中村 俊哉

The objective of this study is to develop a novel aircraft design approach using biomimetics as an alternative to traditional airframes. This approach is primarily inspired by the dragonfly wing, which possesses reinforcement structures composed of cross veins and longitudinal veins. These structures are assumed to regulate deformation and enhance stiffness, respectively. The cross veins were replicated using weighted centroidal Voronoi tessellation (WCVT) based on the out-of-plane displacement of the skin. In contrast, the longitudinal veins were replicated by extracting a centerline from the topology optimization (TO) results on the skin, achieved through image analysis techniques such as binarization and skeletonization. The longitudinal layout effectively reduces compliance by distributing internal loads, utilizing only essential reinforcements on the skin without increasing its mass. The WCVT layout significantly enhances the buckling resistance of the reinforced skin. As a result, the skin reinforced using both cross–longitudinal layouts from TO and WCVT exhibited a buckling load 2.7 times greater while maintaining a lower mass compared to conventional layouts.

An analytical approach was developed for assessing the thermal history of thermoplastic composites during tape placement for in-situ consolidation of the automated fiber placement (AFP) technique. In this study, the heat-conduction equation is developed and solved for internal energy instead of temperature because the diffusivity does not significantly change over a wide range of temperatures in the AFP process, despite the significant change in the specific heat and conductivity. The three-dimensional internal energy history is derived in an integrated form using the Green function technique. The temperature field obtained from the internal energy field agreed with the finite element solution, where the temperature-dependent thermal properties were considered. The effects of manufacturing parameters, such as the placement speed and thickness of the placed laminates, on the thermal history of laminated composites during AFP are discussed using the present approach and finite element analysiss.

Tensile tests of quasi-isotropic laminates with a circular open hole, which include a gap in the hole area, were conducted to demonstrate the effect of the gap on the strength. The test showed that the reduction in the open-hole strength owing to the gap was less significant than that in the no-hole tensile strength. A two-dimensional mechanical model of quasi-isotropic laminates with an arbitrary inclined gap was proposed and analytically solved to obtain a closed-form expression for the stress concentration. An approximate expression was provided as the sum of the global uniform stress field and the local stress field near the open hole, which was solved based on a complex variable method. The present analytical solution agrees well with the corresponding two-dimensional finite element solutions. The solution indicated that the stress increase owing to the gap is limited at the stress concentration area. The analytical results were consistent with the experimental results.

Compression tests of quasi-isotropic laminates with embedded gaps were conducted to reveal the effect of the gap(s) on their compressive strength. Specimens were prepared such that the gaps are located at designated relative positions on a free side edge of the gage section. An analytical solution for a two-dimensional mechanical model of quasi-isotropic laminates with an arbitrary inclined gap is used to estimate the stress changes at the free edge owing to inclined gaps in the arbitrary relative locations. The relative reductions in the compressive strengths of the specimens with different gap arrangements were consistent with the analytically estimated increases in stress. The present results indicate that the significant reduction in compressive strength is caused not only by the waviness of the laminas at the gaps but also by the stress concentration at the free boundary due to the gap ends. Moreover, it is found that the standardized compression test may not be adequate to investigate the effect of the defect introduced by automated-fiber-placement method on the strength of the laminates because the sizes of the gage section of the compression test specimens specified by standard test methods are small compared to the gap spacing.

An engineering system such as an aircraft or a satellite undergoing rigid motion can inevitably include many types of uncertainties. These may be structural parameters such as dimensions, material properties, etc., and motion variables including velocity, acceleration, and direction. Consequently, an evaluation of their effects plays an important role in structural design. In this study, a differential equation is derived for the response of a moving elastic body under uncertain conditions based on the perturbation method. The effect of the uncertainty is represented by the sensitivity of the output with respect to the uncertain parameter. The equation is applied to a rotating Euler-Bernoulli cantilever beam and numerical examples are presented. As the natural frequency depends on the rotational velocity, the strain sensitivity responses are complex.

