GB2514147A - Method for determining the life of a component subjected to fluid loading - Google Patents

Abstract

A method for estimating the life of a component, such as a fuel tank with a baffle, subjected to fluid loading effected by a fluid (18), the method comprising: creating a finite element mesh of the component and a CFD mesh of the fluid from a CAD model of the componentusing the FE and CFD meshes to determine the fluid structure interactions by considering the displacement of the structure caused by the fluid forces calculating a fluid parameter (using a chosen fluid solver) calculating a structural parameter (using the fluid parameter) [using a chosen structural solver] determining the life of the component at least indirectly from the calculated parameters

Description

Method for Determining the Life of a Component Subjected to Fluid Loading The invention relates to a method for determining the life of at least one component subjected to fluid loading effected by a fluid.

Several systems, for example but not restricted to fuel tanks, oil pans, turbines and large tankers are repeatedly subjected to repetitive fluid loading such as but not restricted to phenomena such as aerodynamic loading, sloshing, splashing and slamming. When such systems are subjected to repetitive fluid loading, components inside said systems are subject to cyclic stresses during long periods of time. Said components can be made of metallic materials and serve, for example, for keeping sloshing of fluids inside the systems to minimum. Hence, the designs of such systems and their components are critical. Determining or estimating the life of such components is crucial even before the phase of design.

Note that the terms determine" or "determining" with regard to the lives of the components are to be construed as estimating the lives of the components since an absolutely correct determination of the life of a technical component would not be technically feasible.

Critical problems with regard to the life estimation or the damage estimation of components subjected to repetitive fluid loading before the phase of design occur since there are no computer aided methods or processes until now. Over the last decade computer aided Engineering (OAF) methods have gained profound interest within industries to addressing problems with dealing multi-physics interactions. Several commercial codes exist with predefined multi-physics coupling enabling the user to provide minimal data such as material properties of structures, fluid properties whilst the code solves complex equations addressing the physics during such interactions. Though effective in many industrial situations, the predicament with such codes arises when the user deals with scenarios were large degrees of non-linearities exist in each of the physics within the multi-physics analysis. However, in the recent years commercial codes have become flexible in allowing user-defined coupling. Also, the commercial codes have extended capabilities in allowing coupling with third-party software which provides users with a great degree of flexibility in transferring information between one code to the other and performing extended multi-physics analysis.

Especially in a situation where turbulent two-phase fluid-structure interaction problems arise when expert solvers are required for catering the need for a turbulence and two- phase flow on the one hand and for handling the intensive material in geometric non-linearities due to structures on the other hand. Most fluid-structure interaction problems where the component is subjected to cyclic stresses due to repetitive loads fall under this category. At least one order higher to this analysis is the fatigue and failure that the component undergoes due to such fluid loading which induces the diabolical complexity to the entire modelling situation.

In the prior art several ideas exist on modelling components subjected to fluid loadings by incorporating fluid-structure interaction models where free surface flow exists due to two phases. Such an illustration can be found in US 7930154 B2 where a fuel tank is subjected to impact loading which is being modelled. The fluid-structure interaction methodology was evaluated using a commercial available code (LS-DYNA with Smooth Particle Hydrodynamics (SPH)). The modelling methodology is targeted at the failure of a fuel delivery module (FDM) due to slosh waves arising from impact loading during a crash situation particularly. However, if the FDM was to be subjected to fluid loading in a periodic manner for longer time periods, the use of explicit method in the modelling methodology illustrated in said document will not be feasible. This is due to the inherent nature of the explicit methodology by itself causing the restriction to the process.

A publication by Babar et. al (SAE Int. J. Mater. Manuf. 4(1):969-979, 2011 doi: 10.4271/2011-01-0793) highlights the use of Coupled Eulerian-Lagrangian (CEL) present in ABAQUS software for performing fluid-structure interaction simulations. The simulation approach in the paper propounds the usage of CEL for many analyses that are commonly faced in the automotive community. The approach, for example, demonstrates a partially filled baffled fuel tank subjected to a sinusoidal fluid load. The implication of a fuel tank subjected to such loadings leading to failure of the baffle of the tank has been highlighted in the paper, however, no quantification or numbers illustrating the fatigue has been mentioned. Also, the method by which such a fatigue calculation was made is not highlighted in the paper.

