Abstract of Technical Note (TN-83)

National Institute of Occupational Safety and Health, Japan

Investigation into the Fracture Occurred in a Welded Joint of an Ammonia Compressor

TN-83-1

Etsuji YOSHIHISA

: A reciprocating ammonia compressor for refrigerator was accidentally broken in start-up operation. The fracture was occurred in the circumferential welded joint connecting the flat circular plate to the cylinder of the outlet-header of the compressor. Possible causes are studied in this report.
    The following results are obtained through microscopic observation of the fracture surface of the flat plate and through calculation of the stresses which had been raised in the welded joint.
  1) In the greater part of the fracture surface, granular fracture is observed macroscopically and cleavage fracture microscopically. These are the characteristic features of brittle fracture.
  2) In several parts of the fracture surface, which are adjacent to the inside surface of the flat plate, there exist dimples accompanied by microscopic grains which are regarded as inclusions or precipitations. The existence of the grains indicates that materials of these parts are inferior and that their strength may be rather low. One of them is considered to have been corroded before the accident, therefore it is concluded that the part had fractured under operating condition.
  3) It seems reasonable that there had not been complete penetration in the welded joint and that unfused region remained at the bottom of the joint for a considerable length, had caused stress concentration in the neighboring of the fracture surface.
  4) Taking into account the stress intensification caused by the structural discontinuity in the joint, it is considered that both the thickness of the flat plate and the throat depth of the joint were rather small.
    From these results, the joint had not been well produced by welding and the local stress in the joint had been somewhat high. Then it can be understood that the fragile part in the joint had already fractured under operating condition and that a crack generated as a result of that fracture promoted the brittle fracture of the header.

Flame Quenching Performance of Polyurethane Foams and Their Application for Explosion Protection

TN-83-2

Toshihiro HAYASHI

: In this paper an evaluation of the flame quenching ability of plastic foam is given, with particular reference to its use as flame arrester element. The material tested is polyurethane foam with cellular structure of skeletal three-dimensional network consisting of a large number of interconnecting open cells : less than 3 per cent of the total volume is occupied by the plastic network, the rest being void. In the present work, average hole diameter of the cell is defined as 25.4/N (mm), where N is the number of holes per 25.4 mm on the surface of the foam.
    Experiments are carried out in an explosion tube of steel, 286 mm diameter and 2.2 m long, both ends being closed with blank flanges. A foam under test can be mounted in the tube with variations of the distance from the ignition source which is located at one end of the tube. Throughout the experiments a 0.5 per cent by volume propane-air mixture is used. The pressure of the mixture before ignition is either atmospheric or 3 psig. For a given foam, the minimum thickness required for quenching a flame propagating against the foam is determined by inches.
    Even though there have been observed some thermal damages on foam surfaces by flame attack, result of experiments shows fairly good agreement with the rule of thumb describing that the hole diameter of arrester element must be smaller than half the quenching diameter of the explosive gas mixture: with foams of average hole diameter of 1.3 mm or less, the minimum thickness for quenching is only several inches for most flames under test, but with larger holes than 2.5 mm even a weak flame can pass through the foam of 1.8 m thick.
    Through additional series of experiments, in which the explosion tube is filled with foam except for rather small ignition volume, it is found that the pressure generated is far less than that with no foam, even when a flame has travelled through the cells of foams of larger hole diameter. This result suggests that plastic foams may be used as a filler in some kinds of storage tanks in purpose of preventing the development of explosions.

