A Statistical Adaptable Distribution Function Model for Low Probabilities of Failure

Weibull, log normal, and some other Distribution function models (D.F.M.) have a tendency to deviate from experimental results. This deviation, either exceedingly conservative or nonconservative, is amplified at low probabilities of failure. To remedy such problems a new D.F.M. is derived. It is then used to predict low probabilities of failure. The predictions are consistent with experimental data and are not too conservative or too nonconservative.

Fatigue Behavior of Adhesive Joints Under Biaxial Loading Conditions

Fatigue data obtained under biaxial loading conditions for adhesively bonded joints are used to plot S-N type diagrams to assess the effects of biaxiality in loading. Independently Loaded Mixed-Mode Specimens (ILM MS) are used for data collection purposes. These specimens are basically two (steel) beams bonded to be fatigue loaded under cantilever (opening) mode while a simultaneous but physically separate in-plane (static) shear load is also induced with the aid of a small hydraulic piston embedded in the specimen. Application of such static shear loads results in different S-N behavior for the bonded joint. The model adhesives used are Metlbond 1113-2 and Metlbond 1113 solid film thermosetting adhesives similar to those commonly used in aircraft and aerospace industries. The former is an elastomer-modified epoxy adhesive and the latter is identical except that it containes a synthetic earner cloth. Thus, the effects of carrier cloth in adhesive’s S-N behavior is also assessed. Analytically, the classical linear log-log representation of the adhesive S-N data is explored and modifications necessary to reflect the effects of biaxiality in loading and also the presence of a carrier cloth are assessed. The fatigue failure results are also compared with results obtained under monotonic biaxial loading conditions.

The Doctrine of Manifest Danger and its Relationship to Reliability, Preventive Maintenance and Fail-Safe Design

“Dust thou art, and unto dust shalt thou return.”1 Man has made no observations that would challenge this notion from the Bible and certainly safeguarding systems fall into lockstep. Safety technology has responded to the reality of eventual degradation using four general approaches: reliability design, preventive maintenance, fail-safe design and danger manifestation. The optimum implementation of these approaches will still not eliminate accidents; indeed, no work of manor nature is or can be danger free. Nevertheless, these sophisticated approaches are capable of producing ever-increasing levels of safety, albeit, with attendant ever-increasing cost. It is at once unfortunate and unacceptable that common law2 is not equally sophisticated in dealing with the inevitable failure of safeguarding systems over time. This paper introduces The Doctrine of Manifest Danger which is defined as a design concept using direct cues or indicator devices to communicate to the community of users that the safety of a system has been compromised before injuries occur. Furthermore, the paper addresses a related legal issue by distinguishing between proximate cause and cause of action.

Fatigue Evaluation in Ceramic Materials by Acoustic Emission

Progressive damage due to tension-tension cyclic fatigue loading for three distinct ceramic materials was evaluated using the acoustic emission (AE) technique. The objective of this study was to determine the capabilities of the AE method to detect the imminence of failure and to locate potential fracture sites. Results indicated that the AE technique was capable of predicting failure by showing an increase in energy/count rate prior to failure. Although potential fracture sites can be identified, exact location of the final fracture site can be known only when catastrophic failure takes place.

Probabilistic Failure Prediction of SCS-6/Ti-15-3 MMC Ring

Two-parameter Weibull analysis was used to predict the fracture strength and fatigue life of an SCS-6/Ti-15-3 MMC ring from coupon sample data. The fracture strength and fatigue life of the ring were assumed to be volume dependent. The predicted fracture strengths were determined in terms of maximum allowable ring internal pressure. Two methods were used. One simple method was to calculate an effective volume for an idealized ring on the basis of a theoretical solution approximating the stress distribution. The fracture strength and fatigue life of the coupon samples were then scaled to the effective volume of the ring. The other method utilized finite-element analysis to determine a more realistic stress distribution in the actual, geometrically imperfect ring. The total reliability of the ring was then determined by the product of the elemental reliabilities with coupon samples used as a gage. These approaches were compared with experimental fracture strength results. No fatigue data for the ring were available for comparison. Preliminary results indicate that Weibull analysis of coupon samples shows promise in predicting the fracture strength of metal-matrix composite structures.

Approximate Residual Interface Compression in a Laminated Magnet

Present design of large laminated magnets are based on effective modulus of elasticity for the plate stack that is invariant with interplate compression. Experimental results indicates this is not a valid assumption. This analysis considers the specific design used for compact electron storage ring dipole magnet at the Center for Advanced Micro Devices (CAMD) at Louisiana State University. An iterative technique, using FEM is developed to approximate the effective modulus throughout the magnet structure. This particular magnet is constructed from 1424 steel plates (1.5mm × 610mm × 780mm) by first compressing a 45° 2.93m raduis curvilinear stack to a specified preload and then welding straight and curved straps to the exterior of the stack. Release of the preload allowsexpansion of the stack and forces stretch of the straps, the resultant interplate compression is considerably different from the initial preload and varies throughout the magnet. The analysis technique introduced in this paper is a simplified approach to the interplate compression modeling and consists of the supeposition of two simple analysis; one with straps attached, one without straps attached. An iterative approach is used to incorporate the dependence of effective compressive modulus on the interplate compression. An estimate of the average modulus is assumed for the entire stack and residual compression is calculated. From this residual compression and from experimental data, a distribution of compressive modulus throughout the magnet is computed. From this computed distribution, a new estimate for the compressive modulus is made for each element of the model and the analysis iterated. The net results are the compressive modulus distribution throughout the magnet ia a form suitable in subsequent dynamic analysis.