It is therefore an object of the present invention to provide a complete process by which the life of at least one component subjected to repetitive fluid loading, in particular, undergoing cyclic stresses cannot only be accurately determined but also estimated much faster than performing a conventional fluid-structure-interaction (FS I) simulation.

This object is solved by a method for determining the life of at least one component subjected to fluid loading according to patent claim 1. Advantageous embodiments with expedient and non-trivial further developments of the invention are indicated in the remaining claims.

A method according to the present invention serves for determining or estimating the life of at least one component subjected to fluid loading effected by a fluid. The method comprises a step of providing at least one computer-aided design (CAD) model of the at least one component subjected to the fluid loading. The method also comprises a step of generating at least one finite element mesh of the computer-aided design model for structural analysis. Moreover, the method comprises a step of generating at least one computational fluid dynamics (CED) mesh of the computer-aided design model for carrying out a fluid analysis. The method also comprises a step of determining fluid-structure interaction interfaces from the finite element mesh and the computational fluid dynamic mesh, the fluid-structure interaction (FSI) interfaces characterizing interactions between displacements of the finite element mesh effectable by fluid forces obtained from the computational fluid dynamics mesh. Furthermore, the method comprises a step of defining at least initial conditions for the fluid and material properties of the component as well as a step of choosing a fluid solver for calculating at least one fluid parameter from a set of equations obtained at least indirectly by the computation of fluid dynamics mesh, the fluid parameter characterizing a fluid loading effected by the fluid.

The method also comprises a step of choosing a structural solver for calculating at least one structural parameter characterizing a displacement of the finite element mesh, the displacement being effected by the fluid loading and the structural parameter being calculated in dependency on the fluid parameter submitted from the fluid solver to this structural solver. Moreover, the method comprises a step of determining the life of the component at least indirectly in dependency on the fluid parameter and the structural parameter.

By the method according to the present invention a novel simulation methodology is provided for estimating the life or life cycle of components, subcomponents, systems and the like which are subjected to cyclic stresses due to repetitive fluid loading. The method can include creating a fully coupled fluid-structure interaction (FSI) model, solving methodology, estimating a critical cycle time (CCI) due to fluid loading through simulation, modelling and determining or estimating the life of the component subject to cyclic stresses using the estimated CCI by fatigue analysis.

The method according to the present invention is targeted to circumvent the above mentioned critical problems by allowing for realizing a complete process through which life of components subjected to cyclic stresses through repetitive fluid loadings can be accurately predicted much faster and with a very high robustness through computer aided engineering (CAE) analysis.

The crux of the method according to the present invention lies in transitioning the analysis from FSI to fatigue calculation to cut the turn around time of simulation and to accurately estimate the life of the component subjected to the fluid loading.

By the invention according to the present invention, a complete computer aided engineering (CAE) process can be provided which can include the identification of parts and associated components which are a fully coupled fluid-structure interaction model, solving methodology, estimating a critical cycle time (CCI) due to fluid loading through simulation, modelling and estimating the life of the component subject to cyclic stresses using the estimated OCT by fatigue analysis.

Further advantages, features, and details of the invention derive from the following description of preferred embodiments as well as from the drawing. The features and feature combinations previously mentioned in the description as well as the features and feature combinations mentioned in the following description of the figures and/or shown in the figures alone can be employed not only in the respective indicated combination, but also in any other combination or taken alone, without leaving the scope of the invention.

The drawing shows in: Fig. 1 a flow diagram for illustrating a method for determining the life of at least one component subjected to fluid loading effected by a fluid; Fig. 2 a schematic and perspective side view of a tank for storing a fluid, wherein a component in the form of a baffle is arranged in the tank, the baffle being subjected to fluid loading effected by the fluid stored in the tank; and Fig. 3a-d schematic side views of the tank illustrating sloshing of the fluid in the tank, the sloshing being calculated by a fluid solver used during the execution of the method.

In the figures elements or element having the same function are equipped with the same reference signs.

Fig. 1 shows a flow diagram for illustrating a method according to the present invention, the method serving for determining or estimating the life of at least one component subjected to fluid loading effected by a fluid. The term "determine" or "determining" with regard to the life of the component should be understood as estimating the life since it is not possible to calculate the life of a technical component absolutely correctly.

