Jump to content

John Brink

Moderators
  • Content Count

    5
  • Joined

  • Last visited

  • Days Won

    2

John Brink last won the day on October 7 2018

John Brink had the most liked content!

About John Brink

  • Rank
    Newbie

Profile Information

  • Gender
    Female
  • Are you University user?
    No

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. Chapter 7: Kinematic Conditions 1. How is the default Type2 interface different than the interfaces described in Chapter 6? The default Type2 interface is a kinematic condition rather than a penalty formulation. 2. What is an “Incompatible Kinematic Condition” (IKC)? A situation in which a node is slave to two or more kinematic conditions. 3. What method can be employed to deal with IKC on the Type2 interface? The Type2 interface can be switched to a penalty method with Spotflag=26. 4. How else can a rigid wall be modeled to avoid IKC? Create a simple mesh representing the rigid object and connect all the nodes to a rigid body. Then the interface between any component and this rigid object can be defined with a penalty method Type7 or Type24 interface. 5. What ICoG option would you use if modeling an exterior shell representation of an engine using a Rigid Body (/RBODY)? The ICoF=4 option would apply because the master node coordinates define the center of gravity and the added mass and inertia are put on the center of gravity. The slave node mass and inertia from the shells are ignored. 6. In what situations might you need to define a local skew (coordinate system) when defining boundary conditions? Anytime the boundary conditions need to be applied in a direction not corresponding to the global reference frame. 7. What application might require the use of a moving versus fixed skew? An example would be a follower force. Chapter 8: Load Definitions 1. What nodal group definition can be used to apply a force to all the nodes of a given part? /GRNOD/PART/Part_ID 2. How is the initial velocity definition /INIVEL different than a prescribed velocity /IMPVEL? A prescribed velocity is a kinematic condition, whereas an initial velocity is an initial condition. Chapter 9: Time Step Control 1. How would you describe the conditions required for an explicit analysis to be stable? In order for an explicit analysis to be stable, the solution time step must be less that the critical time step of the model. If information passes across more than one element per time step, then the solution will be unstable. 2. What two factors does the stability condition depend on? Numerically it depends on the size of the smallest element. Physically it depends on the sound propagation speed. 3. What factors control what constitutes a reasonable time step size? The detail required to simulate the physics necessary to capture the results needed to make design decisions along with the amount of CPU power available. 4. Can the time step change during a simulation? How can this happen? Yes, as elements deform the minimum element length can reduce, and the time step is directly proportional to the element length. Also, interfaces can change the time step either due to the nodal time step which is inversely proportional to the interface stiffness or due to the kinematic time step which relates to the velocity of the node penetrating the interface gap. 5. In what situations is it helpful to have the solver automatically switch to small strain formulation? When the elements are getting distorted to the point of reducing the time step too much. 6. If the target nodal time step of a model is 1.5e-06 and the time step scale factor is 0.67, then at what calculated nodal time step would mass be added to the node if constant nodal time step is used as follows: /DT/NODA/CST 0.67 1.5e-06 Because , the time step below which mass would be added is 1.5e-06/0.67 = 2.24e-06. 7. For what types of simulation is Advanced Mass Scaling (AMS) best suited? Quasi-Static simulations where inertia effects are relatively small. 8. If the AMS time step is 10-20 times greater than the natural time step, then what speed-up can be expected? About 3 to 5 times. Chapter 10: Helpful Tools 1. How does a function behave if the simulation ends up with values outside the domain defined? The function is extrapolated from the last two points defined. 2. What /ADMAS type would you specify if you wanted to model carpeting spread over the floor of a car modeled with shell elements? Type=2 3. How would you define a group (i.e. set) of nodes that belong to a particular material? /GRNOD/MAT/Grnod_Id along with a list of Mat_Id’s 4. What are some ways that sensors could be used in a simulation? To delay the firing of an airbag or to turn off rigid bodies at a particular time. Chapter 11: Output Requests 1. Correctly match the request to the RADIOSS input file where the request is defined: Animation Output Starter File Time History Output Engine File The animation output is defined in the engine file, while the time history entities are defined in the starter. In both cases, the time interval for outputting the requests is defined in the engine file. 2. How many animation outputs are recommended to best determine model behavior? About 20 to 50 animations are usually enough. 