The prediction and evaluation of leakage and leak tightness is an important issue in a multitude of high-pressure applications, such as valves, flanges or threaded connections. Through the use of finite element techniques it is in general possible to determine the local contact conditions at the seal on a macroscopic level (to wit the extent of the contact area and the contact pressure in this area). However, the leak tightness of a contact area depends also on the surface topology, which is a microscopic characteristic. Therefore the assessment of leak tightness requires an evaluation criterion relating both scales. Over the years a lot of empirical evaluation criteria have been developed, each with their own application domain. However, in terms of modelling the leakage a big step forwards was made only in the last decades. The Persson method models the contact area microscopically using contact models developed in the field of tribology. This contact evaluation is then combined with results from percolation theory, which state that for a sufficiently large contact area and a uniform contact pressure leakage will occur beyond a well-specified threshold. This has yielded a potent way of evaluating leakage, but the application is limited by the requirement that the contact pressure be uniform. In many applications, such as valves or O-ring seals, the contact is Hertzian and the contact pressure distribution is not uniform but parabolic. This paper will report the first results of an effort to extent such models to Hertzian contact seals. A set of samples for leakage experiments was produced with varying surface topology. The surface of these samples is measured and the leakage behaviour under high pressure is evaluated. At the same time a cellular automata model was built and used to model percolation under non-uniform contact pressures in an effort to adapt the Persson model. Finally the experiments and the modelling results are brought together via a finite element model and compared to each other. This paper will focus on the development of the cellular automata model and the obtained simulation results.

Natural Gas is becoming an important energy source option and the capacity of the world to produce it is surging. Natural gas is usually liquefied for shipping and storage. Fire and explosion are among the most dangerous accidents in facilities at LNG Depot; especially pool fire is the most frequent incidents. At the same time the chain of accidents may lead to extremely severe consequences. In order to avoid such calamity a detail study on accident inversion technology is required to save human lives and prohibit the destruction of LNG Depot. In this thesis a topological network based fire accident inversion method for LNG tank fire accident is proposed. Firstly, analyze the LNG depot with the STAMP/STPA method. Then, the topology model of LNG tank fire inversion is established, and the optimal estimator of the shortest path is proposed according to the weighted edge topological network structure, based on which the fire location can be determined. Case study is applied to a LNG Depot. The results show that the position of the fire source calculated by the proposed method is the same as that of the simulated accident by FDS, which proves the feasibility of the method and provides a basis for reducing the fire losses and preventing accidents of LNG.

The prediction and evaluation of leakage and leak tightness is an important issue in a multitude of high-pressure applications, such as valves, flanges and threaded pipe connections that are used under extreme service conditions that occur in oil and gas exploration and production. Using Hertzian contact theory or finite element techniques it is possible to determine the local contact conditions at the seal on a macroscopic level (to wit the extent of the contact area and the contact pressure in this area). However, the leak tightness of such a contact depends also on the surface topology, which is a microscopic characteristic. Therefore, the assessment of leak tightness requires an evaluation criterion relating both scales. Empirical evaluation criteria have been postulated in the past, each with their own application domain. More recently the Persson method has been developed that models the contact area microscopically using contact models developed in the field of tribology. However, in its current form this model is limited to flat surfaces while in many applications, such as valves, O-ring seals or metal-to-metal seals of threaded pipe connections, the contact is Hertzian and the contact pressure distribution is not uniform but parabolic. This paper provides the experimental results that will be used to validate an extension of the Persson model to Hertzian contact seals. A set of samples for leakage experiments was produced with varying surface topology. The surface roughness of these samples is measured and the leakage behaviour under high pressure is evaluated. This paper focusses on the experimental evaluation of the influence of surface topology on leakage.

With improvement in innovative manufacturing technologies, e.g. automated tow-placement, it is now possible to fabricate any complex shaped structural design for practical applications. This innovative manufacturing technology allows for the fabrication of curvilinearly stiffened pressure vessels and pipes. Compared to straight stiffeners, curvilinear stiffeners have been shown to have better structural performance and weight savings under certain loading conditions. In this paper, an optimization framework for optimal structural design for curvi-linearly stiffened composite pressure vessels and pipes is presented. Non-Uniform Rational B-Spline (NURBS) curves are utilized to define curvilinear stiffeners over the surface of the pipe. An integrated tool using Python, NURBS-based Rhinoceros 3D, MSC.PATRAN and MSC.NASTRAN is implemented for performing topology optimization of curvilinearly stiffened cylindrical shells. Rhinoceros 3D is used for creating the geometry, which later can be exported to MSC.PATRAN for finite element model generation. Finally, MSC.NASTRAN is used to perform structural analysis. A hybrid optimization technique, consisting of Particle Swarm Optimization (PSO) and Gradient Based Optimization (GBO), is used for finding the optimized locations of stiffeners, optimal geometric dimensions for stiffener cross-sections and the optimal layer thickness for the composite skin. Optimization studies show that stiffener placement influences the buckling mode of the structure. Furthermore, the structural weight can be decreased by optimizing the stiffener’s cross-section and skin thickness. In this paper, a cylindrical pipe stiffened by orthogonal and curvilinear stiffeners under internal pressure and bending load is studied. It is shown that curvilinear stiffeners lead to a potential 8% weight saving in the composite laminated skin as compared to the case of using straight stiffeners.

