In order to investigate the difference between the hold-time effect in push-pull low-cycle fatigue and that in torsional one, both types of strain controlled fatigue tests of SUS 316 stainless steel were performed at 600°C with or without hold-time. Significant difference between the push-pull and torsional fatigue test data on the basis of equivalent total strain range of Mises' type was not observed in terms of number of cycles to failure, number of cycles to crack initiation and crack propagation rate. More precisely speaking, however, the push-pull test had a larger hold-time effect on the failure life and on the crack behaviors than the torsional test in lower strain range. That is, slower crack propagation rate was observed in the push-pull test without hold-time than in torsional one, but crack propagation was observed in the push-pull test with hold-time. This crack behavior was discussed from the influence of stress triaxiality near the crack tip on the crack propagation rate and also from the effect of hydrostatic stress.

View full abstract

In this paper we devised a test device for applying dynamic torsion and static tension to a specimen. Using this device, some experiments were performed on mild steels (S45C and SS41) to examine dynamic yield condition. In addition the relations between stress and strain during plastic flow were studied on the basis of a plastic flow theory including a potential function. Consequently the following were obtained. Yield stresses of S45C and SS41 under the high strain rate of 50/s ∼90/s agreed with those by Mises' condition. Equivalent stress-strain relations under dynamic combined stresses also agreed with those under compressive stresses. The relations between Lode's parameter μ and ν seemed to be μ=ν. Consequently the above theory including a potential function seemed to be valid.

View full abstract

The authors propose new X-ray methods of electrodeposition stress measurement ( Front Surface Method, Black Surface Method) in depositing process and derive formulas for various supporting conditions of a plate. As application of the present new methods, they measure the internal stresses in deposited layer and base metal of chromium deposition on one surface of a steel plate in free state.

View full abstract

A compressed granular system was divided into two kinds of zones, so-called "plastic" zone and "elastic" one. The "elastic" zone has no intergrain slip, while the "plastic" zone has it. In the "elastic" zone the force distribution around the grains was assumed to be similar to the stress ellipsoid, and the relations between the stress components and the principal values of the force distribution around the grains were derived. Furthermore the stress-strain relations in the "elastic" zone were obtained, taking translation, rotation and local deformation of granules into consideration.

View full abstract

The impact force is measured usually by the method of Hopkinson Pressure Bar. However, the measurable time range of this method is restricted by the length of the bar. The author suggests a usefull measuring method based on the theory of propagation of longitudinal elastic stress wave. In this method, the external force, ranging from quasi-static to impact, at the end of a cylindrical bar can be estimated by measuring stresses at two suitable points on it. Also the length of the measuring bar can be shorter than in another methods.

View full abstract

The longitudinal vibration of a viscoelastic bar which has random properties with respect to both the stiffness and the viscoelasticity is considered. The viscoelastic bar is subsituted by an n-degrees-of-freedom linear chain and the method to obtain the standard deviations of displacements etc. is given by developing Caravani's method. The correlation of the spring rate and the damping coefficient is also taken into consideration. The standard deviations are calculated to five-degrees-of-freedom in the numerical example. The coefficient of variation of eigenfrequencies remains constant regardless of the order of eigenfrequency. In the forced vibration, the displacement has the period 2π/ω, but the standard deviation of the displacement oscillates with the period π/ω. When the standard deviation of the damping coefficitnt increases, the phase angle deviates.

View full abstract

The hyperbolidal shells of revolution supported on columns have been used effectively as natural draft cooling towers of reinforced concrete. The purpose of this paper is to analyze the dynamic response of hyperbolic cooling towers to earthquake excitations. Taking the propagating velocity of earthquake excitations into account, the author shows that hyperbolic cooling towers oscillate not only in beam vibration modes but also in such higher vibration modes as ovalization of their cross sections. When the value αR/V is considerably large, where α and V are circular frequency and propagating velocity of the earthquake excitation, respectively, and R is the radius of hyperbolic cooling towers on the foundation level, numerical example shows that owing to horizontal earthquake excitations an hyperbolic cooling tower oscillates mainly in an ovalization mode and that owing to vertical earthquake excitations it oscillates mainly in a rocking mode.

View full abstract

This paper deals with a fundamental research for the aseismic design of a 500KV air circuit breaker. In particular, effects of nonlinearity of the damping device, which is connected to stays, on the seismic response of the circuit breaker are theoretically investigated. The model for the analysis is constructed of a simplified dynamic model of the circuit breaker structure, and it is assumed that the seismic wave is a nonstationary Gaussian white noise in the horizontal direction. The Fokker-Planck equation is first established with the aid of equations of motion which are derived without consideration of geometrical nonlinearities. Further, moment equations are introduced with the assumption of Gaussian random responses, ad the time history of the mean squared response is obtained by numerical solutions of the moment equations. From this result, effects of various parameters of the structure were made clear. The transient response to the impulse acceleration is also investigated through simulation by a digital computer.

View full abstract

This paper deals with the shape optimizing problem for a structural beam where its allowable displacement is restricted under multiple loads. A Growing-Reforming Procedure proposed by the authors is employed. The Procedure is extended to the cases where a beam has to support a number of possible loads ; that is, a method to evaluate the envelope of the stress induced by the possible loads. In order to make allowable displacement within a restricted value, a combined method of sensitivity with the Growing-Reforming Procedure is introduced. A computational algorithm based on this method is established. Some typical applications are shown to demonstrate the effectiveness of these methods.