3. What animation request “xxxx” is needed to be able to post-process top and bottom layer stress in shell elements? /ANIM/SHELL/xxxx /ANIM/SHELL/TENS/STRESS/UPPER and /ANIM/SHELL/TENS/STRESS/LOWER Chapter 12: How to Know if Simulation is Valid 1. What action should be taken if the mass error (MAS.ERR) in the engine output file runname_0001.out is 0.10? Best practice is to have less than 5% added mass (0.05), so the time-step requested might be too large. Can also check the animation file with /ANIM/MASS to see if the mass added is in a location that would affect the solution. 2. If the energy error (ERROR) in the engine output file is increasing to a large positive number, what could be the reason? It could be due to incompatible kinematic conditions. 3. If there is a large negative energy error (ERROR) what could be the reason? If the error is negative, it means that some energy has been dissipated. In case of under integrated elements (Belytschko shells, solids with 1 integration point), the Hourglass energy can also explain a negative Energy Error since it is not counted in the energy balance. The normal amount of Hourglass energy is about -10% to -15%. 4. For what type of shell formulations is a slightly positive energy error (ERROR) acceptable? In case of using QEPH shell formulation or fully integrated elements, the Energy Error can be slightly positive since there is no Hourglass energy and the computation is much more accurate. An error of +1% or +2% is acceptable.
  2. Chapter 7: Kinematic Conditions 1. How is the default Type2 interface different than the interfaces described in Chapter 6? The default Type2 interface is a kinematic condition rather than a penalty formulation. 2. What is an “Incompatible Kinematic Condition” (IKC)? A situation in which a node is slave to two or more kinematic conditions. 3. What method can be employed to deal with IKC on the Type2 interface? The Type2 interface can be switched to a penalty method with Spotflag=26. 4. How else can a rigid wall be modeled to avoid IKC? Create a simple mesh representing the rigid object and connect all the nodes to a rigid body. Then the interface between any component and this rigid object can be defined with a penalty method Type7 or Type24 interface. 5. What ICoG option would you use if modeling an exterior shell representation of an engine using a Rigid Body (/RBODY)? The ICoF=4 option would apply because the master node coordinates define the center of gravity and the added mass and inertia are put on the center of gravity. The slave node mass and inertia from the shells are ignored. 6. In what situations might you need to define a local skew (coordinate system) when defining boundary conditions? Anytime the boundary conditions need to be applied in a direction not corresponding to the global reference frame. 7. What application might require the use of a moving versus fixed skew? An example would be a follower force. Chapter 8: Load Definitions 1. What nodal group definition can be used to apply a force to all the nodes of a given part? /GRNOD/PART/Part_ID 2. How is the initial velocity definition /INIVEL different than a prescribed velocity /IMPVEL? A prescribed velocity is a kinematic condition, whereas an initial velocity is an initial condition. Chapter 9: Time Step Control 1. How would you describe the conditions required for an explicit analysis to be stable? In order for an explicit analysis to be stable, the solution time step must be less that the critical time step of the model. If information passes across more than one element per time step, then the solution will be unstable. 2. What two factors does the stability condition depend on? Numerically it depends on the size of the smallest element. Physically it depends on the sound propagation speed. 3. What factors control what constitutes a reasonable time step size? The detail required to simulate the physics necessary to capture the results needed to make design decisions along with the amount of CPU power available. 4. Can the time step change during a simulation? How can this happen? Yes, as elements deform the minimum element length can reduce, and the time step is directly proportional to the element length. Also, interfaces can change the time step either due to the nodal time step which is inversely proportional to the interface stiffness or due to the kinematic time step which relates to the velocity of the node penetrating the interface gap. 5. In what situations is it helpful to have the solver automatically switch to small strain formulation? When the elements are getting distorted to the point of reducing the time step too much. 6. If the target nodal time step of a model is 1.5e-06 and the time step scale factor is 0.67, then at what calculated nodal time step would mass be added to the node if constant nodal time step is used as follows: /DT/NODA/CST 0.67 1.5e-06 Because , the time step below which mass would be added is 1.5e-06/0.67 = 2.24e-06. 7. For what types of simulation is Advanced Mass Scaling (AMS) best suited? Quasi-Static simulations where inertia effects are relatively small. 