|
|
|||||||
|
|
|||||||
|
|
|||||||
|
|
|
|
|
|
|
|
|
|
SPONSORED PROJECT LIST Effective October, 1998 Proposal No. Name Title Oct.
98-01 Susan L. Albin Control of the
Startup Period in Batch Processing DESCRIPTION: The startup periods of batch processes often consist of numerous cycles of setting process variables, sampling product, off-line testing, and process adjustment involving dozens of correlated process and product variables. Our previous work identified a startup-monitoring methodology, based on projection to latent structures, to determine whether a configuration of process variables observed during startup is equivalent to regular production and, if not, which variables need adjustment. Here we develop a startup-control algorithm that recommends the specific new levels for process settings. The significant success of the monitoring methodology when applied to company data strongly suggests the potential value of the proposed combined monitoring and control algorithm.
PROJECT NO.: Oct. 98-02 DESCRIPTION: Products are normally designed for use at specific stress levels. However, many products endure multiple levels of stress during intermittent periods of use, transportation, and storage. Failure mechanisms may differ for a new product as compared to a product that has aged. Corrosion or seal permeability problems may affect electromechanical components over time, for instance. The same is true if the usage environment changes due to force, vibration, temperature, etc. This project develops and tests a set of models to predict the effects of materials aging on component and systems reliability when products are subjected to varying stresses over time. We begin with the case of discrete usage modes and later consider the case of random dynamic stress.
PROJECT NO.: Oct. 98-03 DESCRIPTION: This project involves the development and automation of planning and assessment tools for repairable system reliability with non-constant rate of occurrence of failure. The distribution of failure times is often non-stationary for many products during (1) development testing where there is reliability growth, and (2) extended field usage where there is deterioration. The problem is further compounded because a system is designed with many subsystems and components, each with potentially different failure patterns. This project involves the study and characterization of different reliability figures of merit and planning methods. The project also explicitly considers reliability estimation uncertainty.
PROJECT NO.: Oct. 98-04 PROJECT MANAGER: E.A. Elsayed and Jian Zhao DESCRIPTION: The first phase of this project dealt with the
development of a general reliability model for accelerated life testing.
PROJECT NO.: Oct. 98-05 DESCRIPTION: Many companies have set Parts per Million (PPM) or Parts per Million Opportunities (PPMO) goals in their quest for continuous improvement. Yet, they have no means for assessing low defect levels and determining when improvements have been made. At low PPM levels, sampling is of no benefit, since the sample size necessary to observe a single defect would be prohibitively large. 100 percent inspection is usually not possible and error rates can exceed the PPM level. Thus, there is need for a scheme which can perform the monitoring process without sampling or 100 percent inspection. The use of the Time-Between-Events (TBE) CUSUM is proposed for this project. It has been demonstrated to work well using theory and computer simulations. This proposal suggests examining implementation using a member company.
PROJECT NO.: Oct. 98-06 DESCRIPTION: A logical framework for the development and improvement of industrial processes is the three-phase approach identified as characterization, control, and optimization. Characterization is the process of discovering the specific variables responsible for the variability in the system. Control refers to process stability, and optimization refers to manipulating the most important process variables to obtain the best set of operating conditions for the system. There are three specific research tasks addressed in this project. Two are direct outgrowths of projects performed last year. All research tasks support characterization, control, and optimization of processes.
PROJECT NO.: Oct. 98-07 DESCRIPTION: Many manufacturers produce items that go through a series of processing steps. Following each step, the items are tested or inspected for certain quality characteristics. This testing may be done by automatically testing or measuring one or more characteristic on every unit produced. Alternatively, at each stage only a fraction of the units produced may be tested, where the units tested are selected according to some sampling procedure. This project investigates the performance of multiple sequential inspection and testing systems. Specifically, it is of interest to study the operating characteristics of these systems so that their performance in terms of accepting defective units and rejecting good units may be determined.
PROJECT NO.: Oct. 98-08 DESCRIPTION: IC runs generate data with at least two important characteristics: (1) there are numerous independent, or recipe, variables measured that can be used for predictions, and (2) measurements are taken at a number of sites on each wafer and from a number of wafers at different locations in a furnace. An effective algorithm for process control for this type of data is proposed. The strategy includes some linkage between tools and process steps in a hierarchical control strategy. Simple run means and standard deviations do not capture all of the available information in the data, nor do they adjust for the recipe variables.
PROJECT NO.: Oct. 98-09 DESCRIPTION: Many software reliability models have been proposed to determine whether the target software is ready to be released. However, their usefulness in practice is limited by uncertainties in the underlying models. A drawback of current efforts, such as software reliability modeling, is that they treat software as an independent entity (e.g., ignore hardware platforms) and focus on the number of residual faults in software. While this may be the case for some software, it is clearly not the case for software in embedded applications such as VLSI, transmission controller, electronic control module, which is significantly affected by the environment (e.g., hardware) in which it runs. Given that software runs on a hardware platform and is a component of the overall system, interactions between hardware and software must be taken into account, and the reliability of software must be judged by its contribution to overall system reliability.
|
QRE
Main | About
ORE Cenetr | Faculty | Completed
Projects | Current Projects |