Although the finite element method has come into wide use in the field of structural and continuum mechanics, its error characteristics are not well known. This paper discusses the error characteristics of the finite element method in the case of triangular elements with a linear shape function generally used in plane stress analysis. The paper analyzes the discretization errors of finite element method using the technique proposed by Walz et al., with some generalizations. Here approximations included in the calculations of finite element method are discussed with various element patterns in comparison with the equilibrium equations of the field. The analysis reveals that the approximations in finite element method do not converge to the field equations in almost all the element patterns checked here as the element size is made infinitesimal. From this point of view, it interprets what kind of element patterns should be applied. Errors caused by the approximate strain-displacement relation are also discussed by a similar technique.

View full abstract

An axial load impact fatigue testing machine was newly designed and manufactured, which makes it possible to carry out a series of tests ranging from the ordinary fatigue test to the test under loads including some amount of impact load. Using this machine, plain and notched specimens of 0.11% carbon steel were tested under fully reversed sinusoidal axial load and under impact fatigue load. It was observed that, under impact load, fatigue fractures occurred at the stress levels lower than the ordinary endurance limits for both of the specimens.

View full abstract

Using specimens of SS41 carbon steel, authors made a test on erosion by means of resonant frequency of 22 kc/sec, and simultaneously took accurate measurements of half-value breadth both in the surface and in its innerface through Cobalt Kα radiation and of residual stress. Authors' conclusion is as follows. Both in an erosive surface and in its innerface, residual stress is attributed to compressive stress, after maximum residual stress has taken place in the surface, the weight loss of specimen remarkably increases; consequently there can be recognized a certain time rate in this test duration. The internal residual stress before starting of the weight loss of the test piece being compared with the one after starting, there occurs some change from maximum to minimum nearly at the same depth. The comparative test on half-value breadth at different amplitudes results in that in each period of operation the testing amplitudes relate not to the pressure in the collapse of each one of the bubbles but to the number of bubbles in the collapse.

View full abstract

A reservoir supported horizontally and partially by saddles yields a very large deformation. Tolerable limits of the deformation may vary according to the working conditions of a reservoir. In the present paper, the authors tried to find the shape of a reservoir which does not yield too much deformation under partial support condition. Reservoirs having an oval cross section which is formed by three circular arcs are investigated. Stresses in a reservoir and its deformations are found as the solutions of the differential equations for cylindrical shells introduced by one of the present authors. Experimental study is made along with the theoretical analysis. The results show that there exists an optimum oval form which holds the deformation of a reservoir under desired level.

View full abstract

In this paper, we consider the problem of a semi-infinite thin plate, having a circular hole which is filled with an elastic inclusion of another material, subjected to axisymmetric loading along its straight boundary. The analysis is treated by applying bipolar coordinates with the help of perturbation. Numerical calculations about the stress distribution along the main parts of a semi-infinite plate and an inclusion are also carried out to clarify the effect of an inclusion.

View full abstract

A conception of the equivalent stress and equivalent plastic strain rate is necessary for comparison of the yield surfaces or the stress-strain diagrams, and for estimation of the amount of plastic deformation of the materials deformed plastically under combined stresses. In the theory of anisotropic plasticity based on the maximum shear stress hypothesis, the equivalent stress and equivalent plastic strain rate for work-hardening materials are theoretically developed in this paper with use of the increment of plastic work per unit volume. The conclusions of this paper give the theoretical foundation and a certain restriction to the dissertation published by I. Berman and P. G. Hodge, Jr.. It is assumed here that the principal axes of stress and anisotropy are coaxial.

View full abstract

The authors proposed in this paper the methods for measurement of two principal residual stresses in an orthotropic plate when their directions are unknown. The experimental technique is as follows : A pair of test pieces intersecting at right angles with each other are cut out from a plate in an arbitrary direction, then unknown directions and values of the principal residual stresses are calculated by the authors' formulas from the changes of curvature and twist of the test pieces (Curvature Method), or from the changes of strain of the strain rosette attached on one surface of a test peice (Strain Gauge Method), while the successive thin layers are being removed from one surface or the other. Measurement of an isotropic plate is a spcial case of the authors' present methods and they showed examples of measuring two principal residual stresses in a steel plate by the Strain Gauge Method.

View full abstract

Dynamic behaviour of a machine structure model subjected to two seismic motions one of which has certain time delay interval behind the other is investigated. At first response properties to earthquake for a simple mechanical model of a one-degree-of-freedom system are computed by analog computer. In this procedure it is necessary to obtain wave forms not only of acceleration but also of velocity and displacement. Hence an approximate method to integrate random wave is proposed and discussed. To examine this oepration the frequency characteristics of the transfer function are investigated. It is shown through this analysis that acceleration response amplification and the displacement response show fluctuating characteristics which depend on the natural period of the mechanical model and the time delay interval. Finally statistical investigations are made. Response characteristics by this statistical computation agree fairly well with those by analog computation.

View full abstract

The heat loss characteristics of the hot-wire probes placed normal to a horizontal wind are examined experimentally in the wind-speed range from zero to 150 cm/sec. The reproducibility of the characteristics is clearly shown at extremely low wire Reynolds numbers to which neither the empirical relation due to Collis and Williams nor King's law is applicable. Based on the examination, a linearized constant temperature hot-wire anemometer is designed and constructed which generates a voltage output directly proportional to instantaneous wind-speed from nearly zero to 150 cm/sec. Using the anemometer, practical problems related to the directional response and the effect of the proximity of a wall to the wire upno the reading are examined. An estimation is also given of the effect of velocity reversal upon the reading.

