A two-step methodology for CMOS VLSI reliability improvement: Step one

Reliability improvement of CMOS VLSI circuits depends on a thorough understanding of the technology, failure mechanisms, and resulting failure modes involved. Failure analysis has identified open circuits, short circuits and MOSFET degradations as the prominent failure modes. Classical methods of fault simulation and test generation are based on the gate level stuck-at fault model. This model has proved inadequate to model all realistic CMOS failure modes. An approach, which will complement available VLSI design packages, to aid reliability improvement and assurance of CMOS VLSI is outlined. A ‘two-step’ methodology is adopted. Step one, described in this paper, involves accurate circuit level fault simulation of CMOS cells used in a hierarchical design process. The simulation is achieved using SPICE and pre-SPICE insertion of faults (PSIF). PSIF is an additional module to SPICE that has been developed and is outlined in detail. Failure modes effects analysis (FMEA) is executed on the SPICE results and FMEA tables are generated. The second step of the methodology uses the FMEA tables to produce a knowledge base. Step two is essential when reliability studies of larger and VLSI circuits are required and will be the subject of a future paper. The knowledge base has the potential to generate fault trees, fault simulate and fault diagnose automatically.     

Failure Prevention in Design Through Effective Catalogue Utilization of Historical Failure Events

The science of failure prevention relies heavily on the experience of personnel on a project. As the nation is about to face a tremendous decline in the experienced workforce due to the baby boomer generation’s retirement, it is critical to begin focusing on capturing their knowledge. Cataloging and communicating the knowledge of potential failures is critical to prevent engineering disasters. Many companies have adopted failure-reporting systems that allow them to record their engineering failures to promote failure prevention. While recording this information is vital to learning from past mistakes, often the information is not stored so that engineers and designers can easily recall this valuable linguistic information and use it to improve designs. Therefore, more effective systems for cataloging and utilizing corporate memory of recorded failure events are needed. This article presents the design of a computational linguistic database to support the failure prevention tool, the risk in early design (RED) method. RED promotes failure prevention by identifying failure risks as early as the conceptual phase of product design, where impacts of failure prevention are greatest. It uses a database populated by historical failure event information to present specific areas that are at risk of failure in a product.     

A scheme for functional reasoning in conceptual design

An ideal functional reasoning environment should support designs of any nature, routine or innovative, at any level of detail, as well as through varying levels of detail. In this paper, three existing functional reasoning models are reviewed in this perspective. It has been found that none of these models support all of these requirements. It has been shown that a functional reasoning approach cannot guarantee the generation of solution concepts, which are combinations of known solutions, unless guided by the knowledge of existing solutions. A new model which can support design both across a level of detail and down through levels of detail has been proposed, which, using a divide and rule approach and using recursive problem redefinition while incorporating existing solutions, could support conceptual design. It is also shown that, although the generation of completely new solutions is not supported by the model, the model, when aided by a framework allowing a sustained progress of its knowledge base by transfer of knowledge from existing designs in the form of basic structures and rules of combination, could support generation of designs which otherwise would be considered unsupportable in a systematic way (innovative).     

Numerical analysis and experiment of composite sandwich T-joints subjected to pulling load

This paper presents an investigation into the failure mechanism and alternative design of composite sandwich T-joints subjected to pulling load. Based on a conventional design of sandwich T-joint as the baseline, numerical modeling and analysis using finite element (FE) method was performed to assess the strength against pulling load. The effect of a cutout in the web panel near the joint has been considered. To validate the models, sandwich T-joint samples were manufactured and tested. Detailed FE analysis and inspection of the experimental results indicated that the failure was mainly due to the excessive stress in the adhesive between the cleat flange and the T-joint base panel. The manufacture defects, which reduced the strength of the T-joint test samples had also been investigated. This has been further demonstrated by experimental results of repaired T-joint samples. A very good correlation between the test data and FE results were obtained. An unconventional design of T-joint for simpler manufacture process was proposed. Based on the design, T-joint samples were modeled, manufactured and tested to demonstrate the manufacture process and evaluate the improved strength.     

Failure analysis of seven masonry churches severely damaged during the 2012 Emilia-Romagna (Italy) earthquake: Non-linear dynamic analyses vs conventional static approaches

This study presents a detailed failure analysis of seven masonry churches, which were severely damaged during the 2012 Emilia-Romagna (Italy) seismic sequence, by means of 3D FE non-linear dynamic simulations. The main aims of the numerical investigations are: (1) to carry out a sufficiently wide sensitivity study on different specific case studies to have an insight into the role played by the geometry – which is always unique for churches – and by the irregularities; (2) to validate or address the limits of applicability of the most widespread conventional static approaches recommended by Italian Code, i.e. kinematic limit and pushover analyses. Non-linear dynamic analyses are carried out assuming that masonry behaves as a non-linear material exhibiting softening and damage, frictional behavior and different strength in tension and compression. The set of case studies investigated in this work shows how conventional static approaches are still capable of roughly identifying the most critical macro-elements that usually activate a failure mechanism, but that the results (e.g. collapse acceleration, behavior factor) are affected by a level of approximation that may considerably depend on in-plan irregularity and hypotheses done on the interlocking between contiguous walls. From the comparative analyses carried out, it is authors’ opinion that it is always beneficial to perform different types of analysis in order to have a comprehensive insight into the portions of the structure that can suffer a partial collapse with high probability. Once the active mechanism is identified, non-linear dynamic analyses, with more sophisticated material models and with refined FE discretizations of the critical regions, may be useful to deepen the knowledge of the behavior of such complex structures under seismic actions.     

