A contact concept is proposed to explain the tribology of unlubricated metal-to-metal sliding at high speeds. It is aimed at studying the relationship between relative surface speeds, contact temperature and friction coefficient during oscillatory motion under dry conditions. An experimental set-up was designed and built where two cylindrical rods of crossed axes are used to simulate the contacting oscillating bodies. Four frequencies of 4, 6, 8 and 10 Hz, a constant amplitude of 1.25 mm, and three static contacting loads of 6, 8 and 10 Newtons were tested. Temperatures at four different locations along the surface of contact of one of the rods were measured. An analytical model based upon heat flow equations and friction energy is developed to determine the gradients of both the contact temperature and friction coefficient within the vicinity of contact. The heat flow equations are solved numerically using the Finite Difference Technique. Good correlation between experimental and analytical results was obtained with a maximum deviation of no more than 15%.
In order to solve the Navier-Stokes equations in the case of free surface incompressible flows, a method has been developed on the basis of the finite volume technique applied to a 2D Cartesian grid in the vertical plane. The projection method has been adopted to solve the water phase. The air phase is not solved explicitly but a weighted linear extrapolation of the velocity field computed in water is used, ensuring a divergence free velocity field in the air. The interface tracking is ensured by a level set approach. The numerical implementation of the projection method is carried out based on an original splitting of the unknowns for the transport step, achieving first or second order space accuracy. The projection step is carried out by solving the Poisson’s equation thanks to the iterative GMRES solver. Time integration is ensured by Runge-Kutta schemes. Moreover, the viscous diffusive terms are integrated into the model allowing the explicit computation of internal losses. The implemented model was first validated for pressurized flows based on benchmarks coming from literature. Next, the solver has been validated for both steady and unsteady free surface flows. An industrial application used to reduce damp wind-induced vibrations of high chimneys is illustrated by an experimental sloshing tank.
With funding from NASA's Space Launch Initiative (SLI), Andrews Space & Technology 's approach to RLV development has lead the company to a two-stage-to-orbit concept that would save weight for HTOHL by producing oxygen from air in flight. Once the LOX tanks are filled at rather low altitude (about 10 km) with rather pure oxygen on turbofan power, the motherplane would accelerate and then launch a smaller second stage onto orbit. LOX is generated using liquid hydrogen as a coolant and as a fuel after heating with the incoming air in modified hydrogen-fuelled turbofans. To do the work on the NASA contract, Andrews S&T has, for example, teamed with Pratt & Whitney for turbofan work and UT Corp. for work on air separation technologies. Rather surprisingly perhaps, this in-flight oxygen collection concept has also been investigated since a few years under ESA contracts by the Royal Military Academy of Belgium (RMA) and other Belgian contributors. RMA has just now received new contracts from ESA under the auspices of the Belgian Federal Scientific Services in order to develop hardware for this rather exotic technology, i.e. to do work on a rotary distillation separator and on advanced heat exchangers. Experiments on the influence of a LOX collection plant integrated with the turbofan are still on the list of desired work to be done in the future. For all that, the RMA has teamed with Techspace Aero, the von Karman Institute and the University of Liege. The differences between the US and Belgian concepts will be outlined in this paper. This paper will present a summary and a comparison of the views on both sides of the ocean. It will show which critical technologies have been selected to be studied and implemented in Europe and why and also explain the trade-off done in these choices regarding the available funding. It will also show the latest developments on hardware and test installations and the perspectives on the ESA side. Vehicle performance and engine performance with hydrogen fuel will also be presented with rather details.
The aim of this paper is to present in detail case study, where the deformation of a complex connection system of a modular bridge are measured by means of the digital image correlation (DIC) technique, in order to study the displacements in the joint. The connection system is composed of metallic and composite parts. The paper shows that DIC is a complementary tool of the conventional measurement systems such as LVDT and strain gages. The use of strain gages and LVDT only allows to suggest where the major part of the deformation is located, but does not allow to understand the origin of the deformations. The DIC analysis permits to locate and to quantify the different sources of deformation, which may lead to the improvement of the design and the assembly procedure of the connection. The paper also points out the set-up procedure for the use of the CCD camera, and the precautions which must be taken concerning the contrast pattern to place on the specimen, the light conditions, the calibration procedure and the magnification factor, which must be used especially in this case where different planes are measured.
Static seal is of major concern in spatial technology and because of severe thermodynamic conditions, direct metal/metal contact is often used. This work is a contribution to the study of liquid leakage through a rough metal contact resulting from the tightening of two machined surfaces. Our approach is based on experimental measurements of leak-rates on a model configuration close to a real design on the one hand and on theoretical developments for predictive estimates of leakage on the other hand. Experiments are performed on turned metal samples reproducing the contact surface of a real metal seal. The sample is pressed against a sapphire surface under a controlled contact pressure using a specific experimental set-up. Leak tests are carried out with a liquid and leak-rate is measured versus liquid pressure and contact load using gas chromatography. Starting from the initial topology of the surface, elastic and plastic deformations are computed applying contact pressures corresponding to the ones used in the experiment. Computed deformed surfaces are further employed to form the aperture field on a representative part of the contact on which flow computation is performed. Assuming the flow to be exclusively pressure driven, the equivalent permeability of the contact is computed using the local Reynolds approximation classically employed for flow in fracture with slowly varying aperture. Experimental results are commented and compared to predictions.
