Abstract of Special Research Report (RR-19)

National Institute of Occupational Safety and Health, Japan

Experimental Study on the Methods of Explosion Venting (First Report) --Application for a Certain Type of Commercial Box Dryer--

RR-19-1

Noboru TAGUCHI, Heizaburo TURUMI Toshihiro HAYASHI and Hidenori MATUI

: When flammable gases, vapours or dusts are handled in enclosed equipments, it would be better to consider that potential hazards of explosion exist unless special cautions have been taken for the concentrations of flammables or the ignition sources. Therefore, some methods, not only for explosion prevention but also for explosion protection, are demanded for dangerous equipments to minimize the serious damage of explosions, if they occurred unfortunately.
    Venting explosion is one of the typical methods of protection and safety apparatus for this purpose are called "Explosion Reliefs". They are classified into two types according to the internal pressure of equipments to be used. Explosion reliefs for high pressures are known as Rupture Discs or Rupture Diaphragms, being attached to, for example, reactor vessels of chemical plant ; but their responsibility for rapid change of pressure, such as like explosion, are doubtful and further experiments must be per formed. Another type of explosion reliefs is called "Explosion Door" and it may be useful only for the internal pressure of near atmospheric. With this type, experimental data for practical use are very poor and we can see only a few cases of application. This report deals with the latter type of reliefs.
    Experiments are carried out using a cubical steel vessel of 1 m³ internal capacity, whose upper side is opened and flanges with various vent areas can be bolted to it. Flammables are L.P.G. (consisted mainly of propane and butane), pure propane and methyl alcohol vapour. They are mixed with atmospheric air respectively in the vessel, and having been made homogeneous by the use of rotary blower, the mixture is ignited through a nichrome wire heater set in the center of cubical vessel. Vent covers are vinyl sheet, craft paper and aluminium foil, each of their thickness being about 0.1 mm. Pressure-time relations are recorded by the system ; straingauge transducer-dynamic strain amplifier-electro-magnetic oscillograph.
    Main results are as follows ;
  (1) Vent ratio (ratio of vent area to volume of explosion vessel) and the ratio of vent areato cross section area of the side wall, to which explosion reliefs are attached, are important factors for the effectiveness of explosion reliefs.
  (2) When vent ratio becomes suber than certain value, pressure-time curve shows two evident peaks (p₁ and p₂) and as the vent ratio decreases the ratio p₁/p₂ increases.
  (3) When 60 cm square vent area is attainable and if vent cover is the one that is easily broken by a slight pressure increase, maximum explosion pressure may be depressed under 1 kg/cm², and this value has been evaluated as the limit strength of many of the practical equipments of light construction.
    These results are directly applicated for a commercial type of box dryer, and three tests using propane-air mixture showed that the designed explosion relief of aluminium foil type is effective under experimental conditions. The records of high speed photography show that the expelled unburned substances explode continuously just outside of the relief and the flame reaches as so far as 4 meters.
    It is then concluded that safeguard for external explosion and the selection of setting site of equipments must be carefully dealt with.

Experimental Study on the Methods of Explosion Venting (Second Report) --Application for a Certain Type of Commercial Dust Collector--

RR-19-2

Noboru TAGUCHI, Heizaburo TURUMI Toshihiro HAYASHI and Hidenori MATUI

: Venting explosion is one of the typical methods of explosion protection for many equipments used in hazardous processes and this is performed by the use of explosion reliefs.
    This report describes about the effect of shape and size of vent area on the behavior of explosion pressures when they are vented from enclosed equipments operated under atmospheric pressure.
    Experiments are carried out using a cubical steel vessel of 1 m³ internal capacity, the upper side of which is opened and flanges with various vent areas can be bolted to it. Explosive mixture of propane with air is ignited at the center of cubical vessel. In another series of experiments hydrogen-air mixture is used. Vent cover is vinyl sheet and vent area is ranged between 1920 and 5760 cm^2. Pressure pick-up is wire straingauge type pressure transducer and pressure-time curves are recorded on electro-magnetic oscillograph.
    Main results are as follows :
  (1) For propane-air mixtures (5.0-5.5 vol.% in air), records of explosion pressure show distinct two peaks. First peak pressure (p₁) decreases with the increase of vent area, but second peak pressure(p₂) remains almost constant. The ratio p₁/p₂ decreases with vent area, and when vent areas are larger than 3600 cm² p₁ is smller than p₂, but for smaller vent areas than 2500 cm² a reverse relation occurs.
  (2) For hydrogen-air mixtures (hydrogen concentrations less than 30 vol.%), carried out to obtain information on the behavior of explosions of rapidly burning gases, only one peak pressure (corresponds to ₂) is seen for rather large vent areas. For 60 cm square vent area, uncertain first peak occurred. First peak pressure varies with vent area and the ratio p₁/p₂ is always less than unity.
  (3) The effect of the position of vent area on explosion pressure is not so evident as to be expected. But further experiments must be carried out to ascertain about this effect from practical point of view.
    These results are referred for the design of explosion relief for a certain type of commercial dust collector. Applied experiments for this collector are performed using propane-air mixture and flammable dusts such as epoxy-resin and aluminium powder. These experiments show that the existence of collecting filters and the material of vent cover are important factors for designing explosion reliefs.