This study reconstructs a two-dimensional stress field from measured strain data. The advantage of using stress functions is that the stress equilibrium and strain compatibility are automatically satisfied. We use the complex stress functions given by the finite series of polynomials. Then, we find the proper set of coefficients required to make the best fit to the measured strain data. Numerical examples represent the stress concentration problems around a hole(s) in a plate. It is demonstrated that the present method reconstructs the stress field around the hole(s), and the estimated stress agrees with the finite element (FE) analysis result.

The corrosion resistance of 2024-T3 (UNS A92024) Al alloy with no clad layer and that of friction stir welded (FSW) joint specimens fabricated from the same material were evaluated. The surfaces of both the alloy base material and FSW joint specimens were ground out before being exposed to a 3.0% sodium chloride solution at 60°C for 24, 48, 72, or 96 h. The corrosion pits on the base material samples were found to be randomly distributed, while those on the FSW joint were formed around the edge and center of the weld line. Energy dispersive x-ray spectrometry indicated constituent particles containing Mg at the grain boundaries in the thermomechanically affected zone and stir zone of the FSW joint; this Mg content aggravated the corrosion damage in those regions. The depth and volume of the corrosion pits in the FSW joint were greater than those in the base material. However, the aspect ratios of the corrosion pits in the base material and FSW were similar. Prior-corroded specimens were fatigue tested to evaluate the effect of corrosion damage. The fatigue life of the base material with corrosion damage was slightly shorter than that of the FSW joint specimens with corrosion damage, and the fatigue life of an uncorroded FSW joint specimen was more than 10 times longer than that of a corroded specimen. Thus, corrosion damage has a severely detrimental effect on fatigue life. Further, fracture surface observation revealed that the fracture origins in the FSW joint specimens tended to be multiple corrosion pits; however, the corrosion pits with the greatest depth or volume did not necessarily become fracture origins in the base material or FSW joints.

The purpose of this work is to develop a method for solving inverse transient-state thermal and strength non-linear problems in complex shapes. Non-linearity is caused both by the material temperature-dependent properties and radiation. The proposed algorithm reconstructs the whole transient temperature and thermal stress distribution based on temperatures measured in the element selected points. Measured transient temperature values are generated during a numerical simulation of aerodynamic heating on the atmospheric reentry capsule. Both constant and temperature-dependent properties of the material are assumed. A comparison is presented between the transient temperature distributions obtained based on the material constant and temperature-dependent properties. Finally, the developed method is used to identify the transient temperature and stress distribution in the atmospheric reentry capsule assuming temperature-dependent properties of the material. The proposed approach is expected to be a good solution for improving spacecraft structures.

Crack propagation tests were conducted to clarify the effect of stress ratio on crack growth rate in friction-stir-welded 2024-T3 aluminum alloy. The effect of the distance between the weld line center and the center of the specimen was also evaluated. The result indicates that the peak of the acceleration is not at the area of maximum tensile residual stress. The modified stress ratio that considers the residual stress identifies that the location of the peak modified stress ratio corresponds to that of the peak maximum acceleration. The relationship between stress ratio and crack growth acceleration in friction-stir-welded panels is also evaluated by an analysis using the stress intensity factor range with and without residual stress and the crack growth rate for the master curve obtained by the base material with a different stress range. The maximum acceleration decreases with the increase of the distance between the weld line center and the center of the specimen and with the increase of the stress ratio. Experimental and analytical results showed similar characteristics, although analytical results showed a larger acceleration ratio. Fracture surface observation implies that the small grain size in the stir zone would have prevented the formation of striations in the area.

A transient inverse heat conduction analysis enables the identification of the unknown boundary heat flux from finite number of temperature data obtained, e.g., in the high temperature structural tests or in the actual operation. Since the prediction of thermal load (heat flux) is difficult, the inverse analysis is expected to improve structural design of high temperature components. The present study develops a computational method of transient inverse heat conduction analysis. The developed code is applied to problems of a simple two-dimensional plate and of an atmospheric reentry capsule. Sequential Function Specification (SFS) method and Truncated Singular Value Decomposition (TSVD) are employed to improve the stability of the inverse analysis. Effects of these regularization methods are numerically discussed. © 2014 Elsevier Masson SAS.

