Abstract of Special Research Report (RR-89)

National Institute of Occupational Safety and Health, Japan

The Influence of Size and Aqueous Environment on Fatigue Crack Growth Behaviours in Large Structural Low-Carbon Steels

RR-89-1

Masazumi TANAKA

: Structural elements in large scale machines and constructions are often used under the condition of low cycle fatigue where heavy loads are slowly repeated, and besides, in the corrosive environment such as in rain and sea water.
    In this study, as part of the investigation to clarify their fracture characteristics and ultimately to prevent possible fracture accidents under such conditions, fatigue crack growth experiments have been carried out on large plate specimens of two kinds of low carbon structural steels under low cycle, that is, high stress fatigue conditions. And the effects of width and thickness of specimens and the influences of salt and pure water environments under free corroding condition on crack growth behaviours, the application limit of formula dl/dN = C(ΔK)^m, and the relation of macroscopic crack growth rate dl/dN and microscopic fracture morphology have been examined by means of fracture mechanics and fractography.
    Main results obtained are as follows,
  (1) The relation between macroscopic crack growth rate dl/dN and stress intensity factor range ΔK is not remarkably affected by specimen sizes : width from 70 to 500 mm and thickness from 3 to 28 mm, and can be represented by the following simple formulae;
    dl/dN = 4.9 × 10^-5 × ΔK^2.3 .... for SB42,
and dl/dN = 3.7 × 10^-4 × ΔK^1.9 .... for HT80
respectively in the wide range of dl/dN up to about 2μm/c. For higher rates, the narrower the specimen width is, the earlier the dl/dN value deviates upwards (to higher side) resulting in the narrower range where these formulae are applicable. Final fracture occurs under the net stress condition influenced by the maximum stress intensity factor K_max.
  (2) Macroscopic fractures of thick plate are of rectangular type both in air and in water for SB42. Those for HT80, however, change from rectangular-shear transition type in air to rectangular type in water, showing remarkable environmental effects.
  (3) Aqueous environment under free corroding conditions remarkably accelerates subcritical fatigue crack propagation, regardless of the existence of desolved oxygen and salt in water. And the lower the load frequency, the larger the acceleration effects.
  (4) This acceleration effect relates with the brittle fracture mechanism represented by the brittle striation of SB42 and the quasi-cleavage fracture of HT80.
  (5) The width of ductile striations formed dominantly in air practically coinsides with the dl/dN value. On the other hand, the width of the brittle striations of SB42 is fairly larger than that of ductile striations and also dl/dN at the corresponding AK, increasing with the decrease of load frequency.
  (6) Such brittle fracture behaviours in water environment seems to be caused by local embrittlement of crack tip region due to hydrogen cyclically induced by the plastic deformation.

Fatigue of Fine Grain Grinding Wheels Binded with Organic Bond

RR-89-2

Soichi KUMEKAWA

: Recently, grinding processes are more widely applied for machining parts of computers, video equipment, cameras, etc. than even before. In the process of grinding these parts of newly developed equipment, the grinding wheel are binded with organic bonds.
    In order to reduce ill effects of heat produced in the works by grinding, hardness grade and strength of the grinding wheels are required to be much lower than those of general-type grinding wheels.
    With more application of such fine grain grinding wheels binded with organic bond, their fracture incidents are also being increased in manufacturing workplaces.
    In general, a high-speed rotating grinding wheel as a tool tends to give rise to serious hazards, resulting in a wheel fracture incident. In order to prevent such hazards, investigation the strength of grinding wheels is necessary as a basis of scientific discussion. However, studies on strength of fine grain grinding wheels binded with organic bond have not been reported so much. Particularly, research reports on strength behavior of fine grain grinding wheels, binded with organic bond, under repeating load have not been published to date.
    This study features the method that diametral compression loading, as shown in Fig.2, is applied to repeating load on the specimens. Two types of specimens used in this study. The one is the grinding wheel which is binded with resinoid bond (grain "GC", grain-size 800, hardness grade F) and the other binded with PVA-phenol mixed bond (grain "GC", grain-size 3000, hardness grade H).
    Figure.6 shows the experimental results of statical loading test in the diametral compression method, and experimental results of repeating loading test in the same method are shown in Figs. 7 and 8.
    It is confirmed experimentally that the results of repeating loading test of the fine grain grinding wheels binded with organic bond do not always depend on the effect of time-dependence (delayed fracture) based on experimental results of the dead load type compression test in which a dead load is applied on specimens for 100 hours.
    The author concludes from the experiment conducted that the strength of fine grain grinding wheels binded with organic bond, under repeating load, depend upon the cycle number of repeating load that leads to fracture.

