Abstract of Special Research Report (RR-21)

National Institute of Occupational Safety and Health, Japan

Interruption of Explosions by Flame Arresters (First Report) --On the Quenching Ability of Sintered Metals (1)--

RR-21-1

Toshihiro HAYASHI and Heizaburo TSURUMI

: When a flammable gas-air mixture within certain concentration limits is ignited in a rather long enclosure, it follows initially a slow combustion, and then is accelerated to a deflagration. The deflagration may develop finally to a detonation which is capable of resulting in a serious disaster. To exclude such potential hazards, the explosion should be interrupted as early as possible after the ignition.
    The usage of the flame arresters is a typical method for this purpose of protection, and the theoretical background is dependent on the flame quenching phenomenon in a porous body with narrow passages through which explosion flames are unable to propagate. Although various kinds of flame arresters have been developed, most of them are manufactured and used rather empirically. And from the view point of safety their experimental data are not satisfactory.
    This report describes about the flame quenching ability of sintered metals as constructional elements of flame arresters. The sintered metals tested were commercial filters, and were discs of 2 mm thick with outer diameter of 40 mm. Two kinds of metals, bronze and stainless steel, were tested and proto-shape of particles before sintering was nearly spherical for the former and quite irregular for the latter. For the purpose of this study the sintered metals were specified in terms of filtration diameter, which was generally defined as a minimum diameter of a spherical particle which could not be filtrated through a porous sintered body. The filtration diameter had a range from 120 μ (0.12 mm) to 10 μ, and these values were assumed to be proportional to the proto-particle sizes.
    The disc under test was fitted tightly into a flange (i.e. a mounting flange) and bolted between the end flanges of steel pipe enclosures. One enclosure was "explosion chamber" and the other was "protected chamber". In each mounting flange an orifice was so drilled that the effect of the orifice on the explosion transmission could be determined. The largest orifice diameter was equal to the internal diameter of 1" gas pipe (28 mm), and the minimum was 2 mm. After setting up, the whole assembly was evacuated, and then filled with premixed hydrogen-air mixture at a desired initial pressure. An explosion was initiated in the explosion chamber by a spark plug, and the pressure changes in both chambers were recorded as pressure-time oscillograms. Whether the explosion flame had been quenched in the sintered metal or transmitted through it was shown distinctly by these records.
    For the first series of tests the effect of dimensions of the explosion chamber on the flame quenching was studied, and therefore hydrogen content in the mixture was kept constant(i.e. 30 % by volume in air) ; this was a stoichiometric value and considered to give the fastest speed of flame propagation. The diameter of the explosion chamber was either 1" (28 mm) or 8" (200 mm) and the length was changed in different ways. Initial pressure range was from atmospheric to 2.5 kg/cm² (gauge), with 0.5 kg/cm² steps. The initial pressure which gave ten successive quenchings was denied as limiting safe pressure (L.S.P.) at that condition. The L.S.P., which showed a relative degree of safety from the standpoint of the flame quenching ability, was proportionally increased as the filtration diameter or the orifice diameter decreased, and if compared in terms of L.S.P. the sintered bronzes were shown to be less effective than stainless steel discs for the same filtration diameter. This was probably because of the differences of the particle shape and the method of sintering. The ratio of the length (L ) to the diameter(D ) of the explosion chamber had a considerable influence on the flame acceleration and therefore on the flame quenching phenomena. The results showed that, for a constant D, increasing the L / D gave more dangerous explosions, and when an detonation-like explosion propagated against an arrester, the disc under test was usually fractured or deformed because of a rapidly applied pressure.
For a constant L / D, the larger the pipe diameter the more easily the explosions transmitted into the protected chamber.
    In the other series of experiments hydrogen content was varied between 10 and 60% by volume, whilst the test enclosure assembly was fixed to that of 1" pipe. It was shown that for bronze discs of 120 μ the minimum L.S.P. was given at the stoichiometric concentration, but that for 100 μ bronzes the minimum was obtained at a little lower concentration. For those of smaller filtration diameters and for stainless steel discs, the most dangerous mixture was nearly 20 % hydrogen content. It might be said, therefore, that the speed of flame propagation was not the only predominant factor, but the aerodynamical movement in flames and/or in unburned gases, produced when passing through a sintered metal and entering the protected chamber, had an important influence on the flame quenching.

