Abstract of Technical Note (TN-77)

National Institute of Occupational Safety and Health, Japan

Method of Temperature Measurement of Explosion-protected Electrical Apparatus

TN-77-1

Katsuhiro SAKAMUSHI

: The present paper is concerned with the measurement of temperature of stationary part of explosion-protected electrical apparatus by using a thermocouple.
    In order to fix the thermocouple on the measuring part, the following three methods have been examined on the standpoint of heat-resistivity, reproducibility of data and applicability :
  (1) Method by organic or inorganic adhesive.
  (2) Method by heat-resistant tape.
  (3) Method by mechanically pressing. (A metal fragment is brazed to the tip of thermocouple, and the fragment and thermocouple is pressed on to the measuring part by means of a cylinder type compressed spring balance.)
    On the heat-resistivity, the organic adhesive can be used at the temperature lower than 100 °C ; and the heat-resistant tape, at the temperature lower than 150 °C. The inorganic adhesive and the mechanically pressing methods are appropriate for the measurement of temperature lower than 300 °C.
    For each method given above, a good reproducibility of data would be obtained if sufficient care should be taken according to the requirements inherent to the method.
    On the applicability, the method by heat-resistant tape is the most advantageous of the three because only thing to do is to cut out a length of the type and apply it on the thermocouple. The others are also practicable although they need preliminary treatments before their application.

Experimental Investigation into the Cause of an Explosion in the Flue Duct of an Oil-Fuel Boiler at a Certain Building

TN-77-2

Kinichi KINOSHITA, Michio NAITO, Noboru SUGIMOTO and Kiyoshi FUKAYA

: At about 10 p.m. on 30th November 1976, an explosion occurred in the flue duct of an oil-fuel boiler at a certain building in Tokyo. The explosion was so intensive that the whole wall between the duct and the rooms upward from the 4th to the top, 9th floor was demolished part by part. Of those who had been customers of stores in the complex building at the time, 25 suffered injuries from the scattered building materials.
    The structural shape of the flue duct section was quite unusual, as shown in Fig.1 and the maintenance of a combustion controller had been insufficient. Frequent back- or miss-fires had been also experienced. From the survey in the disaster site, we assumed that the unfired gaseous mixture, resulted from the failure of ignition control, had not exhausted smoothly to the outdoors and that these mixtures stagnated in the high combustible concentrations. The combustible gaseous mixture would be formed from decomposition of fuel oils in high temperature furnace with air. Also, the gas flow would have been disturbed near many girders within the flue duct.
    The aim of this investigation was to confirm this hypothesis and to discuss the possible cause of an explosion with experimental results. The experiments were carried out with use of two model ducts, of which scale were 1/10 of the actual ones. One of models has a recommendable shape in the construction of the flue duct system and the other was relevant to the accident disaster. Instead of determining the concentration of the unfired decomposed gases, we measured optically the smoke concentration from a smoke candle, which is widely used as a simulator of gas flow.
    From the experimental results, we could verify the above hypothesis. Also, it was elucidable that it would be possibly attained the lower explosible limit if the fuel gaseous mixture flowed into flue duct continuously and accidentally more than two times. Then, it would be not necessarily impossible to have an explosion in the flue duct in the presence of suitable ignition sources, like the flame from successive trials of ignition in the boiler and the propagating flame.

Arc Resistance of Solid Organic Electrical Insulating Materials in Pressurized Nitrogen Atmospheres (The First Report)

TN-77-3

Ryuji TANAKA and Tatsuo MOTOYAMA

: Electrical equipment is sometimes installed in artificial environments which differ in composition and/or pressure from the atmospheric air.
    In the present paper, the authors investigate into the high-voltage, low-current arc resistance of solid organic electrical insulating materials in pressurized nitrogen atmospheres.
    The test was carried out according to the method described in the JIS K 6911 or the ASTM D 495, except the atmospheric condition.
    The test apparatus consists of a pair of arcing electrodes which are inserted in a pressure vessel, and electrical networks that operate, in sequence, to supply the electrodes with pre-determined arcing currents.
    The atmosphere employed was nitrogen gas whose pressure was elevated up to 0.6 MPa, but no influence due to pressure-rise was recognized on the change of the stray capacity between the electrodes.
    The materials tested are melamine-glass laminated sheet, phenolic resin, ABS resin and polycarbonate.
    With the increase of nitrogen pressure, the arc characteristic changes ; that is, while the discharge threshold voltage becomes higher, the stable arcing time becomes shorter. The former gives a high density of heat to accelerate the deterioration of material near the electrode, and the latter works to generate less heat to decelerate the deterioration.
    On the other hand, the materials are considered to have their own phase of heat-deteriorating temperatures, depending upon their composition, combination, etc.
    The test results show that with the increase of the pressure, the arc resistance of the materials
  (1) decreases, but increases thereafter, giving a U-shape curve, or
  (2) continues to decrease, giving a L-shape curve.
    The authors conclude that the results obtained are influenced by the discharge characteristics, the heat deterioration of materials and inert nitrogen gas.