Electron Fractographical Study on th Causes of Metals Failure (1std Report) --Observation of the Fracture Surface of Some Metals by Electron Microscope--

RR-19-3

Eiji AKIYAMA, Taiji KONDO and Yoshio KITUNAI

: The knowledge of the causes of metal failure becomes the only effective way to prevent following accidents. But in the case of the service failure of machine or structural members, it is very difficult to determine the causes of destruction.
    Generally, the fracture surface of broken parts display markings which constitute a topographical map, and often reveal the history of events preceding the failure. For a long time visual and low magnification observation of fracture surfaces have aided engineers in their fracture analysis.
    The electron microscope has a large depth of focus, a high resolving power and a large range of magnification. From the observation of fracture surfaces by electron microscope (Electron Fractography or Fractomicrography), some complex fracture causes may be more easily discriminated.
    The main purpose of this experiment is to discuss fracture mechanism and to clarify the accident causes from microscopic features on the fracture surface in service.
    Tensile, Charpy impact and fatigue tests were carried out and the fracture surfaces were observed respectively by electron microscope, in order to obtain some characteristic patterns resulted from several fracture modes.
    The test materials were pure iron, carbon steels (SS 41, S 15CK, S 35C), stainless steels (SUS 27, SUS 38) and cast iron (FC 15).
    Specimens for the electron microscope observation were made by two stage carbon replica method.
    Further some examples of machine parts broken in service were observed by electron microscope and compared with test piece.
    The results may be summarized as follows.
  (1) The tensile fracture surfaces of ductile materials (pure iron, SS 41, S 15CK, S 35C, SUS 27, SUS 38) were characterized by large and uniform dimples.
    However, fracture surfaces of cast iron revealed no characterized features by their fracture modes.
  (2) Charpy impact fracture surfaces depended on test temperature. River pattern was obtained on the cleavage fracture region below the ductile transition temperature, on the other hand, dimple pattern was exhibited on the ductile fracture region above the fracture transition temperature.
    Also both river and dimple patterns were seen on the region between the ductile transition temperature and the fracture transition temperature.
  (3) Striation and Cliff pattern were observed on the high cycle fatigue fractures of ductile materials.
    Such striation pattern was one of the most distinctive features found on many fatigue fracture surfaces. Particularly, fine striation pattern was found in fractured specimens at the stress near the fatigue limit.
    In the case of fatigue test under a constant stress level, the striations were formed vertically to the crack growth direction and the spacing was locally uniform in size.
    For the whole surface, however, the spacing of the striations tended to increase in proportion to stress amplitude and crack length.
    In this experiment the spacing of the striations were observed varying between 0.1 μ and 0.5 μ.
    On the other hand, dimple pattern was observed on the low cycle fatigue fracture surfaces in carbon steels, but the striation pattern which usually appeared in high cycle fatigue fracture surfaces was not found.
  (4) The fracture pattern of a few machine parts, i.e., hook, bolt, joint plate, pin, well coincided with these of test pieces experimentally made in our laboratory.
    As the result of these experiments, it was recognized that Electron Fractography is a very useful tool in the analysis of service failure.
    But the proper analysis with this technique depends on the preservation of fracture surfaces.

Energy of Acrivation for Oxidative Pyrolysis of Several Plastics

RR-19-4

Takashi KOTOYORI

: For the purpose of estimating the behavior of plastics for heat and oxidation, applying so-called Kissinger's method, which is well-known in DTA techniques on endothermic pyrolysis, to exothermic pyrolysis reaction, energies of activation for oxidative pyrolysis, besides frequency factors and specific reaction rates of Arrhenius equation were determined on ll representative thermoplastic resins in oxygen, air and nitrogen at atmospheric pressure.
    The instrument used was Thermoflex type 8021, a differential thermogravimetric analyser, developed by Rigaku Denki Co. Ltd. All samples were kept for more than 2 weeks in a desiccator, in which a saturated water solution of K2C03·2H20 were present. This solution is said to be in equilibrium with a relative humidity of about 40 %. As combustibles must come into contact with oxygen current in the case of oxidative pyrolysis, samples were not mixed with reference substance, but only laid on it. α-Alumina was used as reference, being ignited for 30 min in a crucible before every run. About 500 runs of DTA procedures were carried out on ll kinds of materials. Kissinger's plots obtained in each atmosphere are displayed in Fig.7.1 - 7.3. Table 3 shows calculated values of energy of activation E, logarithm of frequency factor ln A and specific reaction rate k based on these plots.
    Main conclusions gained are as follows :
  (1) Resins containing hydrogen as a-constituent, in general, give rise to dehydrogenation(scorching) reaction and begin to decompose at a temperature little less than 200 °C in the presence of oxygen. Their E values are mostly ranging from 10 to 20 kcal/mol.
  (2) With regard to ln A, so-called compensation effect is observed that ln A increases with an increase in E. Furthermore, it is concluded owing to this effect that specific reaction rates of pyrolysis of most polymers are nearly equal at temperatures at which individual pyrolysis takes place.
  (3) As factor which affects the rate of pyrolysis, one should mainly consider oxygen concentration rather than specific reaction rate.
  (4) In the case of such materials having very low vapor pressures at elevated temperatures as plastics, it has been revealed that an occurrence of ignition is not due to the process that proceeds through stages of pyrolysis followed by evolution of combustible gas and explosion, namely, the general and conventional concept of ignition process, but due to an initiation of oxidation (behydrogenation) reaction and self-heating phenomena on the surface of molten polymer.
  (5) with plastics, normally supposed to be highly hydrophobic, coexistence of trace amounts of water, contrary to usual sense, be able to play a part in lowering energy of activation for oxidative pyrolysis. However, specific reaction rate, the essential value, is evidently larger with drier than wetter sample.
  (6) As it has definitely been indicated that the value of energy of activation expresses pertinently the relationship between molecular structures of the materials and their behaviors to heat (oxidative pyrolysis), there is no doubt that this value would become an important guide in the search for heat-registant polymers.