Evaluation of the Loudness for Quasi-Steady Impulsive Noise

RR-89-3

Yoshiyuki EGAWA

: The relation between loudness and physical properties of noise has been studied in various conditions, thereupon some standards for measurements of the noises have already been established for steady state noises. However, for non-or quasi-steady impulsive noises emitted from press machines or air hammers, we are yet to establish such standards, particularly on noisiness of the impulsive noise that responsible for difficulty in hearing.
    In order to establish a method for the estimation of loudness, this study focuses on examining the loudness of quasi-steady impulsive noise emitted from percussive tools in factories.
    Many impulsive noises emitted from percussive tools were recorded in magnetic tapes, and the wave forms and time intervals between each block of impulsive waves were examined. To examine the loudness for the quasi-steady impulsive noise, the loudness test under the same equivalent continuous A-weighted sound pressure level (=L_Aeq) was carried out using some artificial wave forms which were closely simulated to the waves recorded. Based on the experimental data regarding the relation between loudness and physical properties of waves, the response model of auditory organs for the quasi-steady impulsive noise was clarified.
    As the result of observation of the wave forms and time intervals between each block of impulsive waves for the quasi-steady noise, it was found that the amplitude of the waves were decreased exponentially, and time intervals between each block of the impulsive waves were less than 500 ms. The loudness under the same value of L_Aeq was found to increase with increasing the time interval of the impulse.
    Finally, as the consequence of simulation on the experimental data of loudness test, the author concludes that the loudness is correlated with the physical properties for the quasi-steady impulsive noise, when a time constant in response to auditory organs is 90 ms.

Study on Slip-Resistance Measurements for Footwear (1st Report) --Fundamentals of Measuremant--

RR-89-4

Hisao NAGATA

: Up to now, much interest in safety shoes has been given to the strength of toe cap and the penetration resistance of outsole, and until now little attention has been paid to other properties such as comfortability and slip-resistance of shoes in our country. Dozens of slip-resistance measurements have been developed over the world during the past decades. But no devices and techniques have got international acceptance for universal slip-resistance measurements. In order to attain the goal to establish an internationally accepted universal slip resistance measurement, the methodology of insuring the validity of a meter should be proposed in the slip resistance studies of protective footwear.
    Just as tribologists have emphasized, the coefficient of slip resistance is prone to be dispersive and changeable by many measuring conditions, such as weight of load, test speed, contact area, contact time with flooring materials, temperature and repetitions of measurements.
If we use the meter only in a certain test condition as given between lubricated stainless steel floor and safety shoes, assurance of the meter may be obtained within a limit because of small values of slip-resistance of lubricated surface and their small variance. In order to improve the meter the following important points should be generally stressed:
  1) Our ultimate goals are to establish reasonable criteria to assess the safety shoes in terms of slip-resistance and to estimate risks of fall or slip in working environments, and not to analyze complicated tribological phenomena in slippage of shoe and floor.
  2) The new decisive methodology which deals with the tribological phenomena as a black box is to be introduced into the study of slip-resistance measurements. Tribological methods will be referred to, but not to be decisive.
    The notion of slippage will be divided into slips and slipperiness. In slip-studies, very slippery floors like lubricated surfaces will be adopted, and in slipperiness studies, a wide range of slip resistance of surfaces would be involved. For example, in the latter case the lower limit of slip-resistance is badly needed, but the upper limits is also required to avoid foot twisting or cause falls in stopping, like on gymnasium floors by measuring maximum static friction, these are used mainly in evaluating flooring materials. But in the former case they are used mainly in evaluating safety shoes on slippery floors by measuring dynamic friction in slipping.
    It is necessary to know what values measured from a tester mean, in order to evaluate the validity of a slip-resistance test for protective footwear. The values should indicate the rate of occurrence of actual fall-or slip-resistance induced by foot-slippage. It is therefore very difficult to obtain the actual rate for each combination of floors and footwear. In this report, the psychometric and biomechanical methods are applied to insure the validity of a slip-resistance metef, and the paired comparison methods or ramp tests were executed. As the results, each methods would implicate the means to determine an optimum meter or to understand the properties of slippage.