Computational Simulation of Hot Composite Structures

Computational procedures are described to simulate the thermal and mechanical behavior of high temperature metal matrix composites (HT-MMC) in the following three broad areas: (1) behavior of HT-MMC from micromechanics to laminate via Metal Matrix Composite Analyzer (METCAN), (2) tailoring of HT-MMC behavior for optimum specific performance via Metal Matrix Laminate Tailoring (MMLT), and (3) HT-MMC structural response for hot structural components via High Temperature Composite Analyzer (HITCAN). The complex composite material behavior, static/fatigue life, and failure sequence of SiC/Ti ring was simulated. The observed experimental degradation in strengths of the SiC/Ti composite with increasing temperature was accurately predicted. The static/fatigue life of the SiC/Ti ring starting with the fabrication process cool-down and subjected to simulated internal pressure was predicted in terms of cyclic-stress-to-static strength-ratio versus cycles to failure.

MMC Ring Fatigue and Fracture Life Prediction: An Engineering Model

The framework of an engineering creep-fatigue durability model has been adapted for use in estimating the radial static burst pressure and cyclic low-cycle fatigue macro-crack initiation resistance of continuous fiber reinforced (CFR) metal matrix composite (MMC) rings for application at 800 °F. Rings of circumfrentially wrapped SCS6/Ti-15-3 were manufactured by Textron Specialty Metals and burst tested by Pratt & Whitney as a part of a cooperative program with the NASA Lewis Research Center. Fatigue tests have as yet to be performed. The engineering model is based on a 3-D elasto-plastic micromechanics analysis of the tensile-loaded composite architecture. Use is made of the rule of mixtures, strain compatibility, equilibrium, and the stress-strain relationships of the constituents. Knowledge is required of the mechanical and fatigue properties of the matrix and fibers and how the presence of each affects the sharing of imposed stresses and strains. The model addresses specific issues such as residual fabrication stresses, inelastic deformation within the ductile matrix, multiaxial constraint imposed on the matrix, cyclic relaxation of both residual and applied mean stresses in the matrix, fatigue micro-crack initiation and propagation in the matrix, and tensile fracture of both the ductile matrix and the brittle fibers. In the current application of the model, the specific issues were empirically calibrated through use of tensile and tension-tension fatigue coupons that had been subjected to essentially identical loading as the rings.

The Effect of Fiber Type on the Level of Stress Concentration Created by U-Notches in Long-Fiber Composite Plates

The effect of fiber type on the level of stress concentration created by u-notches were investigated for unidirectional and symmetric cross-ply laminates loaded on or off-axis. The finite element method was used for this purpose utilizing triangular elements under plane stress conditions. Analyses were performed for three different fiber reinforcements of epoxy matrix: glass, graphite and boron. Typical material properties reported in the literature for glass/epoxy, boron/epoxy and graphite/epoxy were used in calculations. The accuracy of the finite element code utilized was first checked by calculating the u-notch stress concentration factors for an isotropic steel plate.

A Life Prediction Method for Matrix Dominated Fatigue Failures of SCS-6/Ti-15-3 MMC

This paper presents a method for determining the matrix dominated cyclic fatigue capability of continuous fiber reinforced SCS-6, Ti-15-3 metal matrix composites (MMC). The basis of the fatigue life prediction system relies on the matrix material dominating the fatigue behavior of the MMC. This assumption has been supported by test data and failure mechanisms identified in unidirectional isothermal and out-of-phase thermomechanical fatigue (TMF) specimens which typically show an area of matrix cracking before failure. Since the matrix fatigue capabilities dominate the fatigue behavior of the composite, an MMC thick cylinder model with an anisotropic coefficient of thermal expansion is used to characterize the state of stress in the matrix. The Smith-Watson-Topper fatigue stress damage parameter is applied to the calculated matrix stresses to determine a “corrected” matrix fatigue stress which is isolated from laminate geometry and laminate stress. This method correlates well to room and high temperature SCS-6/Ti-15-3 MMC test data. Since the “corrected” matrix stress fatigue response is independent of fiber volume, stress ratio, and load history, this MMC Life Prediction System can be used to predict the fatigue behavior of any reasonable laminate and loading in which the failure is matrix dominated. With this life system, life analyses can be performed for laminate geometries and loadings at which no test data exists, which will prove valuable when designing engine components where optimization of MMC capabilities are required but testing is limited.