By the method, a complete computer aided engineering (CAE) process is provided by means of which the life of the component can be determined or estimated accurately and particularly fast. The procedure or method for performing a complete CAE analysis to predict the failure of the at least one component starts with a step Si of identifying or providing a computer aided design (CAD) assembly of all the components being subjected to repetitive fluid loading. Thereafter, procedures or steps S2, S3 are carried out in tandem or simultaneously at least partly. The step S2 comprises a finite element mesh generation of the assembly, the finite element mesh serving for structural analyses of the CAD assembly. The step S3 comprises a computational fluid dynamics (CFD) mesh generation for carrying out fluid analysis. In a next step S4, fluid-structure interaction interfaces are identified or determined. Said fluid-structure interaction interfaces are, for example, boundaries or surfaces where data exchange of forces effected by the fluid and displacements of the structure of the assembly take place, the displacements being caused by the forces of the fluid.

After identifying the fluid-structure interaction (ESI) interfaces, a step 55 is carried out. In the step Sb, initial conditions for the fluid are defined in terms of fill variables such as pressure, a fill level of the fluid in, for example, a tank, velocity of phases of the fluid etc. Subsequently, material properties of the component are defined for carrying out the structural analysis. The momentum imparted in the fluid due to fluid loadings such as but not restricted to aerodynamic loading, sloshing, splashing and slamming etc. could be defined as boundary load by providing the loading in the boundary conditions or by including the momentum source in a set of equations in the force balance terms in the sub-domain equations.

Thereafter, in a step S6, an appropriate fluid solver is chosen, the fluid solver calculating field variables such as pressure, velocity field, volume fraction etc. from a set of equations obtained by a discretization of the CED mesh. A pressure P effected by the fluid and obtained by the fluid solver which is also referred to as CFD solver is passed on or submitted to a structural solver which is chosen in a step S7. In the step S7, the structural solver receives the fluid loading on the FSI interfaces as boundary load and calculates structural variables such as displacements, stress, strain etc. The displacement calculated by and obtained through the structural solver is then passed back to the fluid solver on the FSI interfaces. Said displacements are, for example, displacements of the finite element mesh, the displacements being also referred to as nodal displacements u.

Said process of submitting the pressure form the CED solver to the structural solver and passing back the displacement calculated by the structural solver back to the CFD solver is carried out back and forth and the simulation is carried out until a critical cycle time (COT) t. is estimated or determined. The critical cycle time t. is defined as the time at which the assembly subjected to the repetitive fluid loading experiences a stress level a which is equal to that of the yield stress a., of the material of the component, wherein the determination or estimation of the critical cycle time is carried out in a step SB.

At this point the solution obtained from the fluid loading is critical since the initiation susceptible to damage has just begun and due to such a recurrence in fluid loading phenomenon, it leads to failure as an end result. During the solution process, if the stress or the stress level a in the material is not equal to the yield stress oVld, the simulation process is repeated until the stress level a is equal to the yield stress. In other words, in a step S9 the stress level a is compared to the yield stress aVWd.

The said process can also be performed as a one-way coupled fluid structure interaction wherein the fluid solver passes the fluid parameters to the structural solver whilst the structural solver does not need to pass back the structural parameters to the fluid solver.

Although the one way coupled fluid-structure interaction analysis in general would be less accurate as opposed to the fully coupled two-way fluid-structural interaction process, the stress and the strain evolution can still be realized and the CCT can still be estimated through just the one-way coupled analysis.

In a step Si 0, the stress state obtained at the CCT is taken as an input for a fatigue analysis. In steps Si i and S12, fatigue material properties such as an ultimate tensile strength, fatigue strength coefficients and all characteristics necessary for performing the S-N analysis using the time history input equal to that of the fluid loading is designed for the life estimation of the assembly through a fatigue solver in a step S13. In a step S 14, the output predicted by the fatigue solver is used to estimate the life or the fatigue life of the component due to the repetitive fluid loading, the life being estimated and dependent on at least the number of repeats or cycles until the stress level a equals the yield stress all.

The method is a procedure to carry out the full scale fluid-structure interaction process including the identification of the FSI interfaces variables such as the pressure obtained by the fluid solver and the displacement obtained by the structural solver exchanged during the solution times. During the FSI simulation, the critical cycle time is estimated, the FSI simulation including the solution time during which the stress of the assembly is equal to the yield stress o11,11 of the material and the stress-state information during the COT which is used as an input to the fatigue analysis. A life estimation of the assembly is provided in terms of number of repeats or cycles and through the damage accumulated due to repetitive fluid loading which is obtained from the FSI simulation and by using the stress-state information as an input by performing the fatigue analysis.