8. If the AMS time step is 10-20 times greater than the natural time step, then what speed-up can be expected? About 3 to 5 times. Chapter 10: Helpful Tools 1. How does a function behave if the simulation ends up with values outside the domain defined? The function is extrapolated from the last two points defined. 2. What /ADMAS type would you specify if you wanted to model carpeting spread over the floor of a car modeled with shell elements? Type=2 3. How would you define a group (i.e. set) of nodes that belong to a particular material? /GRNOD/MAT/Grnod_Id along with a list of Mat_Id’s 4. What are some ways that sensors could be used in a simulation? To delay the firing of an airbag or to turn off rigid bodies at a particular time. Chapter 11: Output Requests 1. Correctly match the request to the RADIOSS input file where the request is defined: Animation Output Starter File Time History Output Engine File The animation output is defined in the engine file, while the time history entities are defined in the starter. In both cases, the time interval for outputting the requests is defined in the engine file. 2. How many animation outputs are recommended to best determine model behavior? About 20 to 50 animations are usually enough. 3. What animation request “xxxx” is needed to be able to post-process top and bottom layer stress in shell elements? /ANIM/SHELL/xxxx /ANIM/SHELL/TENS/STRESS/UPPER and /ANIM/SHELL/TENS/STRESS/LOWER Chapter 12: How to Know if Simulation is Valid 1. What action should be taken if the mass error (MAS.ERR) in the engine output file runname_0001.out is 0.10? Best practice is to have less than 5% added mass (0.05), so the time-step requested might be too large. Can also check the animation file with /ANIM/MASS to see if the mass added is in a location that would affect the solution. 2. If the energy error (ERROR) in the engine output file is increasing to a large positive number, what could be the reason? It could be due to incompatible kinematic conditions. 3. If there is a large negative energy error (ERROR) what could be the reason? If the error is negative, it means that some energy has been dissipated. In case of under integrated elements (Belytschko shells, solids with 1 integration point), the Hourglass energy can also explain a negative Energy Error since it is not counted in the energy balance. The normal amount of Hourglass energy is about -10% to -15%. 4. For what type of shell formulations is a slightly positive energy error (ERROR) acceptable? In case of using QEPH shell formulation or fully integrated elements, the Energy Error can be slightly positive since there is no Hourglass energy and the computation is much more accurate. An error of +1% or +2% is acceptable.
  3. Chapter 5: Material Laws and Characterization 1. What is the difference between nominal (engineering) stress-strain and true stress-strain? True stress takes into account the instantaneous area of a tensile specimen. True strain is obtained by summing all the ratios of infinitesimal changes in length during a tensile test. 2. What physical effects influence the stress in a Johnson-Cook material? The strain, the strain rate and the temperature. 3. How can more advanced failure models be added to a material model? By using a /FAIL/Keyword option. The ID of the /FAIL card needs to match that of the material to link them together. 4. How can a Johnson-Cook material model be modified in order to avoid failure in compression while allowing failure in tension? By not using the failure model built into the material law, but rather link a /FAIL/JOHNSON definition to the material. The /FAIL/JOHNSON failure model allows for different failure mode definitions for compression, shear and tension. 5. What is a good application for material Law 27? Material with orthotropic brittle failure behavior such as glass. 6. What is the abscissa value for the stress-strain function for a Law 36 material? Law36 requires plastic true strain values on the abscissa. The ordinate is true stress. . 7. What material law can be used to model hyperelastic materials? /MAT/OGDEN (Law42) can be used to model rubber-like materials. 8. For foam material models (Law70) are engineering or true stress-strain curves used on the corresponding property card? Engineering stress-strain which is total small strain, Ismstr=11. Chapter 6: Interface (Contact) Modeling 1. How is the interface between two colliding parts represented in RADIOSS? An interface between parts is represented using the penalty method which treats the behavior between slave nodes and master segments as springs that generate resistive forces as a function of penetration. 2. What model parameters define the initial stiffness of the interface? Ultimately, the initial stiffness is a function of the material modulus and thickness of a shell or the bulk modulus and volume of a solid. The manner in which these stiffness values affect the initial stiffness of the interface depends on the value of Istf. 3. Why is it good practice to always specify a small Gapmin on Type7 interface when using variable gap, Igap=1? To avoid getting unexpected results when parts are different thickenss which is due to the way the variable gap is defined. 