This paper describes a novel approach for optimizing a district heating distribution network under various flow rate conditions. For district heating systems, the demand or the flow and pressure at each node varies over the time of year. The flow control that affects the operational cost can be based on the variable speed and the on/off control on serial pumping or pressure controlled valves. In the pipe system design, the topology, or the pipe layout, and the pipe diameter is optimized using genetic algorithms. Standardized methods are used for calculating the pipe thickness, supports, anchors and the thermal expansion loops. The interconnection between the pipe system and the pump station design is discussed. The objective is to minimize the total or life cycle cost (capital maintenance and operational cost), subject to ensuring demands or constraints at all points. The results are compared to classical methods where the pump station and the pipe system are designed separately and the improvements are discussed. The problem is formulated by developing an objective function where the optimization parameters define the pump arrangement, pipe system topology, and pipe diameters. The pump station and pipe system optimization consist of selecting components from a pre-defined set of elements and is implemented with discrete decision variables. Optimization of pipe elements consists of optimizing the diameter, after the topology has been defined, and is implemented with discrete variables. Flow distribution and pressure analysis is performed. Thicknesses, pressure classes, supports, expansion loops and anchors are not part of the optimization parameters, but are determined during the evaluation of the objective function. Each time the objective function is evaluated, the pipe system is designed in a sub-optimization according to given loads. The pressure head constraints are used to design the pumping curves. The method is tested on a district heating system in Reykjavik, Iceland.

In geothermal power plants, the steam system is a costly component of the infrastructure in terms of construction requirements. The steam system consists of geothermal wells, well head equipment, pipelines and separators connecting wells to the separators and the separators to the power plant. This paper describes an approach for optimum design of a steam system, with focus on topology and route selection for pipelines transporting two-phase fluid from geothermal wells to a separation station and single-phase pipelines from the separator stations to the power plant. This article proposes a new approach for selecting the locations of the separators and power plant and routes for both two phase-flow and single-phase flow pipelines in a geothermal steam gathering system. Multiple weighted distance maps calculated by uniform cost Dijkstra’s algorithm are used to find the optimum location of a site for a steam separator based on the flow capacity of geothermal wells. The two-phase flow routes are monotonic and the incline is slight in order to minimize the pressure drop and the slug flow conditions in the pipeline. Once the location of separators has been optimized, the optimum location of the power plant can be determined based on the route selection for both the pipelines transporting steam to the power plant and the pipelines transporting the used geothermal fluid from the power plant to the injection wells. The objective is to minimize both cost and visual effects. A comparison of this method to traditional methods shows that when the method described in this paper is used, shorter routes and cheaper systems can be designed with less environmental impact.

Steam turbines need to be safer and more reliable when used in nuclear power plants. In order to ensure long-term reliability of nuclear power equipment, a high safety factor is usually adopted in the design of low-pressure (LP) inner casing of steam turbines. It not only leads to larger self-weight of LP outer casings and fundamental load, but also causes higher manufacturing and transportation costs. In this paper, the stress and deformation behaviors of the LP outer casings of steam turbines are first evaluated using the numerical finite element analysis. Then, two optimization design methods, size optimization and topology optimization are used to conduct the weight reduced optimization design of inner casing, in combination with the design standards, so that the structural efficiency and performance of LP inner casings are achieved. At the same time, the self-weight and related costs are also greatly reduced. This study proposes a more optimized structural design of LP inner casings of steam turbines, and it offers considerable economic benefits.