The Protective Effect of Safety Guards made of Steel Plate

RR-19-5

Eiji AKIYAMA and Soichi KUMEKAWA

: Safety guard is most useful to reduce the accident in rupture of grinding wheel.
    But reports of this type of study have scarcely been published to date.
    Then we observed broken pieces of grinding wheels collided against safety guards by a high speed moving picture camera and investigated the protective effect of safety guards made of steel plate.
    The test pieces and the arrangement of this experiment are shown in Table 1, and the results are as the following,
  (1) Generally, when a rotating grinding wheel broke, it splits into some pieces of various size and they collide against the safety guards.
    We observed the state that the crack produced across a diameter of the grinding wheel and split up two pieces as Fig.3.
  (2) A tongue showed a effective proof for broken pieces.
    We, analyzed time that broken pieces flied from appointed position to arc AB as Fig.5.
    The result are as Table 2.
  (3) We measured the permanent and the temporary deformations of safety guards in the direction of radius shown in Fig.6.
    The result for the permanent deformation are shown in Fig.7, and the results for the temporary deformation are shown in Fig.8 (a), (b), (c).
  (4) We researched of the relation between the deformation of safety guards in the direction of radius and static load when pulled with a pulling speed of 1 mm/min.
    The results are shown in Fig.9 (a), (b), (c).
  (5) Fig.10 shows the static load equivalent to maximum temporary deformation.
    So the shock test of a guard is converted static test.
    And, within the limits of this experiment, the maximum static load is about 980 kg at mark a.

Safety for Electrical Equipment under Artificial Environments --Low Voltage D.C. Spark Iginition of Atmosphere CH4 - O2 and CH4 - N2O Mixtures and of Solid Combustibles in Oxygen Atmosphere--

RR-19-6

R. TANAKA and N. SUGAWARA

: Artificial environments having various compositions, concentrations and pressures are being introduced in the field of human life and activities.
    Recently noticeable are those in spacecrafts, hyperbaric chambers in hospitals and deep sea diving equipment.
    In most artificial environments, flammability or ease of ignition of flammable materials increases under the influence of enriched-oxygen and/or elevated partial pressure of oxygen.
    In this paper the authors reveal the most easily ignited concentration of CH₄ - O₂ or CH₄ - N₂O mixtures, and the minimum igniting currents in low voltage d.c. inductive, resistive and capacitive circuits for atmospheric CH₄ - O₂ or CH₄ - N₂O mixtures are determined, compared with those for CH₄ - air and H₂ - air mixtures, using the IEC-type spark producer. The results are given in Figs.4, 5 and 6 corresponding to each circuit respectively. It was confirmed that CH₄ - O₂ mixtures was lower in m.i.c. than H₂ - air mixtures. In the case of CH₄ - N₂O the m.i.c. was ranked somewhat in the middle of CH₄ - air and H₂ - air in the inductive or resistive circuits, however in the capacitive circuits it held a position a little lower than H₂ - air mixtures.
    In reference to the m.i.c. of CH₄ - O₂ mixtures, examinations are given for the mixtures as substitute for ignition test gas in intrinsically safe circuits in hazardous environments.
    The authors present also the m.i.c., using copies of British Break-flash No.3 Apparatus and Intermittent Break Apparatus, for solid combustibles in oxygen of pressure of 0 and 0 kg/cm² in gauge. The samples are Japanese paper, vinylchloride film and cotton cloth. Results obtained with inductive sparks are shown in Fig.8 and those with resistive in Table 5. It is concluded that spark ignition of solid combustibles in oxygen is comparable with that of atmospheric CH₄ - air mixtures.

JNIOSH