The Vibration Measurement on Some Leg-Type of Rock Drill

TN-77-4

Teizo HAKAMAZUKA and Yutaka MAEDA

: It is known that occupational diseases appear on hands, arms or shoulders of workers who handle portable vibrating tools such as rock drills or chain-saws.
    To prevent the diseases, diminishing the vibration is one of the most effective methods.
    In order to get the fundamental data for development of the vibration arresting technic, some measurements of vibration generated from rock drills of leg type were carried out. 14 types of drills (ten types with usual handles, four with vibration arresting handles) were subjected in this experiment.
    The drills were operated in a certain metal mine. And operating conditions of tools were as following ; wet drilling with single type of rods and bits, drilling depth up to 120 cm, setting leg's angle of 45 degree to horizontal level, keeping rods horizontal, supplied air pressure of 5 - 5.5 kg/cm² . And one skilled worker operated all of the tools.
    Three vibration acceleration pickups of electro-strictive ceramic type were fixed on the handle of each tool in the directions of three axes. Vibration acceleration signals generated from the tool were recorded in a portable magnetic tape recorder and at the same time sound level was measured close by the worker's ear.
    Later, the acceleration data were played back and inputted in a digital computer. They were processed by the method of FFT as 1/3 octave frequency analysis.
    Then they are ; acceleration level is about 145 - 150 dB on usual handles, and about 130 - 145 dB on vibration arresting handles. In high frequency range, the acceleration magnitude of the latter handles is smaller than the formers', but in low frequency range such as 40 - 50 Hz that is not diminished.
Besides, the vibration levels were not almost influenced on drilling depth, speed, and tool weights in this experiment.
    Sound level in working is 113 - 118 phones (A band).

Research on Flame-and Detonation-Arreters (2nd Report) --Effect of Pipe Layout on Detonation Induction Distance--

TN-77-5

Toshihiro HAYASHI and Hidenori MATSUI

: It has been well known that a flame originated in a long pipeline containing detonable gas mixture could transit to detonation after travelling some distance. The distance between a source of ignition and the point, where an transition of combustion wave to detonation occurs in the pipeline, is called "detonation induction distance" (D.I.D.). Although D.I.D. is known to be affected by such factors as the composition, temperature and pressure of gas mixture, and the length, diameter and smoothness of inner surfaces of pipeline in which the flame travels, available informations are not yet satisfactory from the practical standpoint of view.
    This report describes about effects of pipe layouts on D.I.D. Experiments are carried out in 1-inch pipelines, in which the stoichiometric mixture of hydrogen-air or acetylene-air, their initial pressure being atmospheric, is ignited by a nichrome wire heater. Speeds of flame propagation are measured by ion-probe method ; ion-probes are distributed along the pipeline with some 30 or 50 cm intervals. Average flame speed between two ion-probes is defined as the speed of flame at the mid point of those probes. Thus the relation between the distance from ignition source and the flame speed for each point can be obtained. Then, D.I.D. is arbitrarily defined as the distance between the ignition, source and the point where flame speed attains 1,000 m/s on the distance-flame speed diagram.
    Effects of following factors on D.I.D. are discussed ;
  length of the straight pipeline, up to 9 meters ;
  whether an ignition end and/or the other end are closed, partially closed or open to atmosphere ;
  direction of flame propagation, i.e. horizontal, upward or downward ;
  position of ignition source in the pipeline ;
  existence of such fittings as bend or tee near the ignition source ;
  existence of the spiral wire as an obstacle in the pipeline.
    Discussions are also made whether detonation wave could be decayed when travelling through pipe fittings mentioned above.