In the following, the method or procedure is applied to a component in the form of a baffle (Fig. 2) which is arranged in a rectangular tank 12 serving for storing a fluid in the form of liquid fuel or water. The method is used to estimate or determine the life of the baffle subjected to repetitive fluid loading effected by the fluid stored inside the tank 12.

Using the method, the end user can accurately predict the life of the tank 12 or other systems within hours which in theory would have taken months to estimate. The tank 12 in the form of a rectangular container has a height H of 250 mm, a width W of 250 mm and a length L of 600 mm. The baffle 10 arranged in the tank 12 is deformable and has a thickness of 1 mm, a width of 100 mm and a height of 160 mm. Afill level 14 of the fuel is indicated by a dotted line, the fill level being approximately 40 per cent of the capacity of the tank 12. The tank 12 containing the baffle 10 is subjected to a periodic fluid loading by the fluid. The objective of the simulation is to estimate or to determine the life of the baffle inside and enclosed in the tank. The tank 12 is subjected to a repetitive fluid loading given by the following equations: The displacement is given by equation a: = a*sin(w*t) The acceleration corresponding to such a displacement is given by equation b: = _aw2 *sin(w*t) The direction of the acceleration given by the equation b is indicated in Fig. 2 by a directional arrow 16.

The acceleration input given by the equation b is taken as the repetitive forcing function subjected to the tank 12 in the direction indicated by the directional arrow 16. 40 per cent of the rectangular tank 12 is filled by fuel or water, wherein the rest of the tank 12 is filled by air. The force acting upon the tank 12 ideally creates a free surface flow in the tank 12 in subsequently subjecting the tank 12 and the baffle 10 to violent slosh modes in different slosh cycles due to repetitive fluid loading. The method comprising the steps Si to S14 illustrated by Fig. 1 is carried out for estimating the life of the baffle 10.

The tank 12 with the baffle 10 is identified as the assembly experiencing fluid loading (step Si). A CFD mesh of the tank 12 is prepared through commercially available software STARCCM+ ver. 6.06.011 (step S3). However, several other codes such as but not restricted to ANSYS Fluent®, COMSOL Multiphysics® etc. can also be used for the same illustration (step 53). Subsequently, the finite element mesh is prepared through a commercially available software Hypermesh ver 10.1. Other commercial codes such as but not restricted to ANSYS®, ABAQUS®, COMSOL Multiphysics® can also be utilized for the same demonstration (step 52).

The baffle 10 inside the tank 12 is identified as the FSI interface (step S4) since the information exchange on pressure and the nodal displacement between the fluid and the structural solver takes place in order to estimate the life of the baffle 10 subjected to the fluid loading. Subsequently, the CAE solution process (step S5) is carried out by solving the fluid equations by providing the equation b by a body force approach to accommodate for the acceleration of the tank 12. The volume of fluid method (VOF) is used to distinguish the two phases namely water and air in the system in form of the tank 12 with the baffle 10. All boundaries of the tank 12 are taken to be impenetrable with imposing the no-slip boundary condition (step 56).

The material properties (in this case steel EN 100200) are taken for the purpose of demonstration. The elastic-plastic material model has been chosen. Since the process is transient, all boundaries are treated to be free for deformation (step S7).

Ideally, a SIMULIA CO-SIM engine® that serves as a data mapper between the fluid solver (STARCCM+ (R) ver 6.06.11) and the structural solver (ABAQUS (R) ver 6.11-1) has been employed such that the information is exchanged and mapped between the CFD mesh and the structural mesh (finite element mesh) of the baffle 10. In this case the fluid solver is taken as LEAD, meaning the fluid solver starts at first and exchanges the data to the structural solver and then the nodal displacement from the structural solver is passed back to the fluid solver accounting for the deformation in the baffle 10 through a morpher present in the fluid solver. However, a user-defined code could also be written for such instantaneous data exchange between the two solvers.

The stress distribution in the baffle lOis monitored at each solution time. Once the von mises stresses at any instance in time and at any location in the baffle 10 crosses the yield stress of the material (145 MPa for the chosen material) the FSl process is stopped.