View full abstract

Noise generated by a fan consists of the rotating noise and the turbulent noise. These two kinds of noise have entirely different characters with respect to decay in an axisymmetric pipe. In this paper the generation and decay of these two kinds of noise are experimentally examined by changing various factors, such as the suction pipe geometry, the location and the thickness of struts, and the combination of the number of impeller blades and the number of struts. The rate of decay of the rotating noise is predictable and in most cases the noise quickly decreases in proportion to the distance from the rotor, providing that the number of impeller blades and the number of struts are properly chosen. Therefore, the noise observeda way from a low pressure axial flow fan is mostly the turbulent noise.

View full abstract

This paper deals with the surge pressure and pressure fluctuation in a pipe caused by the instantaneous operation of a valve in a hydraulic system made up of an accumulator, a directional control valve and a pipeline. Theoretical and experimental results indicate that when the valve attached to one end of the pipe is operated rapidly and a step change in pressure is applied to the pipeline, the surge pressure which is caused by the spool-type directional control valve is proportional to the spool velocity. Moreover it is shown that the period and the logarithmic decrement of the pressure fluctuation at the other end are influenced only by a dimensionless quantity defined as the square of the radius of the pipe times the natural frequency of the fluid column divided by the kinematic viscosity ; in other words, the pressure fluctuation is similar in cases where the values of this dimensionless quantity are equal to each other.

View full abstract

An experimental investigation into vertical upward air-water two-phase flow was performed using a smooth tube and three spirally grooved tubes. All the test tubes are about 13 mm in inner diameter and 8m in length. The experimental data on each test section include pressure drop, radial distribution of void fraction, cross-sectional mean void fraction, and mean film thickness in annular flow regime. The summary of the results is as folows : (1) The transition boundaries of flow patterns are very similar to those of smooth tube. (2) The pressure gradients do not always depend on the magnitude of relative roughness of test tube, but a ribbed tube, having the greatest relative roughness, shows the greatest values in pressure gradient. (3) The variation of grooves produces remarkable difference in radial distribution of void fractions. (4) Mean film thickness for grooved tubes become definitely greater than for smooth tube.

View full abstract

The condensation heat transfer of a binary liquid metal vapor with and without a non-condensable gas is studied experimentally. The liquid metal vapor, which is generated from a mercury-lead-bismuth alloy, is forced to flow downward in a vertical tube, the outside of which is cooled by water, oil or air. The radial temperature distributions in the tube wall and in the vapor are measured as well as the vapor pressure. The heat flux is found to be proportional to the temperature difference between the vapor and the wall surface. In the presence of a non-condensable gas in the vapor, the condensation heat transfer coefficient is found to decrease with an increase of the weight fraction of the non-condensable gas. The coefficient obtained experimentally agrees well with the theoretical one of single component vapor predicted by assuming a local thermal equilibrium in the vapor.

View full abstract

To determine ignition lag, burning rate and drag coefficient for a fine fuel droplet, behavior of an array of tetralin droplets (120∼170 μφ in diameter) falling freely into the hot air (800∼1000°C) in an electric furnace was studied by high speed photography. The following results were obtained. (1) Ignition lag is decreased remarkably by raising the air temperature, but it is little affected by the droplet size. (2) In the droplet array, a droplet falling ahead promotes the ignition of a following one. This effect is varied with the distance between them. (3) For the burning rate of a single droplet, an empirical formula presented formerly by Sami and Ogasawara is modified so as to make it available over a wider range of Reynolds numbers, down to about 1. (4) In the range of our experiment, drag coefficient of a burning droplet is reduced to about 30% of that of a solid sphere.

View full abstract

In order to realize a low compression ratio diesel engine, the problematic points, for example the engine noise resulting from the maximum rate of pressure rise and exhaust emission (NOx), have to be improved. In this paper, the auxiliary fuel injection was tried using a direct injection type engine for improving the problematic points stated above, and further some factors which affected the effects of the auxiliary fuel injection, for instance the combustion chamber configuration and the compression ratio, were discussed. As a result of our experiments, the auxiliary fuel injection in a low compression ratio engine was found a favorable means of suppressing the emission of noise and NOx, and in this case the optimum timing of auxiliary fuel injection was near the top dead center of exhaust stroke, i.e. Vigom injection, with respect to specific fuel consumption and noise emission. On the effects of auxiliary fuel injection, especially with respect to improving the maximum rate of pressure rise, the extent of the effects of auxiliary fuel injection was intensive, as the injection rate of the main injection was large and the turbulence in combustion chamber was strong. Besides, on the relationships between the compression ratio and the effects of auxiliary fuel injection, the efects of auxiliary fuel injection were intensitve at lower compression ratio with respect to the maximum rate of pressure rise and noise emission. Consequently, the adoption of auxiliary fuel injection in a low compression ratio engine might be very effective.

View full abstract

The dimensional factors which affect the die filling are examined and reported in this paper. A plasticine is used as a model material and the square grid technique is employed to measure the dislacement. The components having H-shaped sections are tested as one of the typical geometries under a plane strain condition. It has been proved that an overall forging process is divided into five stages according to the mode of deformation which occurs in the regions joining the rim and the web where the mode appears to change markedly during the process. The sequence of these deformation modes is so important for improving the die-filling that it provides an approach to the optimum design of the die and to an estimation of the optimum volume of materal. The dimensional factor which relates closely to this sequence is the ratio of flash width to height.

View full abstract