View full abstract

The stability analysis of a slide drive servo system of numerically controlled machine tools is important to improve its performance, but is very difficult since an actual system has a lot of nonlinearities. When the slide of the system is driven, the analysis is much more difficult because of stick-flip phenomena etc. In this paper, a semi-graphic solution method using describing functions is proposed, which is similar to the one proposed in the 1 st paper in which the slide is not driven (1) But, in this paper, particular treatments are given to frictional nonlinearities and digital quantization nonlinearities because the slide is driven. By using this method, the nature of stability and the mechanism of self-excited oscillations (containing stick slip) observed in the actual system are clearly explained. The analysis and experiments are done on the software servo system in which a mini-computer is used as a digital controller.

View full abstract

A finite-difference method for solving the time-dependent three-dimensional laminar flows is proposed, with the aim of reducing the computer time. The method is a p-u method (p : pressure, u : velocity) using a convective-difference scheme, and the accuracy is increased by applying the predictor-corrector method. The solutions satisfy sufficiently the condition of continuity throughout the flow region. The method is suitable to the strongly time-dependent flow problems. In the numerical examples, it is shown that the computer time by the present method can be saved as compared with the existing method.

View full abstract

Fundamental equation for pulsating turbulent flow in a circular tube, in which the Reynolds stress is described in terms of eddy viscosity expressed with friction velocity, is approximated by a finite difference equation. Numerical solutions of the cross-sectional velocity distributions are obtained in two cases of time-dependent friction velocity and of time-independent one. It is shown numerically or experimentally that for very low frequencies the flow is in quasi-steady state. At low frequencies, only the results of the analysis with time-dependent friction velocity agree well with the experimental results. At high frequencies, the results of the analysis with time-dependent friction velocity agree with those those with time-independent one, and both of them coincide with the experimental results.

View full abstract

When a non-uniform flow accompanied with a secondary circulation enters a bend, there develops an oscillatory or spiral type of secondary flow. In this paper, this type of flow is studied theoretically and the results are compared with the experiments.

View full abstract

At first, the effect of axial velocity distribution of the pipe flow entering the rotor on the metering accuracy of a turbine flowmeter was studied theoretically and experimentally. An analysis of the relation between the revolution speed of rotor and the velocity distribution was presented. Characteristics of the meter set behind a bend or an orifice were measured and the deviation of metering accuracy due to the difference of upstream velocity profiles was discussed. Secondly, a multi-tube flow straightener was proposed to reduce the errors caused by distorted velocity profiles downstream of a single bend. Then, it is shown that a metering accuracy better than ±0.2% could be attained by using a flow straightener assembly of the overall length of 10 pipe diameters preceding the meter.

View full abstract

From the observation of axisymmetric water-water jets, it is found that their structure is almost the same as that of gas jets, although the Schmidt number of water is several hundred times as large as that of gases. Therefore, it is concluded that molecular diffusion contributes very little to mass transport in free turbulent jets. The spread of a jet in a cross flow in its symmetric plane is found to be similar to that of an axisymmetric jet, whether the cross flow is a uniform-velocity flow or a shear flow. On this ground, using the empirical relation of axisymmetric jets for jet spread, a method for predicting the shape of an axis of a jet emerging into a cross flow is proposed. Calculated shapes of jet axes in a uniform flow and a uniform shear flow agree well with he experimental results over a wide range of velocity ratios.

View full abstract

Energy equations for both a cooling fluid and a heating fluid are solved numerically with a finite difference method in the case of a uniform flow and no heat resistance in a heat exchanger surface. Temperature distributions in the direction perpendicular to the heat exchanger surface are obtained as well as in the flowing direction. Consequently, effects of flow arrangements (parallel-flow, counter-flow and cross-flow), ratios of fluid capacitance rates and ratios of heat resistances are shown by using a nondimensional duct length X₀ instead of the number of heat transfer units NTU. Here, X₀ has a form in which an overall heat transfer coefficient K in NTU is replaced with a thermal conductivity λ.

View full abstract

An experimental investigation has been made on the turbulent thermal boundary layer in the entrance region of a smooth pipe with a uniform wall temperature. The temperature profiles in the entry region are well described with the logarithmic "law of the wall". The relation of universal temperature to universal velocity becomes linear in the turbulent part of the wall region. The empirical formula of the eddy diffusivity for heat has been derived from the measurements of the eddy diffusivity and the turbulent Prandtl number in the wall region. Inseparably associated with the large-scale variation of the turbulent boundary layer thickness with time, the thermal boundary layer in the entry region fluctuates with irregular frequencies less than 2 Hz, and this has a direct effect upon the mean temperature profile.

View full abstract

The depth of plastic indentation on roller's surface is measured in order to apply to that of the tooth surface of spur gear in gear rolling. Considering this result and tooth profile error, the tooth profile after rolling is analysed theoretically. The gears are rolled at a low speed, and the calculated tooth profile and the measured one are compared. The two profiles resemble each other with the exception of pitch point. Moreover there is a closer resemblance between the two in the gear rolling by alternate reverse method. The change in tooth surface roughness, pitch error and tooth trace curve during rolling also is studied.

View full abstract