Experimental Studies of Laminar Flame Propagation in Dust Suspensions (I) --Polyethylene Dust-Air Flame--

TN-77-6

Toei MATSUDA

: Considerable data of characteristics of dust explosion hazards have been accumulated and a large amount of information is now available on the subjects. However, a fuller understanding of the flame propagation process in dust suspensions is still insufficient and needed for predicting more accurately the explosion behavior in the dust handling industrial plant.
    The propagation of laminar dust flame is studied experimentally, using polyethylene powder as a fuel in the first place. Measurements of lower limits of flammability, flame speeds and gaseous products after combustion are made initially, and appearance of the flame is observed in order to examine the nature of dust flame propagation. Then, a Mach-Zehnder interferometer and schlieren techniques are applied to the propagating flame with employment of high-speed cine photography.
    A vertical tube of cylindrical glass about 2m long and 75 mm in diameter was used for observation of upward propagating flames. For the steady flame propagation, the top of the tube was closed, leaving the end opened just before ignition.
    The relationship between the lower flammability limit, (L mg/l) and the mean dust falling velocity at the limit, (V cm/s) was obtained as (V - 114)·(L - 0.7) = -1008. The flame speed was measured as function of dust concentrations and particle sizes, and appeared to relate to the ratio of the experimental to the calculated carbon dioxide contents in the combustion products. The limiting flame speeds at the lower limits seemed to remain essentially constant for the particles greater than about 125 μm, but with decreasing particle sizes decreased to the value of premixed gaseous flames.
    Observation of the flames at near-limit showed a change of flame appearance with dust particle sizes. In case the mean particle diameter was at least below 90 μm, the flame propagated discontinuously with burning individual or groups of particles of a faint blue color at the front, usually following with burning groups of particles, yellow shinning. At the diameter around 125 μm, the flame was like a coherent yellow flame which contained burning individual or groups of particles, scatteringly in itself. The flame appeared to be composed of the dark, but also coherent flame head and the yellow trails of flame separate from the front, above the diameter around 180 μm. These flame appearances also transferred to the longer flames with increasing dust concentrations above the lower limit of flammability.
    The optical records indicated that any change of falling dust particles was undetected ahead of the flame, and that the pyrolysis of particles could occur in the preheating zone for a blue flame appearance, and in the flame itself for large part of particle sizes. The particles mostly kept falling into the flame at their speed unchanged, and in case the particle sizes were not so large, a small number of particles was seen to reverse their direction upwards in the flame.
    On these observation and experimental results, the flame propagation mechanism was considered qualitatively as follows.
    When the particle size is small below 90 μm and the dust concentration is near the lower limit, the heating will be enough for some particles to melt and oxidative-pyrolyse in the preheating zone of the approaching flame from the bottom, as the velocity of falling dust is comparatively slow. However, the pyrolysis occurs almost completely near, or at the particle surface, owing to the small size particle, and the localized pre-mixture of pyrolysate and air may be formed. Then, the appearance of blue sparks is observed with ignition at the flame front. The flame propagation is likely to be supported with buoyancy of the hot exhausted gas. With raise of the particle diameter and dust concentration, the pyrolysis takes place more rapidly in the flame, and the process will be seen of the upward diffusion of the pyrolysate into the preheating zone, rich of fresh air, accordingly the whole bright diffusion flame being observed. Moreover, as the particle sizes, or the dust concentrations are increased more than the above case, the pyrolysis of the particles becomes not to be sufficient in the flame head, due to less heating, but soon to be fast in the flame inside, thereby the bright trails of burning particles being obtained. For this kind of flames, the more thicker flame front, or flame head could be given if the heat flux from the lowerside diffusional flame is attained with higher dust concentration. In conclusion, the mechanism is affected by the rate of pyrolysis of the particles and the dust sedimentation velocity.