A study on the Time Intervals between Accident (5)

RR-89-5

Shigeo HANAYASU

: The accident frequency rate has been widely used as a measurement of safety performance in many undertakings over a long period of time.
    In order to explore the significant changes in accident situation in succeeding intervals of time, the time intervals between occupational accidents were utilized as a useful indicator to give expression to safety performance in undertakings having a certain accident risk.
    This paper deals with the stochastic treatment of the time intervals between occupational accidents to evaluate the safety performance of working places. Emphasis was placed on the probabilistic analysis of the time intervals between accidents and the number of accidents, when the accident frequency rate varies and depends on a certain probability function.
    The main features of this paper are as follows :
  (1) From the analysis of frequency distribution of the number of occurrences of accidents wihtin a fixed interval of time, it was found that the frequency distribution of many accidents agree with the negative binomial distribution as well as the poisson distribution simultaneously. Particularly, negative binomial distributions have many good agreement with actual accident distributions rather than the poisson distribution.
  (2) Depending upon the results of frequency distributions of accidents, it was recognized that many accidents take place at random. However the accident rate itself depends on the probability distribution whose probability function can be expressed as the gamma distribution. Hence, the frequency distribution of the number of occurrences of accidents as well as the frequency distribution of time periods between accidents are represented as the mixture of probability density functions. This mixing of the probability function of accident rate to the distribution function of the number of occurrences of accidents ultimately reaches the negative binomial distribution. Whilst, the mixture of the frequency distribution of accident rate to the frequency distribution of time intervals between accidents finally reaches the composite exponential distribution.
  (3) If sufficient accidents data are available, parameters of the probability distribution function of the negative binomial distribution or the composite exponential distribution, can be estimated by making use of these accident data.
  (4) In case sufficient data are not available, the Bayesian statistical inference can be employed as an alternative method to estimate the parameters of the distribution functions. By making use of Bayes' theorem, a posterior distribution of the accident rate can be obtained as the gamma distribution whose parameters are the number of accidents and duration of time investigated. Then, the frequency distribution of the number of occurrences of accident within a given period of time, can be obtained by way of mixing the probability distribution function of accident rate to the poisson distribution, which yields the negative binomial distribution. Similarly by mixing the accident rate distribution with the exponential or gamma distribution, the frequency distribution of the time periods between accidents, which takes into consideration the variance of accident rate, can be represented as the composite exponential or gamma distribution.
  (5) Utilizing Bayes' theorem, a renewed posterior distribution of accident rate can be estimated at the time when new accident data are acquired. Then by making use of this up-to-date accident rate distribution, the renewal of the probability distribution functions of the number of occurrences of accidents as well as the frequency of the time periods between accidents, can be achieved.