The solution at this time is identified as the critical cycle time solution. By means of software, the simulated sloshing of the fluid in the tank 12 with the deformable baffle 10 can be illustrated. In the illustrated example the solution to the COT was found to be t = 4 seconds (steps SB and 59). The accumulated stress distribution is taken as an input to the fatigue solver for further analysis (step Si 0). The S-N analysis (step S12) along with the corresponding fatigue material properties (step Si 1) are chosen, the time history for the load provided in the equation b is taken as an input. The fatigue solver (step S13) cycles through the equation b with the accumulated stress distribution obtained from the FSI (step 58). A commercial code (n-code HBM) can be used for the fatigue analysis.

Once a predefined set of cycling, for example, 3E6 cycles, has been completed, the life is evaluated in terms of number of repeats (step S14). In this example, the life of the baffle 10 is estimated to be 1.102E6 repeats for the given fluid loading. Thus, the life of the baffle 10 subjected to the repetitive fluid loading is evaluated by the method illustrated by Fig. 1.

Fig. 3a-d each show the tank 12 with the fluid designated by 18 sloshing in the tank 12, the sloshing being simulated by said method.

The illustrated CAE process provides a very fast way of obtaining the life of the assembly being -in the present example -the tank 12 with the baffle 10, the assembly being subjected to repetitive fluid loading. The life can be determined or estimated in a much faster way comparing with a conventional FSI simulation for the entire slosh load cycles and physically testing the system through experimentation. In other words, the information obtained by FSI through CAE is passed onto calculations for damage estimation so that the life of the component can be estimated. The process is robust and hence there are currently no impediments with simulating the degree of non-linearity in the system from both the fluid and the structural points of view. Ideally, when the life cycle evaluation of the component is carried out through physical testing, the component is damaged at the end of the test. This is eliminated with the above described CAE process. All cost associated with prototyping and testing are reduced or avoided. The illustrated CAE does not depend on the data from the test as an input to carry out the simulation. The CAE process can be used even as a design solution for pre-designing and for evaluation on material reduction in the assembly and for obtaining an optimized design.

Claims (7)

1. Claims A method for determining the life of at least one component (10, 12) subjected to fluid loading effected by a fluid (18), the method comprising the steps of: a) providing at least one computer-aided design model of the at least one component (10, 12) subjected to the fluid loading (Si); b) generating at least one finite element mesh of the computer-aided design model for structural analysis (S2); c) generating at least one computational fluid dynamics mesh of the computer-aided design model for carrying out a fluid analysis (S3); d) determining fluid-structure interaction interfaces from the finite element mesh and the computational fluid dynamics mesh, the fluid-structure interaction interfaces characterizing interactions between displacements of the finite element mesh effectible by fluid forces obtained from the computational fluid dynamics mesh (S4); e) defining at least initial conditions for the fluid and material properties of the component (10, 12, S5); f) choosing a fluid solver for calculating at least one fluid parameter from a set of equations obtained at least indirectly by the computational fluid dynamics mesh, the fluid parameter characterizing a fluid loading effected by the fluid (S6); g) choosing a structural solver for calculating at least one structural parameter characterizing a displacement of the finite element mesh, the displacement being effected by the fluid loading and the structural parameter being calculated in dependency on the fluid parameter submitted from the fluid solver to the structural solver (S7).h) determining the life at least indirectly in dependency on the fluid parameter and the structural parameter (S14).
2. 2. The method according to claim 1, characterized in that the steps f) and g) are carried out, wherein the fluid solver calculates the fluid parameter and is submitted to the structural solver whereas the structural solver does not pass back the structural parameter to the fluid solver (one-way coupled fluid-structure interaction) (S6, S7).
3. 3. The method according to claim 1, characterized in that the steps f) and g) are carried out at least twice, wherein the structural parameter is submitted back from the structural solver to the fluid solver, the fluid solver calculating the fluid parameter in dependency on the structural parameter, two-way coupled fluid-structure interaction (56, S7).
4. 4. The method according to any one of claims 1, 2 or 3 characterized in that the steps f) and g) are carried out until a critical cycle time is reached, the critical cycle time characterizing a time at which a stress level of the component is equal to a yield stress of the material of the component (10, 12, S9).
5. 5. The method according to claim 3, characterized in that before step h): a fatigue analysis of the component is carried out in dependency on the critical cycle time (Sb, 513).
6. 6. The method of claim 6, characterized in that the fatigue analysis (510, S13) is carried out in dependency on an S-N-analysis of the component and/or fatigue material properties of the component (Si 1, S12).
7. 7. The method according to any one of claims 4 or 6, characterized in that the life is determined in dependency on the fatigue analysis (S14).