Experimental Studies of Laminar Flame Propagation in Dust Suspensions (II) --Cork Dust-Air Flame--

TN-77-7

Toei MATSUDA

: The previous paper described the observation of flame propagation in suspension of polyethylene dust which, chosen as a representative industrial dust having relatively simple chemical composition and structure, was dispersed in a vertical tube, 75 mm in diameter. In this paper, experimental results on the propagation of laminar cork flame in the same tube with open-end ignition and the top closed is reported and then its propagation mechanism is examined by comparison with that of the above dust. Cork dust is of high flammability and high dispersability, and R.H.Essenhigh et al. have reported the nature of flame propagation in clouds of the cork dust showing some distribution of particle sizes (1958).
    Here, the dust was sieved strictly into five samples, of which mean particle diameters were, respectively, 90, 125, 180, 255 and 360 μm. Thermal gravimetric analysis of the dust samples showed 65.7 % weight loss and 34.3 % carbon residue on heating to about 600 °C in nitrogen atmosphere. The lower limit of flammability (0.037 mg / l) and the limiting flame speed at that limit (about 40 cm/s) were ascertained to be independent of dust fineness. This characteristics is contrary to the general tendency for the flammability of dusts to increase as the particle sizes become smaller. Then, measurement of the mean dust falling velocity on the limit was made at the base of the tube, and their values were shown to be within a relatively narrow range for these samples. This may imply that the propagation mechanism is common to all the samples at the limit.
    Direct photographs of near-limit flames propagating upwards showed many independent spherical diffusion flames of individual particles, irrespective of dust fineness and except for the sample of mean particle diameter 90 μm formed agglomeration, and that the flame length was between about 2 and 10 cm. The burning process of a particle at the flame front could be followed on cine-filming with application of a Mach-Zehnder interferometer. It showed the two stage burning for individual particles corresponding to the first stage of the combustion of volatile matters and pyrolysates, and then the second stage of the surface combustion of carbon residues of the particle. The increasing ratios of the maximum visible flame diameter to initial particle mean diameter were obtained with decreasing particle diameters for the flame of a single particle near the flame front. They appear to be comparable with that of combustion of fuel drops in suspensions.
    Photomicrographs of the particles quenched on the tube wall while burning indicated that the material around the particle contained needle-like crystallines in some cases. They were probably celin or friedelin involved in the cork, as predicted from TGA.    The motion of particles ahead of the flame was analysed in the records on films of silhouettes of unburned particles against the back-illuminating light, where the distribution of particle velocities was gained for downward and upward movements. As expected, the ratio of number of particles travelling upwards to that for falling down was increased with increasing dust fineness at the limit, and also dependent on flame speed, or dust concentration. The particles moving up were likely easy to ignite in the flame front, as they would have to reverse their direction of falling for upward propagating flames. These movement of particles could be attributed to the flow of air in front of the flame, similar to that of laminar propagation of gaseous premixed flame in a tube.
    From the above observation, the mechanism that the flame propagates discontinuously from one particle to the other is considered to apply to cork dusts of all the particle sizes near the limit.
However, the burning particles became to merge forming a continuous yellow flame front with separate burning particles in the coherent underside of the flame, with increasing dust concentration above the limit for the particle sizes over 180 μm. This shows that the flame propagation mechanism transfers to that of the whole upward diffusion flame, like that observed previously in one of the polyethylene dust flames.

Research on Flame and Detonation-Arresters (3rd Report) --Detonation-Arrester Model Composed of a Perforated Plate and Wire Gauzes--

TN-77-8

Toshihiro HAYASHI and Hidenori MATSUI

: It has been well known that a flame originated in a long pipeline containing detonable gas mixture could transit to detonation after travelling some distances. Flames should be quenched as earlier as possible after their origin, but in many cases safety devices are required to come against detonations. Such devices are called as detonation arresters.
    In this report, studies were made on quenchings of acetylene-air detonation waves travelling in 1-inch pipes, and some models of detonation arrester were proposed. Because of their simple design and low pressure drop, those arresters, composed of a perforated plate with narrow holes and a layer of 100 mesh stainless wire gauges, might be used in practical fields of application.
    At first, direct quenching of detonations by wire gauges only were tested, and then the effects of the perforated plate inserted in test pipe ahead of wire gauges were studied. To decrease pressure drops across arresters, the diameter of arrester casing, which contains a perforated plate and wire gauges, were enlarged up to 4 inches.
    For atmospheric initial pressure, detonations were quenched directly by 30 gauges and the quantity of gauze required for quenching was independent of casing diameter. When a perforated plate was inserted, such factors as the diameter and number of holes drilled in the plate, the distance between the casing entrance and the perforated plate, and the distance between the perforated plate and the layer of wire gauges affected largely on detonation quenching behaviours. It was also found effective for quenching to blind some part of holes near the center of the plate so that predominant waves could not pass through the plate. For 4-inch casing with a perforated plate with 300 holes of 3 mm diameter, detonation of atmospheric initial pressure was quenched by only 15 gauges, and the corresponding pressure drop was below 10 mmAq for air flow of 100 l/min.