Study on Test Methods and Evaluation of Hazardous for Dangerous Materials

RR-89-6

Hidenori MATSUI, Takayuki ANDO, Yasuhiro FUJIMOTO and Shigeru MORISAKI

: In recent years, it has widely been discussed on the definition or classification of dangerous substances or materials for the purpose of their safe production or transportation. From a safety point of view, it is preferable to determine by appropriate tests whether or not a chemical substance has dangerous properties such as detonability and highly combustibility, or in which group the chemical is classified. Though various kinds of test methods are presented so far for the evaluation of dangerous substances, it is not easy to carry out all these tests.
    In Japan, the term of "Dangerous substances" is defined by the Fire Prevention Law and the Industrial Safety and Health Law. The table of the dangerous substances cited in the Fire Prevention Law has recently been revised to include the test methods, and it is thereby necessary to examine in industrial establishments whether the chemicals handled/treated are referred to as dangerous substances or not.
    In this report, the test methods and the criteria of classification proposed by the Fire Prevention Law were experimentally examined to investigate the possibility of revision of the table of dangerous substances described in the Industrial Safety and Health Law. In the experiments, the common dangerous substances defined in the Industrial Safety and Health Law and the Fire Prevention Law were tested according to the test methods by the Fire Prevention Law, and then the other dangerous substances were tested by our original test methods.
    When test methods are introduced to evaluate chemical hazards, the following conditions will be indispensable to propagate the test methods extensively to industry;
  (1) The test methods are those capable of estimating latent hazards of chemicals.
  (2) Operation of the test methods is easy and the results are reproducible.
  (3) A required quantity for the test is small.
  (4) The test can be conducted safely.
    From the results of this study, it is recommended that the following test methods which satisfy the above conditions will be adopted according to each group of dangerous substances.
  (1) Thermal analysis(DSC) for explosive substances.
  (2) Auto-ignition test and reaction test with water for ignitable substances.
  (3) Drop ball shock-sensitivity test for oxidizable substances.
  (4) Flash point test(Seta type) for inflammable substances.
  (5) Diffusion burning test for combustible gases.
  (6) Small flame test and dust explosion test(Hartmann type) for inflammable solids.

Static Electrification of Dielectric Liquids Caused by Agitation

RR-89-7

Tsutomu KODAMA and Yasuyuki TABATA

: Agitation of flammable liquids is often conducted in many manufacturing factories for various purposes such as suspension, emulsion, solution, blend, mix and so on. It is well known that when low conductivity liquids contain impurities like solid particles and water droplets a large amount of static charge accumulates in the agitated liquids and thereby fires and explosions may take place due to the discharges of statics electricity.
    The purpose of this study is to investigate experimentally the mechanism of static charging caused by agitation and the anti-static effect of additives for suppressing static electricity.
    In the experiments, liquids are stirred in a cylindrical stainless vessel of 310 mm in diameter and 315 mm in depth with an impeller of paddle type and measurement was made of the electrostatic field strength with a static field meter set on the top cover of the vessel. This field strength was transformed into the charge density by multiplying a certain factor calibrated from the experiments.
    The results obtained from the experiments are as follows:
    Neither pure liquids (kerosene, methanol, ethanol, acetone, xylene, n-hexane) nor the mixtures of pure liquids (xylene/methanol, xylene/ethanol, Xylene/n-hexane) yielded a significant level of static charge by agitation.
    When the water content of kerosene was above 0.3% vol, the charge densities increased with the increase of water content of as far as 30% vol and reached a high level of about 10μC/m³ at higher rotating speeds than 400 r/min. After the stop of agitation under the condition as above, the static potentials continued to increase for a certain period of time, and as the result the relaxation time of static charge also became longer up to about three times as long as the value calculated from the conductivity of liquid in the exponential-decay theory. These phenomena were caused most likely by the charge separation at the interface between the continuous phase (kerosene) and the dispersed phase (water), which settled down to the bottom owing to the difference of specific gravity.
    The kerosene deteriorated by the corona discharge with or without water was also charged up to the same level as the pure kerosene with a high content of water during and after agitation, which suggested the presence of impurities dispersed in the deteriorated kerosene.
    Addition of a conductivity additive "ASA-3" to any kinds of kerosene with or without water could greatly reduce the charge density down to a level of 0.1μC/m³ during and after agitation when the conductivity of the liquid was just raised up to about 1000 pS/m.
    From the discussion on the generation and leakage of static charge, it was guessed that the static electrification of liquids with the dispersed phase caused by agitation was mainly ascribed to the preferential adsorption of ions, which were produced a lot at the interface between the dispersed phase and the continuous phase, to the wall of the vessel.