Accident Analysis --Explosion of a High Pressure Carbon-Steel Pipe in a Fuel Oil Desulfurization Plant--

TN-83-3

Masazumi TANAKA

: In the evening of March 31st in 1982, a serious accident happened in the Kashima Petroleum Combinat in Ibaraki Prefecture. A 6 inch pipe made of carbon steel in a fuel oil desulfurization plant suddenly ruptured, and the contents blew off, then, exploded and fired. As the result, five of eight workers who had hastened to the spot were killed and three of them seriously burnt.
    To clarify the cause of this accident, a wide range of investigation was carried out concerning 1) witness of survivors, 2) management conditions of safety and hygiene, 3) maintenance, 4) process operation, 5) damage state of plant, etc.
    Main results obtained are as follows.
  (1) The cause of this accident was recognized to be the rupture of 6 inch pipe, and not the other factor such as inadequency of process operation.
  (2) The cause of the rupture of pipe was "Hydrogen Attack" in pipe material. Because some its characteristics were found such as decarburization, intergranular microcrack and void formation, and the existence of methane and hydrogen gas in large blisters formed in the damaged zone.
  (3) Though a few probable reasons which might bring about the hydrogen attack were proposed and discussed from various points of view, decisional reason couldn't be identified in this investigation.
    Finally, to prevent the serious accident of this kind, some safety measures were proposed, and further it was stressed that some institutions should make active effort to solve the important problems remained in researches on hydrogen attack phenomenon.

Accident Analysis --A Study on Capacity of Protection of Safety Helmet--

TN-83-4

Kiyoshi FUKAYA

: A timbering worker who wore a safety helmet was killed as a result of accident that a flying object fell on his head, because its kinetic energy was greater than the limit of the capacity of protection of the safety helmet.
In order to estimate the impact load against the helmet, as well as speed and mass of the flying object, a few kinds of test were made on the capacity of protection of safety helmets which were the same type as the one which the victim wore. These were shock absorption tests of safety helmet and its parts, i.e. shell and cradle, and strength tests of shell and cradle. By comparing the degree of damage of the helmet which the victim wore with that of helmets which were destroyed by tests, the impact load, speed and mass were estimated to be at 660 kgf - 980 kgf, 6.3 m/s - 4.4 m/s and 5 kg - 10 kg, respectively.
The result of the tests made in connection with the accident show the limit of the capacity of protection of safety helmets.

The Method of Analysis on Industrial Accident by Causal Tree Analysis

TN-83-5

Yoshimi SUZUKI

: It is commonly recognized that occupational accidents occur as the result of a concatenation of various factors. From this point of view, it is necessary to investigate the process of occurrence of the accident in order to establish the effective safety measures. As an important clue to these investigations, what is brought to attention is the systematic and quantitative analysis on the relationship between various factors of accident, and on the interaction of these factors.
    This paper deals with the CTA (Causal Tree Analysis) method in the hope to apply its usefulness not only for the countermeasure of occupational accidents, but also for the method of analysis on the process of occurrence of the accident. Though the CTA is an a-posteriori method (Table 1), it can be considered to be one of the approaches to predicate clearly and facilely the relation between various incidents which will lead to an accident.
    In this paper the principal of CTA and its necessary steps in operation are firstly presented in outline, i.e., the formation of a diagram of accident (causal tree), the notion of PFA (potential factor of accident) (Fig.1 - 4, Table 2 - 3).
    Secondly, as an example of usage of the CTA, some accidents which occurred in several sites of construction work are studied, their diagram of accident being indicated (Fig.5 - 9).
    Then, some interactive types of PFA are arranged and classified, and these were presented as a matrix (Table 3).
    Finally, 50 accidents caused by slope-failure and rock-fall in cutting work sites are studied by using the CTA method. Concerning these accidents, the frequency of factors of accident and the relationship between these factors are obtained (Table 4). With these results, it is noted that, although this type of accident is principally dependent upon the factors of physical environment, there is yet other factors, occupying about 60 per cent of total number of all factors which took parts in the process of occurrence of accident. These results were discussed in detail, comparing with an independent study, as one of the simple applications of the CTA. It is clarified that the results obtained in this research would be the fundamental data for the further examination of the internal structure of occupational accident ; for example, examining the difference of the structure according to the industrial branches, the sort of work and the type of accident.

JNIOSH