Multicriterion Optimization: A Tool in Advanced Materials Technology

This paper will demonstrate on two advanced materials — a fibre-reinforced composite laminate (FRC) and a transformation toughened ceramic (TTC) — the importance of multicriterion optimization in the production of useful advanced materials with enhanced mechanical properties. In a previous paper (Thomsen et al., 1994a), the authors have demonstrated the application of single-criterion optimization to these materials which are based on a brittle matrix and thus prone to cracking at very low applied stresses. The optimization process aims at altering their microstructure so that all their desirable mechanical properties are enhanced. Currently, the advanced materials technologists must take a heuristic approach to meeting the often competing requirements. The present paper will show how multicriterion optimization can come to the aid of the technologists and reduce their reliance on empirical approaches.

Artificial Neural Network and the Taguchi Method Application for Robust Wheatstone Bridge Design

The primary function of the Wheatstone bridge is to measure an unknown resistance. The elements of this well-known measurement circuit will take on different values depending upon the range and accuracy required for a particular application. The Taguchi approach to parameter design is used to select values for the measurement circuit elements so as to reduce measurement error. Next we introduce the use of an artificial neural network to extrapolate limited experimental results to predict system response over a wide range of applications. This approach can be employed for on-line quality control of the manufacture of such device.

Design System for Blow Moulded Products

In recent years the spectrum of applications for thermoplastics has experienced constant growth. They do not only exhibit good specific materials properties and lend themselves to costeffective processing, but it is also possible to manufacture products of great complexity. Numerous possibilities for structural configuration are thus presented, enabling product material and form to be adapted to specific requirements. Computer-aided simulation techniques are among the tools utilized in developing optimal components or products. This led to a design system which can accelerate the entire development process of a product or component. It starts from the idea and results in the production stage making use of multidisciplinary combination of C-technologies, structural analysis procedures, material data bases, production, inspection and testing. The propagated concept is significantly furthered by the rapid development in the areas of hardware and software. A comparison of a traditional development process with the development process utilizing the design system illustrates the possibilities which exist to exploit simulation thechniques in every stage of product development in order to enhance quality and reduce costs.

Modeling of a Leg System to Illustrate the Feasibility of Energy Recovery in Walking Machines

A key consideration in the design of walking machines is the minimization of energy consumption. Two main avenues of research have been pursued in the past (a) finding an optimal gait which reduces energy consumption or (b) the employment of energy storage devices to recover energy from one step to the next. This study follows the latter approach, which has hitherto concentrated on hopping machines. Several practical design considerations for energy recovery in multi-legged walking machines, where a stance phase prevents the immediate recovery of energy from one step to the next, are investigated. Two schemes, passive and active locking, are introduced. The simplified models presented serve to illustrate the feasibility of these schemes.

Structural Shape Optimization Wth Error Control

A robust and automatic shape optimization procedure is presented in this paper, which incorporates recent developments in the field of computer-aided design (CAD) of mechanical structures, such as geometric modelling, automatic selection of independent design variables, sensitivity analysis using reliable mesh perturbation schemes, error estimation and adaptive mesh refinement. A numerical example is given to show the efficiency of the procedure.

Robust Parameter Design of the Precision Injection Molding Process

The most important problem, causing defective parts, in the injection molding process, is nonuniform shrinkage of molded parts. This leads to an iterative trial-and-error cycles of modification of mold cavity and core to arrive at the right dimensional size required which can occasionally to complete retooling. For this process, there are many factors that can be thrown out of control. Using the traditional scientific approach, engineers have longed to understand the mechanics of the process to control it, with limited success. In this paper, a design of experiments setup, using the Taguchi Methods, was done to reduce the nonuniform shrinkage. The company where the experiment was carried out is a precision parts molder for their own product lines. By using the internal experts from the company, a list of independent process parameters with no interactions which were thought the most responsible for dimensional size were listed. As there were 13 such parameters, it was decided to use the L27 orthogonal array. The optimum value that the company experts thought would produce the right part were used as the settings for the initial experiment. The 27 experiments were then performed, allowing sufficient time to let the machine stabilized between the experiments. The S/N ratio calculation for 27 experiments was explained. Next the calculations for the percentage that each parameter contributes to the dimension was determined. Finally, a confirmation experiment was performed to verify the results.

Optimization of Stirling Engine Performance Based on Multipoint Approximation Technique

The ideal Stirling working cycle has the maximum obtainable efficiency defined by Carnot efficiency, and highly efficient Stirling engines can therefore be built, if designed properly. To analyse the power output and the efficiency of a Stirling engine, numerical simulation programs (NSP) have been developed, which solve the thermodynamic equations. In order to find optimum values of design variables, numerical optimization techniques can be used (Bartczak and Carlsen, 1991). To describe the engine realistically, it is necessary to consider several tens of design variables. As even a single call for NSP requires considerable computing time, it would be too time consuming to use conventional optimization techniques, which require a very large number of calls for NSP. Furthermore, objective and constraint functions of the optimization problem present some level of noise, i.e. can only be estimated with a finite accuracy. To cope with these problems, the multipoint explicit approximation technique is used.

Derivation of Multiple Assembly Sequences From Exploded Views

The original exploded view method of assembly planning (Mohammad and Kroll, 1993b) attempts to automatically generate the “simplest” assembly plan for a product. Criteria for optimal assembly sequences, however, are sometimes unclear and dependent on specific circumstances. It may therefore be advantageous to present the designer with more than one solution to the assembly planning problem, or allow him or her to select the appropriate criteria. This paper introduces an extension to the original method, where all possible assembly sequences are generated. This completely automatic capability is demonstrated by manually applying the method to a simple example and then comparing the results to those generated by the algorithmic procedure of the C. S. Draper Lab.

A Simulated Annealing-Based Approach to Three Dimensional Component Packing

This paper introduces a computational approach to three dimensional component layout that employs simulated annealing to generate optimal solutions. Simulated annealing has been used extensively for two dimensional layout of VLSI circuits; this research extends techniques developed for two dimensional layout optimization to three dimensional problems which are more representative of mechanical engineering applications. In many of these applications, miniaturization trends increase the need to achieve higher packing density and fit components into smaller containers. This research addresses the three dimensional packing problem, which is a subset of the general component layout problem, as a framework in which to solve general layout problems.

Composite Laminate Stiffnesses and Their Sensitivities

The great number of possible stacking orders to form laminates suggests to apply optimization, more frequently than usual, in the design of structures made of composite materials. One of the columns upon which optimization of structures is built is the mathematical search procedure for locating a minimum (or maximum) of a constrained function. Efficient algorithms will require the evaluation of derivatives of the object function as well as of the constraints. In that context the sensitivities of laminate stiffness matrices may be required. In order to meet such a requirement the derivatives with respect to both ply thicknesses and ply angles, of laminate stiffnesses, including transverse shear stiffness, will be presented in this report.