Linking design process to customer satisfaction through virtual design of experiments

This work seeks to better understand how design processes affect design outcomes. Design process data were collected from journals kept as a part of mechanical engineering capstone design projects at Montana State University. Student processes were characterized by time coding journal entries using a 3 × 4 matrix of process variables. The data were modeled using a principal components artificial neural network, and the model used in a virtual designed experiment to obtain estimates for design process factors that significantly affect client satisfaction. Results indicate that greater client satisfaction is achieved through: greater problem definition (PD) activity and idea generation at conceptual design levels, and PD and engineering analysis activities at the system design level. Whereas, design activity at the detailed level associates with lower client satisfaction. These results support some aspects of existing models of “good” design process, and suggest adaptations of the models for novice designers.     

The reliability physics approach to failure prediction modelling

Empirically based failure rate modelling methodologies employed in reliability prediction handbooks, and deterministic modelling methods are both critically examined using microelectronic packages as vehicles. As an alternative, a coupled mechano-stochastic approach to reliability prediction modelling is presented. The goal is to use physics of failure principles with appropriate failure probability density distributions to design for failure-free operation and predict failure times for components now available, as well as new components resulting from new materials, technologies and processes. In addition, an approach for extending the model to aid in logistics support analysis is discussed.     

A probabilistic approach to uncertainty quantification with limited information

Many safety assessments depend upon models that rely on probabilistic characterizations about which there is incomplete knowledge. For example, a system model may depend upon the time to failure of a piece of equipment for which no failures have actually been observed. The analysts in this case are faced with the task of developing a failure model for the equipment in question, having very limited knowledge about either the correct form of the failure distribution or the statistical parameters that characterize the distribution. They may assume that the process conforms to a Weibull or log-normal distribution or that it can be characterized by a particular mean or variance, but those assumptions impart more knowledge to the analysis than is actually available. To address this challenge, we propose a method where random variables comprising equivalence classes constrained by the available information are approximated using polynomial chaos expansions (PCEs). The PCE approximations are based on rigorous mathematical concepts developed from functional analysis and measure theory. The method has been codified in a computational tool, AVOCET, and has been applied successfully to example problems. Results indicate that it should be applicable to a broad range of engineering problems that are characterized by both irreducible andreducible uncertainty.     

TREE-EXPERT: a tree-based expert system for fault tree construction

This paper addresses the topic of automatic fault tree construction, utilizing an expert system with Artificial Intelligence (AI) techniques and presents the related software tool, TREE-EXPERT—an expert system for automatic fault tree construction. In the light of the features involved in developing a fault tree, a new and more reasonable structure of knowledge representation, which is knowledge tree based, has been established. The knowledge tree provides the means by which component failure behaviors can be described by a group of particular fault tree modules instead of production rules. By introducing the conditional branch function, the new design of the knowledge base incorporates many good features such as strong expressivity, flexibility and ease of extension and it takes advantage of the user's familiarity with the field of fault tree analysis. Additionally, the design of the inference engine is original in that it deals with nodes, which it treats, as special components, so that many complicated engineering cases, such as the application of success criteria, and the problems of flow diversions and flow reversals in a process system, can be well managed and the function of the expert system is improved as a whole. TREE-EXPERT can be used to deal with large-scale and complicated engineering systems, and many engineering factors can be considered, e.g. more than one system parameter and the effect on them switching of the system operating states, bi-directional inference, human error failure, common-cause failure, maintenance and test, etc. On the other hand, the software uses P & ID (Pipe & Instrument Diagram) type interface to describe the system topology, which provides an easier man-machine interface with powerful graphics functions. This software can handle not only ‘process’ systems but also, with appropriate additions to the generic knowledge base, electrical systems and other similar systems.     

Uncertainty analysis of reliability predictions for brittle fracture

Reliability analysis of ceramic components under stationary or transient loading is generally performed on the basis of a Finite Element stress analysis from which the failure probability according to the multi-axial Weakest Link theory is calculated with the help of a suitable post-processing routine. We use the STAU post-processing routine and the general purpose Finite Element code ABAQUS. Due to scatter in the material parameters, the resulting failure probability is also prone to statistical uncertainties. We present a method of assessing this scatter using so-called resampling simulation methods. The analysis leads to confidence intervals for the failure probability which is a novel and important result especially for the purpose of design sensitivity considerations and the assessment of pooling procedures. In a simple example using a four-point bend specimen, the effect of pooling (i.e. grouping of results from different experiments by suitable scaling procedures) on the numerical result and on the scatter of failure probability is demonstrated. Here, pooling is done using results of inert strength measurements at various temperatures and scaling to room temperature values. A technologically more relevant example deals with a ceramic component in a model clutch under thermo-mechanical frictional loading. As a first step, the local risk of rupture is calculated which leads to the identification of the most critical regions of the component. As a second step, resampling confidence intervals for the failure probability are determined. As resampling data base, we use inert strength values at different temperatures as well as material data for sub-critical crack propagation.     

