Hot dip galvanized bridges have become popular for prevening corrosion due to exposure and reducing maintenance costs during the life time of steel bridges. The reliability of galvanizd friction type joints with high strength bolts may differ from that of ordinary joints of uncoated members. This difference is caused primarily by the iron-zinc alloy and pure zinc layers of the surface of the member. The present paper reports the results of investigation on the behavior of hot dip galvanized friction type joints with high strength bolts under short-duration static load and the secular change of bolt tension. The results are then applied to evaluate the reliability of galvanized joints having bolt tension lower than the specified value as well as of slip-load occurrence. The main conclusions derived from this study can be summarized as follows: After a month, the loss in bolt tension was 10-13% for the hot dip galvanized surface as compared to the initial bolt tension. The secular change of bolt tension and the slip coefficient calculated by using the residual bolt tension could be expressed by use of the regression method based upon the exponential creep law. The probability of slip occurence decreased with increasing duration after the bolt clamping of he joint. The maximum value of the probability calculated was rather small, ranging from 4.8×10-7 to 6.6×10-7.
A disaster simulation program by using a personal computer has been developed in order to investigate the safety evaluation for accidents in plants. The devised matters for program are as follows; (1) The difference of the effects of location, capacity and kind of tanks on disaster can be calculated. (2) The effect of dyke height on disaster can be calculated. (3) The computational results can be graphically shown by personal computer. (4) The affected region for human body by heat and explosive pressure can be calculated. If a tank and pipes in plant have been destroyed by earthquake and so on, the gas or liquid in it will be leaked and fire or explosion will occur. The effects of the damaged tank on another ones are calculated by both heat and pressure of explosion. As an example, the damage progression with time for a model petrochemical plant has been calculated. As a result, the effects of oil dyke and location of tanks on disaster have been analyzed. It is revealed that this program is very useful for safety evaluation and selection of site of fire monitoring system for plants.
When subjective uncertainties are taken into account in structural reliability analysis and subjective evaluations are given in terms of fuzzy sets, we encounter the problem of how to combine a number of evaluations. In this paper, several combination methods used in the field of structural reliability engineering were first examined with respect to the following two points: (i) if the result of combination is independent of order of combination, and (ii) if the result of combination is reasonable when a variety of evaluations are combined. No method examined was found to satisfy both of the conditions. Then, a new combination method was proposed. The method was obtained by modifying Nishimura et al's combination method which was based on Dempster's rule and the concept of vagueness. Modifications were made in treatment of empty sets and vagueness. The method gives the same result independent of the order of combination, and also gives a reasonable result when a variety of evaluations are combined. Furthermore, by using the combination method proposed and preparing five membership functions and three vagueness functions for choice, a way was given for taking account of subjective uncertainties in structural reliability analysis.
This paper attempts to develop an expert system for assessing the reliability of existing structures. Similar to the usual expert system, this system consists of interpreter, rule-base and working memory. However, this system has such a remarkable feature that it includes a fuzzy manipulation system which can treat fuzzy sets in the process of data handling, rule representation and inference procedure. Using this system, it is possible to deal with various kinds of uncertainties and ambiguities involved inherently in the data, rules and inference process in a unified and simple manner. An illustrative example is presented to demonstrate the applicability of the system developed herein.
For application of ceramics to mechanical and structural components, a data base system for fracture strength of ceramics was developed. In this data base system, various fracture strength test data of ceramics are compiled with the data of their manufacturing process and the test environment, which are used to retrieve various fracture test data. The fracture strength of ceramics obtained by each test varies usually in spite of being same in material and test environment. For estimating the strength as a material constant independent of each testing method, this system performs a unified analysis for all the fracture strength test data retrieved. This unified analysis method can estimate the material constant expressed by the delayed fracture strength normalized in terms of cummulative effective hold time Nfteff=1sec and effective volume Veff=1mm3. Furthermore, a reliability design system was developed for analyzing the strength reliability of the structural components of ceramics against the actual service loads. It is possible by this system to determine the allowable design stress by applying the normalized strength properties to the probabilistic analysis. This system has a wide applicability in performing the reliability-based design of ceramic structural components as well as in developing ceramic materials.
Unidirectional fibrous composites are supposed to give maximum performance under an optimum fiber orientation angle. But the performance is very sensitive to design variables and loading conditions. Therefore, the best configuration under a deterministic condition still has some problems to be solved from a probabilistic viewpoint. In this paper, a method for evaluating the reliability of unidirectional fibrous composities under probabilistic conditions is proposed, and the effects of various factors on the reliability are investigated. The proposed method is based on a recent structural reliability theory, which is called AFOSM (Advanced First-Order Second-Moment) method. It is found that the proposed method is useful and that the variations of the principal strength, applied stress and orientation angle reduce the reliability. An emphasis is placed on the orientation angle along which the maximum reliability is obtained, and it is found that the optimum angle varies with the variation of the applied stress in some cases.
This paper deals with a method for representing the distribution of fiber orientation in fiber reinforced composites. It is well known that the mechanical properties of fiber composites depend on the elastic properties and volume fraction of the constituent materials, and are affected by the fabrication processing. In particular, the fiber orientation and content can be strongly influenced by the material flow in molding. Therefore, a knowledge of the fiber orientation distribution is a requirement for predicting the mechanical behavior of the fiber composites. In this paper, the fiber orientation density function using an incomplete beta function is proposed in order to express the distribution of fiber orientation for the fiber composite, and the mold flow of fibers in SMC laminates is discussed by the proposed formulation. It is clear that various distributions of fiber orientation can be expresssd by selecting proper paramater values.