A Study on Pendant Supported Crane Boom's Bucklimg Strength out of Plane and an Accident Analysis of a Crane Boom

TN-77-9

Yutaka MAEDA and Teizo HAKAMAZUKA

: About pendant supported latticed crane boom, several modes of buckling are to be considered, i.e., every chord member's. buckling, buckling in the plane Which contains boom's axis and vertical line, buckling out of the plane, and so on.
    Buckling out of plane has been little considered up to resent days in Japan. A buckling of boom of a truck crane took place.
    The outline of the accident was as follows ; a load was to be lifted by a truck crane of which maximum rated load was 127 tons. The load was estimated about 27 tons in weight and boom length was intended to be 54.86 meters. But the actual load was heavier than that of estimated, and booms were jointed by mistake to be longer than that of intended. And when the load was lifted, the boom buckled.
    In this case, the operator made several mistakes, however, crane boom was expected to be proof against these trivial errors, therefore, the structural competence of the boom was reconsidered, and it was proved that the boom's strength out of plane was not sufficient to support such overload, as this case.
    Buckling experiments of miniature booms were carried out by use of steel booms of which size was 5 × 10 × 500 (mm). Boom angle was about 35 and 42 degrees. The results were approximately equal to the calculated values from following equations ;
      W_cr = P_cr sin(θ₁ - θ₂)/ sin (θ₂ + π/2 )
      P_cr = (kl)² EI/l²
      tan(kl) = kl(1 - c/l)

Lime Treatment of Slay Minerals

TN-77-10

Ikuo MAE and Yoshimi SUZUKI

: For considering the safety of various methods of soil stabilization or chemical grouting, the investigations from chemical and geochemical points of view about these methods are necessary.
    In this study, in order to get the fundamental data of chemical changes about lime treatment of clay minerals, three kinds of clay mineral (kaolinite, montmorillonite and pyrophyllite) are chosen.
    Reagent grade Ca(OH)₂ was used for preparing lime-clay mixtures at few levels of lime content based on the dry weight of the clay.
    After addition the calculated amount of distilled water and kneading thoroughly, lime-clay mixtures were compacted to maximum density at optimum moisture content by the use of Havard Miniature Compaction Apparatus.
    Compacted specimens were sealed carefully to minimize moisture loss and carbonation during curing and they were cured at the room temperature or high (60 °C) temperature.
    Physical properties such as unconfined compressive strength of compacted specimens were measured and mineralogical changes of these specimens were followed by the X-ray diffraction analysis. These experiments were made for the each kind of clay, of lime amounts and of curing time.
    The results of these experiments may be summarized as follows.
  (1) Unconfined compressive strength of each compacted specimens of lime-clay mixtures were generally developed according to the length of curing time ; curing period for strength appearance was 2-10 weeks at room temperature curing and 2-4 days at high temperature curing. (Fig.5-10)
  (2) The amounts of lime in the lime-clay mixtures could be measured by means of the relative strength of diffracted X-ray. (Fig.11)
  (3) X-ray diffraction patterns for each lime-clay mixtures proved that the relative strength of diffracted X-ray for lime weakened according to the length of curing time and that reaction products were formed in kaolinite-lime mixtures at high temperature curing and montmorillonite-lime mixtures. (Fig.12-20)
  (4) These reaction products were calcium aluminate hydrate in kaolinite-lime mixture and calcium silicate hydrate in montmorillonite-lime mixtures. In pyrophyllite-lime mixtures, any reaction products was not detected. (Fig.12-20)
  (5) The decreasing amounts of lime in lime-clay mixtures had tendency to be constant having no relation to initial amounts of lime in the mixtures. (Fig.21)
  (6) The relations between the changes of unconfined compressive strength and the changes of relative strength of diffracted X-ray for lime were indicated as straight lines. (Fig.22-25) The point of relation between both characters was moved on the line to upper according to the length of curing time. This indicate the possibility of conclusions about the final strength of lime-clay mixtures and about the initial amount of lime or the curing time to obtain the requiring strength.

JNIOSH