Numerical simulation of priming near vacuum of liquid rocket engines inherently requires the simultaneous modelling of pressure waves, phase change rates, cavitation and multi-phase compressibility. This complexity is required to pin down the origin of a strong pressure peak observed during start-up of the upper-stage engine AESTUS on flight 510 of Ariane 5. The present study focuses only on the priming of the fuel dome. The simulation allows both the evaluation of pressure peaks at critical points inaccessible for experiments and the temporal flow direction at the injector plate.
Model updating techniques using frequency response function (FRF) data are studied in this paper. The numerical techniques are discussed for implementation with a large commercial finite element (FE ) code. System equivalent reduction expansion process is adopted to reduce the complete FE solutions onto the experimental degrees of freedom. The rank-deficiency difficulty with this method is overcome using either of two numerical techniques: diagonal perturbation and singular value decomposition. This second technique is also used in solving the updating equation. Experimental FRF data are compared with the FE solutions, and the updated model parameters are obtained via an iteration procedure. A simplified frequency domain assurance criterion is proposed to evaluate the correlation between the FE model and the measured structure at the chosen frequencies. After verifying the efficiency of the methods with several benchmark tests, the program is applied to an aeroplane model test. Some conclusions are given and remaining problems illustrated.
This article is intended as an introduction to and an overview of Pneumatic Artificial Muscles (PAMs). These are pneumatic actuators made mainly of a flexible and inflatable membrane. First, their concept and way of operation are explained. Next, the properties of these actuators are given, the most important of which are the compliant behavior and extremely low weight. A classification and review is following this section. Typical applications are dealt with in the last but one section and, finally, some concluding remarks are made.
Robots employed in machining applications (e.g fettling of castings, deburring,...) are often programmed via a teach-in procedure. This is often not economical for the machining of small series and / or complex work pieces. The time needed for the teach-in programming is often much higher than the real production time. CAD/CAM programming as done for classical multi-axis NC-controlled milling operations is a potential solution to cope with small series. Based on the CAD model, the CAM-system generates a tool path, which is output as a CLDATA file (Cutter Location DATA) and in a next phase converted by a NC-postprocessor to a specific machine (robot) program. An important task of the postprocessor is the conversion from CLDATA to robot co-ordinates. Today the programming of such multi-axis machining operations is still a complex task. The risk for having collisions between moving machine (robot) components during operation is high and the consequences drastic. Within this paper, an innovative concept is proposed. A robot simulation system (virtual machining environment) has tightly been integrated with the postprocessor. Each postprocessed position is directly checked for collision and if it occurs, a collision avoidance algorithm is applied. The developed programming concept has been demonstrated on two robot cells: one for the manufacturing of models for casting applications and one for the polishing of ship propeller blades.
Virtual environments (VEs) are becoming important tools for simulating manufacturing systems. Various virtual environment constructing techniques and toolkits have been devised in recent years. This paper provides an inside view of those methods. Based on the insight, a futuristic VE construction approach that aims to integrate manufacturing application knowledge with environment data is proposed and its implementation discussed.
Thin walled metallic structures are widely used. They can offer an optimal weight strength ratio, but their design is complicated because of the importance of stresses and deformations caused by torsion and warping. Earlier, it was unusual to check the influence of torsion on load carrying structural elements. The continuously growing accuracy of numerical methods helps to reduce the uncertainties associated with the structural modelling and contributes to the use of thin walled cross-sections. Different theories, which differ significantly for open and closed sections, have been established to study the behavior of thin walled beams. This paper presents a unified finite element formulation for the analysis of three-dimensional thin walled beams with arbitrary both open and closed cross sections. The theory, derived from Prokic's work [6-9], presents a new warping function valid for arbitrary cross-sections without neglecting the shear strains in the mean surface of the thin wall (contrary to Vlassov assumptions). The performance of this formulation is evaluated by comparing solutions of problems with De Saint-Venant, Vlassov and Benscoter theories.
Many experimental machines using legs to achieve a good mobility on very uneven ground were built and tested in the last decades. The advances in the fields of automatic controls and actuators lead to the idea that time was ripe for machines built following the configurations typical of the animal word to succeed. These machines would display the excellent mobility of living beings. This statement is critically discussed in the present work and simplified, not zoomorphic, configurations which can still be used with advantage are described.
This paper deals with the numerical simulation of the acoustic wave propagation. It is well known today that the standard finite element method (FEM) is unreliable to compute approximate solutions of the Helmholtz equation for high wavenumbers due to the pollution effect, consisting mainly of the dispersion, i.e. the numerical wavelength is longer than the exact one. Unless highly refined meshes are used, FEM solutions lead to unacceptable solutions in terms of precision, while the use of very refined meshed increases the cost in terms of computational times. The paper presents an application of the Element-Free Galerkin Method (EFGM) and focuses on the dispersion analysis in two dimensions. It shows that it is possible to choose the parameters of the method in order to minimize the dispersion and to get extremely good results in comparison with the stabilized FEM. However, to lead to those very accurate results for 1-D and 2-D problems, the EFGM needs an important computational time, mainly due to the computation and the assembly of the stiffness and mass matrices. Thus, in order to reduce this computational time, it is suggested in this paper, as a first step, to take advantage from the developments of computer hardware, currently moving towards multi-processor machines, by computing and assemblying the matrices simultaneously on several processors. This is called a parallel assembly algorithm. The paper presents the numerical assessment of the CPU performance of the parallel implementation vs the sequential one.
