H. de Boer

Refined semi-analytical design sensitivities

Efficient structural optimization routines require availability of gradient information. Semi-analytical (SA) design sensitivities are rather popular, as they combine ease of implementation with computational efficiency. Their main drawback however, is their well-known inaccuracy problem for shape design sensitivities. It was found that the inaccuracies are especially unacceptable for slender structures and become more pronounced when relatively large rigid body motions can be identified for individual finite elements. Based on these observations, the authors recently developed a refined SA method taking full advantage of analytical differentiation of rigid body modes. The present article presents a sound and unified formulation of refined semi-analytical (RSA) design sensitivities for linear, linearized buckling, geometrically nonlinear and limit point analyses. Numerical results are presented in order to demonstrate the efficiency of the proposed method. It is concluded that the refined SA method possesses the advantages of the traditional SA method, whereas it does not exhibit its unacceptable inaccuracies.

Elastic stiffness analysis of a thermo-formed plain-weave fabric composite: Part I: geometry

The present paper investigates experimentally the influence of thermo-forming on the geometry of a plain-weave carbon-fibre thermoplastic laminate. Specifically, effects of the fabric stretching and simple shear deformation modes on yarn crimp and cross-section geometry are examined. For that purpose, specimens are stretched and sheared in a heat chamber above the melting temperature of the resin. Microscopic geometry analyses are carried out on laminate cross-sections prepared along both yarn directions. It is found that both deformation modes significantly influence yarn crimp and cross-section geometry. The observed deformations are schematised in order to enable stiffness modelling in Part II of this paper. To verify the proposed schematisations, additional laminate cross-sections are cut from thermo-formed products which contain several areas with nearly constant shear deformation. Fabric stretching and simple shear occur simultaneously in the products. Observed yarn deformations correspond well with the proposed schematisations.

Error analysis of refined semianalytical design sensitivities

In the recent past inaccuracy problems that arise when computing shape design sensitivities by the semianalytical method have been reported. Therefore, the authors have developed a refined semianalytical method. This method is based on the fact that the contribution to the pseudo-load vector corresponding to the rigid body motions can be evaluated by exact differentiation of the rigid body modes. In the present paper the effectiveness of this refined semianalytical method is studied by means of the beam model problem. This model problem has been investigated earlier by Barthelemyet al. (1988), Barthelemy and Haftka (1988), Pedersenet al. (1989) and Olhoff and Rasmussen (1991).

Post-Tensioned Glass Beams

At Delft University, glass researchers have developed a revolutionary safety concept for glass beams. This safety concept shows some analogy with reinforced concrete; glass beams are reinforced and/or post-tensioned by adding (stainless) steel to the layout of the beam.

Elastic stiffness analysis of a thermo-formed plain-weave fabric composite—part III: experimental verification

The effects of thermo-forming-induced deformations on the membrane stiffness of plain-weave reinforced thermoplastic laminates are investigated experimentally. Tensile tests on plain-weave glass and carbon reinforced laminates are presented. Stretched and sheared reinforcement configurations are considered. In order to investigate the effects of these deformation modes individually, laminates are heated above the melting temperature of the thermoplastic matrix and subjected to nearly pure stretching and shear deformations. After re-consolidation, tensile test specimens are cut out in, predominantly, the yarn and bias directions. Furthermore, test specimens are obtained from an actual thermo-formed product, which was used for geometric analyses in part I [Compos. Sci. Technol. (2000) 1041]. Simple shear leads to drastic changes in mechanical properties. Fabric stretching leads to a significant difference in stiffness between the straightened and the crimped yarn directions. With results obtained, three classical laminate theory based stiffness models are validated. These models have been reported in part II [Compos. Sci. Technol. (2000) 1249]. In-plane Youngs modulus is predicted reasonably well. Poissons ratio and shear coupling ratio are generally underestimated by the models.

Shape optimization of thermoformed continuous fibre reinforced thermoplastic products

An approach to automated shape optimization of Continuous Fibre Reinforced Thermo-Plastic (CFRTP) products is reported. The methodology combines numerical modelling of the production process and subsequent finite element analyses of the structure. As an optimization technique, the multipoint approximation method is adopted. It is assumed that CFRTP structures are produced by means of a thermoforming process. The CFRTP material consists of single-layered plain weave fabric embedded in a thermoplastic resin matrix. The thermoforming process is modelled by means of a simple kinematically based technique, which considers only the in-plane shear deformation of the fabric and omits the effect of the resin matrix. The effective laminate properties of a CFRTP composite are calculated taking into account the fibre reorientation, as predicted by the thermoforming simulation algorithm. Several numerical examples involving thin-walled CFRTP products restrained to linear structural behaviour are solved to demonstrate the proposed approach. It should be noted that this approach is also applicable to woven preforms without resin matrix, e.g. using Resin Transfer Moulding (RTM).

Initial results from COMPTEL onboard GRO

Abstract COMPTEL is the first imaging telescope to explore the MeV gamma-ray range (0.7 to 30 MeV). At present, it is performing a complete sky survey. In later phases of the mission selected celestial objects will be studied in more detail. The data from the first year of the mission have demonstrated that COMPTEL performs very well. First sky maps of the inner part of the Galaxy clearly identify the plane as a bright MeV-source (probably due to discrete sources as well as diffuse radiation). The Crab and Vela pulsar lightcurves have been measured with unprecedented accuracy. The quasars 3C273 and 3C279 have been seen for the first time at MeV energies. Both quasars show a break in their energy spectra in the COMPTEL energy range. The 1.8 MeV line from radioactive 26 A1 has been detected from the central region of the Galaxy and a first sky map of the inner part of the Galaxy has been obtained in the light of this line. Upper limits to gamma-ray line emission at 847 keV and 1.238 MeV from SN 1991T have been derived. Upper limits to the interstellar gamma-ray emissivity have been determined at MeV-energies. Several cosmic gamma-ray bursts within the field-of-view have been located with an accuracy of about 1°. On 1991 June 9, 11 and 15, COMPTEL observed gamma-ray emission (continuum and line) from three solar flares. Also neutrons were detected from the June 9 and June 15 flares.

Maximum Likelihood Method Applied to COMPTEL Source Recognition and Analysis

As the instrumental resolution of high-energy astronomical experiments increases, the observer is confronted with a ’space of outcomes’ (hereafter dataspace) with at most a few events per bin (particularly if time resolution comes into play). This implies that the sought signal is not only contaminated by additive noise components (e.g. instrumental, earth’s atmosphere), but is also masked by relatively large intrinsic statistical fluctuations. In the case of the 1–30 MeV imaging telescope COMPTEL, the number of counts per bin is of the order 1, if the full resolution is to be explored.

Accurate design sensitivities for closed-filled structures – application to plastic bottles

Abstract.Many efficient structural optimization algorithms require gradient information. Semi-Analytical (SA) design sensitivities are rather popular, as they combine ease of implementation with computational efficiency. For closed-filled structures also the effects of internal pressure changes and solubility of the gas in the fluid may be important. Typical examples for which these aspects are relevant can be found in packaging, e.g. edible-oil bottles. The present paper focuses on design sensitivities for such closed-filled structures. The design sensitivities are based on a refined SA formulation for the structure, whereas the contribution related to its contents is evaluated analytically.

COMPTEL images locations of gamma‐ray bursts

The γ‐ray telescope COMPTEL onboard GRO has so far located 6 gamma‐ray bursts which occurred in its ∼1 sr field of view. The positions of the sources were derived by the maximum‐entropy method. Systematic and statistical uncertainties for the four strongest bursts are approximately 1° to 2° and can be reduced in future analysis.