Operational load and stress data are useful for structural integrity management and damage prognosis of aerospace systems. Identifying aerodynamic loads by monitoring strain is not easy because the loads are distributed continuously over the structures surface. In this study, we propose a flexible method for interpolating a continuous load distribution in order to identify the full-field aerodynamic load from strain data acquired at a number of discrete points. Our method uses the conventional finite element method and pseudo-inverse matrix, and we further extend it by coupling with an aerodynamical equation. Numerical simulations show that this extension improves the estimation accuracy when only a limited amount of strain data is available. The effects of measurement error are also discussed. It is concluded that the rank reduction method improves the estimation accuracy and that use of a proper aerodynamical restriction can suppress the adverse effect of measurement error. © 2011 Elsevier Masson SAS © 2011 Elsevier Masson SAS. All rights reserved.

In this paper, metallographic observations, hardness measurement, and static and fatigue tests were conducted to investigate the discontinuity states which become crack nucleation sites in friction stir welded butt joints in 2-mm-thick 2024-T3 aluminum alloy and static and fatigue properties of the joint. Because different types of surface finish can be used depending on the application of the joint, several types of surface conditions were tested to evaluate their effect on crack nucleation sites and static and fatigue life. Indentation hardness tests revealed that typical hardness reduction is not necessarily observed on the section of the welding line. Based on fatigue test results, it was confirmed that there are several types of crack nucleation sites for friction stir welding (FSW) joints depending on the surface finish, and the features of the fracture surface also differ depending on the site. Furthermore, the type of discontinuity state affects the fatigue life of the FSW joint. © Springer-Verlag London Limited 2010.

Numerical simulation of fatigue crack propagation based on fracture mechanics and the conventional finite element method requires a huge amount of computational resources when the cracked structure shows a complicated condition such as the multiple site damage or thermal fatigue. The objective of the present study is to develop a simulation technique for fatigue crack propagation that can be applied to complex situations by employing the continuum damage mechanics (CDM). An anisotropic damage tensor is defined to model a macroscopic fatigue crack. The validity of the present theory is examined by comparing the elastic stress distributions around the crack tip with those obtained by a conventional method. Combined with a nonlinear elasto-plastic constitutive equation, numerical simulations are conducted for low cycle fatigue crack propagation in a plate with one or two cracks. The results show good agreement with the experiments. Finally, propagations of multiply distributed cracks under low cycle fatigue loading are simulated to demonstrate the potential application of the present method. Copyright © 2007 by ASME.

A finite element code for the transient combined radiation-conduction analysis is developed. The Gaussian quadratures technique and linear trial functions are used to approximate the scattering integration and the radiation intensity distribution, respectively. The Galerkin method is employed to yield a finite element formulation for radiation analysis. Numerical examples are presented and are compared with published solutions. It is found that the present code is equivalent to the other methods. A series of experiments with ceramic tile insulation is also conducted to examine the validity of the developed code. The transient temperature and the radiative heat flux at the back surface are measured during heating. Agreement is good between the experimental and the analytical results.

A nonlinear damage model, previously developed for creep-fatigue life evaluation of Mod. 9Cr-1Mo steel and 316FR stainless steel in a high vacuum environment was applied to 2 1/4 Cr-1Mo steel. In the damage model, the damage accumulation process is considered to be composed of three basic processes : Fatigue, creep and creep-fatigue. The fatigue damage process consists of a crack initiation period and a crack growth period. The creep damage process consists of nucleation of creep cavities. Damage interaction is fatigue cracks propagate from creep cavities. Application of this damage model to creep-fatigue tests under complex waveforms and complex strain histories showed a good correlation with the experimental results within a factor of 2.