A linear flow resistance model (LFRM) of multi-governing valve system performance was built using computational fluid dynamics to determine the distribution of flow rate through all parallel-placed steam valves at different opening ratios. A four-valve configuration connected to a water-feed pump turbine was systematically separated into three sections: valve chambers, diffuser passages and governing stages. The steam flow through each individual section was computationally modeled, and revealed that pressure drops were dependent on the flow rate. A numerical simulation strategy based on shear stress transport (SST) turbulence modeling was validated by the experimental measurements from a single valve test rig, which showed favorable agreement with the measured pressure drop at different flow rates. Subsequently, an LFRM was built to consider the geometric topology. Here, the pressure drop’s dependency on the flow rate along each section in an individual valve passage was regarded as a transfer function module. A performance map of the multi-governing valve system was obtained to predict the flow rate distribution under the opening conditions of different valves. Finally, the three-dimensional steam flow of the full multi-governing-valve system was numerically simulated to obtain the steam flow rate through different valves, and found to be in good agreement with the prediction gained using the LFRM. The proposed model can potentially be used in planning operation control strategies.

This paper presents a numerical methodology and simulation for three-dimensional transonic flow in Safety Relief Valves. Simulation of safety relief valve flows is very challenging due to complex flow paths, high pressure variation, supersonic flow with shock and expansion waves, boundary layers, etc. The 3D unsteady Reynolds averaged Navier-Stokes (URANS) equations with one-equation Spalart-Allmaras turbulence model is used. A fifth order WENO scheme for the inviscid flux and a second order central differencing for the viscous terms are employed to discretize the Navier-Stokes equations. The low diffusion E-CUSP scheme used as the approximate Riemann solver suggested by Zha et al. is utilized with the WENO scheme to evaluate the inviscid fluxes. Implicit time marching method with 2nd order temporal accuracy using Gauss-Seidel line relaxation is employed to achieve a fast convergence rate. Parallel computing is implemented to save wall clock simulation time. The valve flows with air under different inlet pressures and temperatures are successfully simulated for the full geometry with all the fine leakage channels. A 3D mesh topology is generated for the complex geometry. Detailed simulations of air flow are accomplished with inlet gauge pressure 0.5MPa and 2.1MPa. The simulated air mass flow rate agrees excellently with the experimental results with an error of 0.26% for the inlet pressure of 0.5Mpa, and an error of 2.5% for the inlet pressure of 2.1MPa. The shock waves and expansion waves downstream of the orifice are very well resolved.

The flow above the free end of a surface-mounted finite-height circular cylinder was studied in a low-speed wind tunnel using particle image velocimetry (PIV). The cylinder was mounted vertically in the wind tunnel, normal to a ground plane. The approaching flow was in the x -direction and the cylinder axis was aligned in the z -direction. Velocity measurements were made above the free-end surface in several vertical ( x-z ) planes and several horizontal ( x-y ) planes, for finite circular cylinders of aspect ratios AR = 9, 7, 5 and 3, at a Reynolds number of Re = 4.2×10 4 . The relative thickness of the boundary layer on the ground plane was δ/D = 1.7. In the vertical symmetry plane, the mean velocity measurements show the prominent separation from the circumferential leading edge, the mean recirculation zone above the free-end surface, the arch vortex inside the recirculation zone, and reattachment of the flow onto the free-end surface. Experimental evidence is found for a leading-edge separation bubble, a flow structure which has been reported in some numerical simulations in the literature. As AR decreases, the reattachment point and the centre of the arch vortex move downstream, the recirculation zone becomes thicker, and the centre of the arch vortex moves higher above the free end. Away from the symmetry plane, the recirculation zone becomes thinner, the arch vortex centre moves upstream and closer to the free-end surface, and the reattachment point moves upstream. In the horizontal planes, measurements made very close to the surface can approximate the mean surface streamline topology, revealing the pair of foci representing the termination points of the arch vortex, the prominent curved reattachment line, reverse flow beneath the mean recirculation zone, and the reattachment and separation saddle points on the free-end centerline.

Hydrogen storage remains a key issue for the high scale deployment of fuel cell applications. Gaseous hydrogen storage at high pressure with type IV vessels is the best technology. But it is necessary to reach a significant cost reduction of these storage systems. An optimization of the composite structure can be reached by numerical simulation. The goal of the OSIRHYS IV project is to develop and validate models and methods for composite high pressure design and optimization with behavior uncertainties knowledge. It was decided to limit this study to a particular topology, material and winding process. First burst simulations have been performed and results of linear static computations have been compared to experimental data. The numerical simulation models are compared with regards to vessel component masses, burst pressure, burst mode and local displacements. Results show that linear static analyses using axisymmetric and volume FE models could already predict with a reasonable accuracy the radial behavior of the tank in the case of a safe burst mode. Nevertheless, improvements of partner models are needed to reach better agreement with test data. These improvements need to be based on material and vessel geometry knowledge, behavior modelling and the FE model.