The critical values of the modified Kolmogorov-Smirnov goodness-of-fit tests for the extreme-value and two-parameter Weibull distributions were given by means of the Monte-Carlo simulation for the case when the parameters are estimated from a complete sample. As a graphical plotting technique for estimating the parameters, the combination of the median ranks on extreme-value or Weibull probability paper and the least squares method was used. In order to determine the critical values, the Monte-Carlo simulation was applied to generate 100000 sets of samples for each sample size. The critical values were tabulated for sample sizes 3(1)20, 25(5)50, and 60(10)100.
The safe operation of high speed railway rolling stock such as the Shinkansen Line vehicle is achieved by securing the reliability of important parts of a car like axles and wheels. The strength design method for the car axle is based on JIS E 4501. According to this method, the design load is calculated by using a dynamic factor based on the vibrational acceleration of a car. This paper presents a method to determine a new dynamic factor based on the running test results. The design stress calculated by the new dynamic factor is smaller than that calculated by the old one, and is closer to actual stress. The maximum value of the actual stress is 60 to 90% of the design stress. It seems that the design safety factor based on the design method specified in JIS will give a value on the safe side. With the train speed up, the actual stress becomes larger; therefore the structural development of a light weight car and the improvement of fatigue strength of the axle are important measures to reduce the unsprung mass.
In order to obtain fatigue strength data in long life range over 108 stress cycles under multiple two-step varying load conditions with high frequency of Hi/Lo block cycles, an automatic load changing device for uniform rotating bending fatigue testing machine was developed. With this device can be conducted the repeated Hi/Lo varying load fatigue tests with short interval of the higher stress level of about 100 cycles. It was confirmed, through experiments done by using JIS. S45C round bar specimens, that the level of performance of the newly developed load changing device was high. Furthermore, the discussion on the results from the viewpoint of the linear cummulative damage laws showed that, in some cases, the effect of understressing on fatigue lives appeared in two opposite manners, i.e., fatigue lives became considerably longer than estimated values or vice versa, especially when the number of understressing cycles was sufficiently large. To reveal such a inconsistent effect of understressing, and further, to clarify the fatigue behaviors in long life range up to 2×109 stress cycles, the authors intend to conduct a series of experiments systematically.
The relation between the range of stress intensity factor ΔK and striation spacing S was obtained by using CT specimens of various kinds of structural steels with different hardness, and the actual fatigue stress of square or round bars under repeated 3-point bending load, which was assumed to be an actual machine part, was evaluated from the Vickers hardness HV and the striation spacing measured in the bars. The material constants, Cs and ms, obtained from the relation between S(mm) and ΔK (MPa√m) for the CT specimen, S=Cs(ΔK)ms, were able to be approximately expressed as Cs=1.68×10-10(HV)1.88 and ms=19.1(HV)-0.53, respectively. The actual fatigue stress of the square bar under 3-point bending agreed relatively well with the stress calculated from the striation spacing measured on the bar and the values of Cs and ms, which were obtained from the above equation by substituting the hardness. For the round bar, however, the calculated stress was larger or smaller than the actual fatigue stress when the crack length is short or long, respectively.
To clarify the interaction effect of stress changing on the crack propagation behavior, the fatigue tests of which stress sequence had high to low and low to high were carried out on high strength steel JIS SNCM439. The results obtained were as follows: 1) The acceleration behavior of crack propagation rate was observed just after changing the stress level from low to high, and the crack propagation rate became two times of that under the constant loading. This acceleration behavior continued for 500-1000 cycles of high stress loading. 2) The retardaton behavior of crack propagation rate was observed just after changing the stress level to low from high. The number of cycles of low stress loading needed for the crack propagation rate to return that of constant loading was about 10000-70000. 3) It was difficult to elucidate the acceleration and retardation behaviors only by the change of crack opening stress level. 4) It was necessary to consider the accerelation effect of crack propagation rate for the estimation of fatigue life under service loading conditions when a higher stress is intermittently loaded.
The reliability evaluation system for ceramic gas turbine components was developed. This system ‘GFICES’ is based on the statistical strength theory using 2 parameter Weibull distribution. Main functions of this system are fast fracture analysis, static fatigue analysis, dynamic fatigue analysis and the evaluation of the effect of proof testing. In addition, several other functions such as application of bimodal Weibull theory and consideration of aging degradation of ceramic components are available. Being applied to the first stage ceramic bucket, this system has been found to be effective for the strength evaluation of ceramic structural components.
A rotor disk of jet engine and/or gas turbine is subjected to the intermittent stress waves, that is, the repetition of centrifugal stress at the start and stop of machine with superimposed vibrating stress. The serrated part and bolt hole of the rotor disk are important parts needed for fatigue strength evaluation. A number of studies have been conducted on the fatigue strength under the intermittent stress wave and the fatigue strength estimation of notched specimens. The fatigue strength evaluation method under the intermittent stress proposed by authors is being used for the fatigue life prediction of piston crown of a diesel engine. However, the adaptability of the method to the rotor disk has not been verified yet. The present paper aims at the verification of the adaptability of the method for a rotor disk made of martensite stainless steel FV535. The following conclusions were obtained from this experimental study. 1) The fatigue crack initiation life of a notched specimen under constant stress amplitude was well predicted by Koe's or Neuber's method. 2) The fatigue life prediction method proposed by the authors may well predict the fatigue life of a smooth specimen in the region of higher stress level than the fatigue limit by using the interaction coefficient C of 0.3. However, in the region of stress near the fatigue limit, the method predicted very conservative life and the coefficient C was more than 5.