A performance evaluation of three inference engines as expert systems for failure mode identification in shafts

This paper aims to present performance evaluation of three different inference engines (rule based reasoning, fuzzy based reasoning and Bayesian based reasoning) for failure mode identification in shafts. This research was done with a focus on the validation cases and results after their use in failure cases from several industries where the three systems were tested under the same conditions.Each system was implemented using the same user interface and knowledge base, with different frameworks and techniques as follows: rule based inference reasoning (prolog, C#), Mamdani-fuzzy based reasoning (C, MATLAB®) and Bayesian based reasoning with a variable elimination algorithm (C, MATLAB®).The best performance was obtained using the Bayesian inference engine. The conditional probabilities give flexibility when evidence is not listed, while the fuzzy and classical IF-THEN systems depend on the rules in the inference engine.The process presented in this paper could be used for validation of any expert system or for comparison with other expert systems (inference engines) when the knowledge base is the same.     

Failure investigation of crown wheel and pinion

The crown wheel and pinion are the critical components in the transmission system of an automobile. Failure of these components has drastic effect on the vehicular movement. This in turn leads to increased downtime for repairs. The cost of these components adds to the criticality in addition to its function. A fractured gear was subjected to detailed analysis using standard metallurgical techniques to identify the cause for failure. The study concludes that the failure is due to the compromise made in raw material composition by the manufacturer, which is evident by the presence of high manganese content and non-existence of nickel and molybdenum. This resulted in high core hardness (458 HV) leading to premature failure of the vital component of transmission system in a vehicle.     

Crankshaft failure analysis of a boxer diesel motor

This paper reports a failure mode analysis of a boxer diesel engine crankshaft. Crankshafts are components which experiment severe and complex dynamic loadings due to rotating bending combined with torsion on main journals and alternating bending on crankpins. High level stresses appear on critical areas like web fillets, as well as the effect of centrifugal forces and vibrations. Since the fatigue fracture near the crankpin-web fillet regions is one of the primary failure mechanisms of automotive crankshafts, designers and researchers have done the best for improving its fatigue strength. The present failure has occurred at approximately 2000 manufactured engines, and after about 95,000 km in service. The aim of this work is to investigate the damage root cause and understand the mechanism which led to the catastrophic failure. Recommendations for improving the engine design are also presented.     

Reliability assessment of passive isolation condenser system of AHWR using APSRA methodology

In this paper, a methodology known as APSRA (Assessment of Passive System ReliAbility) is used for evaluation of reliability of passive isolation condenser system of the Indian Advanced Heavy Water Reactor (AHWR). As per the APSRA methodology, the passive system reliability evaluation is based on the failure probability of the system to perform the design basis function. The methodology first determines the operational characteristics of the system and the failure conditions based on a predetermined failure criterion. The parameters that could degrade the system performance are identified and considered for analysis. Different modes of failure and their cause are identified. The failure surface is predicted using a best estimate code considering deviations of the operating parameters from their nominal states, which affect the isolation condenser system performance. Once the failure surface of the system is predicted, the causes of failure are examined through root diagnosis, which occur mainly due to failure of mechanical components. Reliability of the system is evaluated through a classical PSA treatment based on the failure probability of the components using generic data.     

The effect of the threshold stress on the determination of the Weibull parameters in probabilistic failure analysis

In the analysis of brittle materials and components the probability of failure is commonly modelled using a two-parameter Weibull distribution. Occasionally, a three-parameter model is used when the material shows significant threshold behaviour. In this paper two methods for determining the three-parameter constants are discussed. Two theoretical two- and three-parameter distributions are then analysed to examine the number of samples needed to determine the parameters accurately. The two-parameter models are the best fits of the three-parameter models and their failure distributions are very similar to the three-parameter distributions. It is concluded that far more specimens need to be tested than is usually the case to be confident that the correct distribution has been found.     

Reliability analysis techniques for mechanical systems

This paper discusses the application of known system reliability analysis techniques and identifies problems encountered in the practical implementation of these methods, revealing that no single technique is sufficient or even feasible in the case of complex mechanical systems. A new functional analysis method as well as a (new) criticality quantitative approach and failure mechanism analysis are presented and used to analyse an aircraft fuel system. A rigorous and detailed FMECA is still required. Besides its main function it will supply much of the valuable information for many other techniques.     

Analysis levels within the SINTAP defect assessment procedures

Within the SINTAP defect assessment procedures, there are a number of levels of analysis. An increase in level corresponds to increased complexity and requires more detailed data but also leads to improved accuracy. Some levels of analysis are based on existing procedures such as the R6 approach used in the UK nuclear industry and the engineering treatment model developed in Germany. Other levels have been derived during the course of the SINTAP project. In this paper, the levels of analysis for homogeneous components are described and the background to their derivation is discussed. The resulting approaches are illustrated by examples for a range of engineering metals.