In recent years we have seen a development of novel experimental techniques that enable one to non-destructively characterize polycrystalline microstructures. These techniques hold significant advantages over approaches like serial sectioning since the specimen is not destroyed in the characterization process. This is of immense value in advancing our understanding of materials and multiscale computational models. In particular, processes at the small length scales like the initiation and early development of grain boundary damage can now be measured more closely while the resulting simulations can now be directly compared to the experimental data. The task is, however, far from being simple as extremely complex geometry needs to be coupled with advance constitutive models for the bulk grain material and the grain boundaries themselves need to be combined. In this work a model, based on a X-ray diffraction contrast tomography data of a stainless steel wire with a diameter of 0.4 mm is presented. 3D topology and crystallographic orientation of individual grains are directly transferred into a finite element model. Grain boundary damage initialization and early development is then explored for a number of cases, ranging from isotropic elasticity up to crystal plasticity constitutive laws for the bulk grain material. In all cases the grain boundaries are modeled using the cohesive zone approach. Also, the stability of the simulations and measures aimed at improving it are reported upon.

Turbulent flow between a flexible wall and a solid surface containing a backward-facing step (BFS) was investigated using digital particle image velocimetry (PIV) and high-speed photography. Motivation for this study stems from paper manufacturing industry, where high-speed, wall-bounded jets of air are employed in air clamp devices to precisely position moving sheets of paper at the specified locations with respect to other equipment. In the current investigation, stationary sheet of paper under tension was positioned in proximity to the BFS. The incoming air flow emerged from a planar nozzle that was located in the solid wall upstream of the BFS. Curvature of the nozzle wall resulted in a favorable pressure gradient condition. As a result, the flow upstream of the BFS was attached to the solid wall due to the Coanda effect. Flows corresponding to two values of the Reynolds number (3000 and 3600) based on the step height and the maximum flow velocity at the step location were characterized in terms of patterns of instantaneous and time-averaged velocity, out-of-plane vorticity, streamline topology, and turbulence statistics. In addition, paper sheet oscillation was characterized using high-speed photography. Frequencies of the natural vibration modes of the flexible wall (paper sheet) were well separated from the hydrodynamic frequencies corresponding to the oscillations of the shear layer downstream of the BFS, which resulted in the absence of resonance in the system and low characteristic amplitudes of the paper sheet oscillation. In the time-averaged sense, interaction between the separated flow downstream of the BFS and the flexible wall (paper sheet) resulted in deformation of the paper sheet and formation of diverging channel geometry between the sheet and the solid wall. As the inflow velocity increased, the paper sheet was pulled closer to the surface of the air clamp, which resulted in increased confinement of the incoming Coanda jet. The flow reattachment length calculated on the basis of time-averaged flow patterns increased with the increasing Reynolds number.

This paper reports on two criteria for burst pressure prediction of pipe with the areal surface defect. The criteria deal with not circumferential defects in internally pressured pipes without significant global bending loads. On the basis of the experimental analysis relatively simple methodologies of prediction were designed. They are based on finite element (FE) analysis and representing transition between analytical approaches and full failure simulations employing material failure models. Concept of presented methodologies is based on similarity of stress, strain and material state when plastic collapse starts in the defect domain. The similarity is expressed in terms of simple one-parametric criteria: (i) geometry and topology of the domain at which the plastic state of material is reached, (ii) amount of specific energy dissipated by plastic deformation in the regions, where the ratio of energy dissipated by plastic deformation compared to elastic stress energy is too high. Calibration of these methodologies is based on burst tests of DN800 pipes made of X60 steel with artificial defects. The ranges of defects dimensions are: longitudinal from 120 mm to 340 mm; circumferential from 60 mm to 180 mm; remaining thickness from 2.7 mm to 6.1 mm.

Acoustic power generated by turbulent flow over a coaxial side branch (deep cavity) resonator mounted in a rectangular duct is calculated using a semi-empirical approach. Instantaneous flow velocity is decomposed into an irrotational acoustic component and vorticity-bearing hydrodynamic field. The total velocity at several phases of the acoustic oscillation cycle is measured using digital particle image velocimetry. The acoustic velocity field is calculated numerically. The emphasis is on the effect of the accurate geometry representation for the acoustic field modeling on the calculated acoustic power. Despite the generally low levels of acoustic radiation from the coaxial side branches, when the main duct is incorporated into the model for calculation of the acoustic velocity, the acoustic velocity exhibits substantial horizontal (streamwise) components in the vicinity of the cavity corners. This streamwise acoustic velocity correlates with hydrodynamic horizontal velocity fluctuations, thus contributing to the calculated acoustic power. In addition, spatial structure and strength of the acoustic source changes as the distance between the side branches varies. The transformation of the acoustic source structure is characterized in terms of patterns of instantaneous and phase-averaged flow velocity, vorticity, and streamline topology.