Fundamental deformation and deformation-recovery behaviors of Ni-Ti-Nb shape memory alloy were experimentally investigated, especially their dependency on test temperature. Monotonic tension tests and deformation-recovery tests were conducted at low and high temperatures ranging from 253 K to 473 K. Stress-strain response above 323 K is quite different from that below the temperature. The yield stress shows the positive temperature-dependency below 323 K and the opposite trend is observed above this temperature. From this result, it is estimated that the dominant mechanism of inelastic deformation is the plastic deformation above 323 K and the pseudo-elastic one below it. The temperature at which strain recovery starts strongly depends on the maximum prestrain and prestraining temperature. Residual strain is always observed in the deformation-recovery tests. Also it is found that the larger recovery strain can be obtained for the larger prestrain and at the lower prestraining temperature.

Novel and consistent relaxation behaviors have recently been found with several steels at homologous temperatures less than 0.5, on a Titanium alloy, on an Aluminum alloy and on the polymer Nylon 66 at room temperature. A strong dependence of the relaxation rate on prior loading rate was observed. At equal relaxation times the stress at the end of the relaxation period associated with fastest (slowest) prior strain rate has the smallest (largest) magnitude. It is shown that the stress rate term in the evolution law of the equilibrium (back) stress enables the viscoplasticity theory based on overstress (VBO) to naturally model this newly found relaxation behavior. If the stress rate term is absent, as is the case for other state variable theories, this relaxation behavior cannot be modeled without modifications. Analyses valid at the beginning of and during the relaxation tests and numerical experiments illustrate the properties of VBO and other "unified" state variable models at low and high homologous temperatures.

Creep-fatigue tests were conducted using 316FR stainless steel at 650°C in air under a variety of strain time programs including symmetric continuous, unsymmetric continuous and strain hold cycles. By comparison of the results with the test data obtained in vacuum, it was found that the life reduction which occurred for a slow tension - fast compression strain wave and the tensile strain holding wave was smaller in air than in vacuum, and that the symmetric continuous cycle gave the largest life reduction from the data obtained in vacuum. Following the creep-fatigue tests, failure surfaces were investigated by SEM. When time/rate-dependent life reduction occurred, the fracture mode in vacuum was intergranular, but in air, striations were observed and intergranular fracture was rarely observed.

The viscoplasticity theory based on overstress (VBO) has been employed to model the uniaxial cyclic deformation of 316FR steel at high temperature. The experiments show that the stress strain response is rate independent in monotonie tension or the initial stage of cyclic straining, while the observed cyclic strain hardening is rate dependent. Also, the subsequent holding of strain results in a particular stress relaxation. These interesting deformation behaviors relating rate effects cannot be well reproduced by the conventional viscoplasticity theory. In the present paper extensions of VBO are introduced to reproduce these behaviors. The first extension admits a rate independent asymptotic solution. The second one produces the time (rate) dependent cyclic hardening. Good agreement is found between the experimental data and the numerical simulations by the present model.

The ratcheting behavior of the “unsymmetric two-bar system” was investigated by numerical experiments. The two bars are restrained to the same length and are subjected to a constant load. One bar sees cyclic temperature variations, while the other bar is kept at constant temperature. The material models employed are rateindependent plasticity (kinematic hardening) and the viscoplasticity theory based on overstress (VBO) matched to represent the cyclic neutral 6061 T6 aluminum alloy elastic and inelastic deformation behavior. For simplicity, temperature-independent material properties were assumed. Numerical analyses were performed to investigate the effects of rate of thermal loading and temperature range. Elastic-inelastic shakedown is ultimately achieved due to work hardening. There is a strain range increase until it reaches a steady value. Kinematic hardening and VBO predict almost the same strain range, which, for the case of VBO, is nearly rate-independent. The behavior for both material models is very different for the mean strain. For VBO, the number of cycles to shakedown is rate-dependent and is considerably larger than for kinematic hardening. Finally, the steady-state mean strain and strain range are computed directly for VBO. © 1997 by ASME.