Nuisances on Electronic Device due to Electromagnetic Wave Radiated with Electrostatic Discharge

RR-89-8

Yasuyuki TABATA and Hajime TOMITA

: An electrostatic discharge "ESD" is known to cause various nuisances on sensitive electronic components and devices, such as electric overstresses, latent failures and malfunctions. An electromagnetic noise radiated with ESD also affects adversely electronic circuits, both analog and digital, using ICs and LSIs. The latter, an electromagnetic interference "EMI" of electronic circuits due to ESD, is increasing as electronic devices become more and more complex.
    From the background described above, an electrostatic damage level and immunity test of the ICs and LSIs have been studied and the nuisances almost solved from a view point of product reliability of the electronic solid circuit in industry. However, characteristics of electromagnetic fields radiated with ESD, as a radio frequency noise "RFN" due to ESD, have hardly been investigated, since a transient phenomenon as ESD is difficult to measure accurately; as a result, the noise level induced by ESD in the electronic devices has not been covered quantitatively yet.
    The purpose of this study is to provide a frame of the electromagnetic field and RFN due to ESD. The method of approach is simple; the discharge current wave form of the ESD occurring between a charged metallic sphere and grounded one is observed with a frequency-wide-band oscilloscope, and then the electromagnetic field strength around the ESD source and the RFN level induced in electronic devices are deduced from numerical analyses applying the discharge current wave form to the Maxwell Equation and Antenna Theorem, respectively, because the ESD occurring between conductors is the most generally accepted one for determining theoretically and experimentally the electromagnetic field strength and RFN level due to ESD.
    This paper discusses the main factors of the ESD controlling the electromagnetic field strength and RFN level. The results obtained from the numerical analyses and experiments are summarized as follows:
  (1) The discharge current reaches immediately a peak value after the ESD occurs between metallic spheres, and its rise time, a few ns, is somewhat longer than a time constant of the discharge circuit used in the experiment.
  (2) The electromagnetic field radiated in free space depends on the discharge current, and the maximum values of the field strength deduced from the numerical analysis are about 10 V/m and 100 mA/m, respectively in the case that the on-set voltage of ESD is about 10 kV.
  (3) The ESD causes the EMI in an electronic circuit in the vicinity of the ESD source, and the radiated electromagnetic field controls primarily the RFN induced in the electronic circuit in the region more than about 1.0 m apart from the ESD source.
  (4) The RFN level due to ESD, the voltage induced in the electronic circuit, depends primarily on the high frequency elements of the discharge current and on-set voltage.

Threshold of Convulsion for Electric Shocks due to Underwater A.C. Uniform Electric Fields

RR-89-9

Eiki YAMANO and Tatsuo MOTOYAMA

: Working in the water, as a sea, river, etc. and use of electricity in such areas are being increased in recent years. These situations may involve divers in hazards of underwater electric shocks. This paper reports an investigation of such hazards.
    An experiment was conducted with rabbits to find the threshold of convulsion about electric shocks due to a.c. uniform electric field under water. In the experiment, a hair-clipped rabbit was immersed in plexiglass water tank, and exposed to underwater electric field for 20 min per test. During the test, it was observed whether the convulsion (i.e. extension in hind legs-or whole-length) occurred or not. The extension of short duration was also counted as a convulsion.
    The applied electric field was 50 Hz sinusoidal wave, and the 3.5 % salt water, with a temperature of 30°C and a conductivity of 6 S/m, was used. Water level in the tank was adjusted before the experiment to such an extent that the rabbit was able to walk, which was 14.7 cm in depth on the average.
    From the results, the cumulative distribution curves were derived for maximum non-convulsing and minimum convulsing field intensities, and the threshold field intensity for the general population was estimated by the curves.
    The results of investigation are summarized as follows:
  (1) The average of individual minimum convulsing field intensities tested on 9 rabbits was 5.33 V/m and the range of data was between 4.03 and 7.27 V/m.
  (2) The average of maximum non-convulsing field intensities tested was 4.5 V/m and the data ranged from 3.03 to 6.30 V/m.
  (3) The threshold convulsing field intensity, i.e. the limit of motility, for the population will be in the range between 1.1 and 1.9 V/m, in the uniform electric field under water. The values of 1.1 and 1.9 V/m are the 0.5 percentile ranks of the individual maximum non-convulsing and minimum convulsing field intensities, respectively.

JNIOSH