In this paper, the Maximum Opposite Angulation Approach (MOAA) for 3-D including the topology optimization is discussed. The MOAA algorithm is developed to generate meshes in 2-D and 3-D. The basic principles of the algorithm both in 2-D applications and in 3-D applications, is to pre-set uniformity to the initial data set to form point pairs yielding possible shortest line segments. These line segments are connected with the points providing the maximum angle for the vertex of the triangular mesh to be constructed. Thus, the algorithm provides triangular meshes having well balanced interior angles and good aspect ratios. The MOAA algorithm can be proved similar to the Delaunay’s approach in 2-D from the principle and with the quickest speed. In 3-D, it was also shown that it is much more efficient than many Delaunay class algorithms with mesh architectures preserving the topology, for uniformly organized data points. In this study, the topology optimization together with the MOAA algorithm is presented to improve the precision of reconstruction of the original surface. In this context, topology judgment for intersection problem in 3-D, distortion phenomenon, the possibility of loosing some characteristics of the original surface is thoroughly investigated.

Digital particle image velocimetry is employed to investigate acoustically-coupled flow past a coaxial deep cavity (side branch) resonator mounted in a duct. The emphasis is on the effect of the separation between the coaxial side branches on the interaction between separated shear layers that form across the side branch openings. Various resonator geometries are characterized in terms of patterns of instantaneous and time-averaged flow velocity, vorticity, and streamline topology at several phases of the acoustic cycle. In addition, phase-averaged images of the flow in conjunction with unsteady pressure measurements are evaluated in order to provide insight into the mechanisms of acoustic power generation. Generally speaking, the acoustic source undergoes a significant transformation as the distance between the coaxial side branches changes. When the side branches are located relatively far away from each other, each of them forms an independent acoustic source. As the distance between the side branches decreases, interaction between the associated oscillating shear layers results in formation of a single acoustic source of complex spatial structure.

The engineering structures composed from the rods (frames, trusses, etc.) are widely used in the various industrial applications. There are many publications presenting a theoretical modeling of the rod-based structural systems especially for the designs constructed from the thin-walled components. Usually, the single bar element of such system is taken as the basis for the general model development. The growth of the use of fast computer techniques has resulted in the extensive development of numerical methods where the main role belongs to the finite-element analysis. This paper presents a new approach to the numerical simulation and dynamic analysis of the rod systems based on the use of graphs in the numerical solutions for this class of problems. The main idea is that the solution of the global 3D problem for the whole system can be split into the number of simple one-dimensional problems for the individual elements (bars). Then, at each step of the numerical integration, these one-dimensional problems are linked with the common solution procedure for the original system using additional systems of equations written for the nodes (bar joints). In this way, the global coupling for all the original problem variables is guaranteed. Implementation of this method is illustrated in this paper for the flat two dimensional rod systems and incorporates finite-difference methods. The rod systems under consideration could be composed from an arbitrary number of the rectilinear bars/rods with different geometrical and physical characteristics. Also, the random topology of the bar joints and reactions in the system nodes can be realized. The advantage of the proposed approach is that not only dynamical behavior of the whole system, but also the transition nonstationary processes in the system components could be simulated for various combinations of the load applied and boundary conditions imposed.

The work aims to study experimentally the flow structure, heat and momentum transport in the wake of two staggered circular cylinders. The cylinder center-to-center pitch ranged from 1.2 to 4.0, and the angle between the incident flow and the direction through the two cylinder axes was between 0° and 90°. In order to characterize the heat transport of this flow, both cylinders were slightly heated so that heat generated could be considered to be a passive scalar. The velocity and temperature fluctuations were measured using a movable three-wire (one X-wire plus a cold wire) probe across the wake and a fixed X-wire, which acted to provide a reference signal. Measurements were conducted at x/d = 10 and 20 and a Reynolds number of 7000, where d is the cylinder diameter and x is the longitudinal distance from the mid point between the cylinders. Based on phase-averaged vorticity contours and sectional streamlines, the flow structure behind two staggered cylinders is classified into five typical patterns. The flow behaviours for each pattern are examined in detail, including its topology, vortex strength, stability and decay rate. The different behaviours between the patterns are further linked to the different initial conditions. The momentum and heat transport are not presented due to limited space.