A series of creep-fatigue tests has been conducted with modified 9Cr-1Mo steel at 873 K in a high vacuum environment of 0.1 mPa. In order to investigate the accumulation of creep-fatigue damage, the creep-fatigue test programme includes changes in strain waveform during the test: from creep-fatigue type to fatigue type and from fatigue type to creep-fatigue type. The conventional linear cumulative damage rule for fatigue and/or creep-fatigue damage fails in evaluating the creep-fatigue life under the present complicated strain wave history. The linear summation of the life fraction is smaller than unity when the prior loading is creep-fatigue type and larger than unity when the prior loading is fatigue type. Scanning electron microscope (SEM) observation of the fracture surface was also conducted. In the case where the strain waveform changes from prior creep-fatigue type to subsequent fatigue type, the crack mode changes from transgranular to intergranular with an increase in the prior creep-fatigue loading history. In the case where the strain waveform changes from prior fatigue type to subsequent creep-fatigue type, the primary crack mode is generally intergranular regardless of the prior fatigue loading history.

Creep-fatigue tests were conducted with Modified 9Cr-1Mo steel at 600°C in a very high vacuum of 0.1 μPa in order to investigate a 'pure' creep-fatigue behavior which is free from the environmental effect of the air. On the whole, the time-dependent life reduction occurs when the duration of tension is longer than that of compression. A good correlation was found between the creep-fatigue life and the fracture mode. When the fracture mode is transgranular, no time-dependent life reduction occurs. In contrast, when it is predominantly intergranular, a significant life reduction is observed. The overstress was experimentally analyzed to evaluate the creep-fatigue damage. The relationship between the overstress and the inelastic strain rate can be fitted, being independent of the strain range.

A series of creep-fatigue experiments has been conducted with 304 stainless steel at 650 °C, 2 1 4Cr-1Mo steel at 550°C and modified 9Cr-1Mo steel at 600°C in air and in a very high vacuum environment of 0.1 μPa. A damage model based on the overstress concept was employed to evaluate the "pure" creep-fatigue life observed in this high vacuum environment which is completely free from the environmental effect of air. This damage model for the vacuum data was also applied to data obtained in air in order to evaluate the environmental effect of air on the creep-fatigue interaction. It was found that the fatigue damage is highly affected by the air environment, resulting in a time-rate-dependent life reduction. This life reduction is mainly controlled by the strain rate or the time duration of the compression loading stage. The environmental effect of air on the creep damage is complicated. The analyses based on this damage model suggest that the air environment accelerates the creep damage in 304 stainless steel, gives no effect in modified 9Cr-1Mo steel and reduces the creep damage in 2 1 4Cr-1Mo steel. © 1994.

Following creep-fatigue tests with Modified 9Cr-1Mo Steel in a high-vacuum environment, creep-fatigue surfaces were observed by SEM. In the cases where the creep-fatigue life is time- or rate-independent, the primary fracture mode is the transgranular type, but some differences were observed in the number of dimples being dependent on the strain wave form. When time- or rate-dependent life reduction takes place, the primary fracture mode is found to be of the intergranular type. This intergranular fracture mode is mainly observed on the inner side of the specimen, and the fracture mode becomes transgranular on the outer side. The area being fractured in the intergranular manner is correlated to the time-dependent damage variable calculated by the creep-fatigue life evaluation model based on the overstress.

A unified inelastic constitutive equation is proposed to describe the cyclic inelastic deformation of 2 1 4Cr-1Mo steel at an elevated temperature of 550°C. In the present study, the evolution of overstress is discussed which is shown to lead to a time-independent term in the inelastic strain rate. This term is proportional to the stress rate. Another significant point proposed in this study is the distinction of the inelastic deformation depending on the evolution of overstress. A mathematical model has been developed based on these assumptions. Many numerical experiments were conducted to examine the validity of the present model, and in comparison with the experimental results, the present model has been shown to give good descriptions for a wide variety of loading conditions, including rate-dependent plasticity, stress relaxation, and cyclic softening. © 1993.

A unified inelastic constitutive equation is proposed to describe the cyclic inelastic deformation of 2 l/4Cr-lMo steel at 550° C. In the present study, an evolution of the overstress is discussed which leads the time-independent term into the inelastic strain rate. A mathematical model is developed and some examples of the numerical simulation are presented. © 1993 by ASME.