determined by the program according to which the machine calculates the coordinates of the points which are the centres of the elements. The picture can be broken down into 256 × 256 = 65,536 elements. Thus, complete information about a picture can be considered as the array of numbers ϱ [ x, y] in which the indices x and y can go from 1 to 256. The transformer can distinguish between 16 levels of darkening. A white field corresponds to the greatest value of ϱ, equal to unity. There is an additional controlling cathode-ray tube for the visual observation of the motion of the light spot over the picture. Below we examine a program designed to distinguish ten typewritten numbers and the results of experiments are given. The program is defined in algol-60 [3]. The program structure is explained in sufficient detail in the comments.