Experimental Study on Micro-Seismic Noises in Rocks (2nd Report) --Characteristecs in the Generation of Micro-Seismic Noises in Rhyolitic Tuff (Oyaishi)--

RR-21-2

Ikuo MAE and Yoshimi SUZUKI

: There is a fact that rocks in instability under stress generate micro-seismic noises (a kind of elastic shock wave) with micro fractures.
    It is expected that micro-seismic noise detection is an effective method for predicting accidents such as roof falls and pillar failures in quarrying works.
    In order to apply micro-seismic method to the abovementioned purpose, the necessary studies of fundamental facts are being made in the laboratory and in the quarry.
    The present paper describes the results obtained from the experiments of the micro-seismic noise activities during the course of failures of the specimens under increasing load, sustaining load and repeating load.
    The several results are summarized as follows.
    The increasing stress is applied to specimen, occurrence of m.s. noises increases immediately after the stress application, thereafter activity of m. s. noises reduces at the middle stage, however it again begins to increase abruptly before the failure.
    The pattern of occurrence of m.s. noises under increasing stress seems to be related to the properties of rhyolitic tuff.
    The sustaining constant stress is applied to specimen, there is no high micro-seismic activities at the early or middle stage, and preceding to the failure m.s. noises occur frequently.
    Under these stress condition, the frequency of m.s. noise depends on the increment of strain in rock.
    The frequency distribution of number of the m.s. noise with respect to the energy possesses a statistical regularity which is expressed by the exponential equation.
    There is a relative relationship between the distance from the source of m.s. noises and attenuation, of them through the rock.

On the Dynamic Loads Imposed on Crane Structures --Shock Loads at Load Lift Up--

RR-21-3

Teizo HAKAMAZUKA, Yoshimasa KAWAJIRI and Soichi KUMEKAWA

: It is generally supposed that some causes of crane accidents are breaking of hooks or similar tools. And the failures occur by shock-load and fatigue. To prevent these failures, the observation of the load behaviours are necessary.
Standing of the view, we measured the load behaviours in overhead-travelling crane practically. The measurements are on strains or accelerations of crane-structure and load which is instantaneously lift from floor up. On the other hand, in carrying out analysis to these phenomenum, an analogue computer was used. That is, we consider the load lifting behaviour as mechanical system which have damped. 2 mass points and 2 springs. One spring is non-linear, another is linear. [Fig.3]
The equilibrium of forces yields the equations as follows.

where m₁ = mass of load
        m₂ = equivalent mass of trolley and girder
        x₁, x₂ = vertical displacement of load, trolley
        x₀₂ = predisplacement of x₂
        φ = rotated angle of motor
        r = radius of winding drum
        n = reduction ratio of pulley block
        f = reduction ratio of gears
        k₁', k₂ = spring constant of wire-rope system, girder block (at center)
        E = force for floor to support the load
        c₁, c₂ = coefficient of viscous resistance
    The result of simulation are very similar to the practice measurings.

Studies of Explosive Chracteristics of Hydrogen (2nd Report) --Explosion Pressures of Hydrogen-Air Mixtures at High Pressures--

RR-21-4

Shozo YAGYU and Toei MATSUDA

: In many cases, hydrogen is industrially used at elevated pressures. The previous paper dealt with the effect of high pressures on the flammability limits of hydrogen. The present paper describes the effect of high pressures on the explosion pressures of hydrogen-air mixtures.
The explosion pressures, times to attain maximum explosion pressure and average rates of pressure rise were measured at room temperature and pressures from atmospheric to 50 kg/�p² to specify explosive characteristics of mixtures of hydrogen and air. The experimental apparatus used is shown diagrammatically in Fig.1. The mixtures were ignited at the bottom of a 7.5 cm diameter and 50 cm height cylindrical bomb. The explosion pressures were measured by strain gauge transducer and recorded on two-channel oscillograph.
The results showed that the ratios of maximum explosion pressure to initial pressure are nearly constant for the same Concentrations of hydrogen in the mixtures, and that the maximum value of the ratios is about 7.5 when the hydrogen concentrations are 30-35 per cent. Near the upper and lower limits of flammability of hydrogen-air mixture, the ratios are 3.0 and 1.0, respectively.
Hydrogen Concentration is about 10 per cent in lower fuel side, which is equivalent to the ratio of 3.0. The relation between hydrogen concentrations and times to attain the maximum explosion pressure corresponds to that of explosion pressure ratio. The pressures increase relatively slow in the region less than 10 per cent of hydrogen, but develop rapidly within 100 ms beyond that concentrations. As above mentioned, the explosion pressure is determined by multiplying the initial pressure at a certain concentration. This means that the average rates of pressure rise enlarge with the increase of initial pressure.