4. What is the result if a slave node lies on a master segment in a Type7 interface? If the slave node is past the master segment? If a slave node lies on the master segment, an error will result. If the slave node is past the master segment, the parts could be non-physically locked together. 5. What parameter could be used on the Type7 interface if you are unable to correct a model that has nodes too deeply penetrated (even for Inacti flag to fix) so that the model will run. Fpenmax can be used to automatically deactivate nodes that are too deeply penetrated. 6. What can happen to the interface time step as the slave node penetrates further into the gap for a Type7 interface? Because the Type 7 interface is a variable stiffness, the time step can decrease as the stiffness of the interface increases due to penetration into the gap. 7. What value would be calculated for the gap of a Type24 interface if two parts modeled with shells (one of 2.0 mm thickness, the other of 4.0 mm) came into contact with each other? The gap value would be the average of the thickness values of the two parts, which is 3.0 mm. 8. How can a press fit be modeled with a Type24 interface? By specifying Inacti= -1 which will take all initial penetrations into account. 9. What can happen to the interface time step as the slave node penetrates further into the gap for a Type24 interface? The time step of a Type 24 interface is constant because the stiffness definition is linear. 10. What is the difference between the stiffness representation between Type7 and Type24? Type7 stiffness in nonlinear (more physical for nodal location) and Type24 is linear (less accurate nodal location). 11. How would you choose whether to use a Type7 or Type24 interface? Type 7 interface provides more accuracy for describing self-contact and buckling problems. It works well for shell to shell or shell to solid contact. Type 24 is best suited for solid to solid contact problems or if press fits are being modeled.
  4. Chapter 2: Intro to Explicit Methods 1. What types of physics are best analyzed using implicit methods? Why? Simulations that are static (slowly applied loading), because there is little or no inertia involved. 2. What types of physics are best analyzed using explicit methods? Why? Simulations that are dynamic, highly non-linear, and transient (short duration). 3. How is the solution time step related to the speed of sound? The solution time step is inversely proportional to the speed of sound. So, if the material is stiffer (higher modulus of elasticity), the speed of sound increases resulting in a lower solution time step. 4. What physical parameters control the speed of sound? What modeling parameter affects the solution time as a result? Both material modulus and density control the speed of sound. Thus, to keep a stable solution, the sound (or shock) wave through a medium must not pass across more than one element per time step. This implies that the length of the element is the modeling parameter that affects the overall solution time. Chapter 3: Running a RADIOSS Simulation 1. What is the purpose of the RADIOSS Starter file? The RADIOSS Engine file? The Starter contains definition of the model (e.g. nodes, elements, properties), while the Engine contains the run controls (e.g. termination time, animation output requests). 2. What advantage is there to having two input files: Starter and Engine? The setup makes it easy to run a restart file. Also, just the starter can be run first to check for warning messages before having the engine run the simulation. 3. What does a block define in the RADIOSS input deck? A RADIOSS block defines one feature of the model (e.g. node definition, definition of a material). 4. What character denotes the start of a new block? Each block starts with the character “/”. 5. What characters can be used to put a comment in the input deck? Either a “#” or “$”. 6. What file should you look at to view any errors or warnings after running the Starter? The input deck runname_0000.rad will generate an output file from the Starter called runname_0000.out that will contain any errors or warnings related to the model definition. Chapter 4: Element Formulations 1. What assumption(s) are being made when using shell elements? Shell elements do not calculate a stress through the thickness of the shell (normal stress). 2. What causes hour-glassing of a shell element? To be more CPU efficient, explicit codes typically have one integration point on the face of a shell element. Non-physical “hourglassing” modes can result due to numerical instability of the shell. 3. How many integration points through the thickness of a shell are recommended to capture bending? At least five integrations points should be used through the thickness of a shell. 4. When would small strain formulation be useful for solid elements? When defining foam materials (Law70). There are also situations where switching to small strain formulation can help the time step from dropping too much due to excessive deformation. 5. For what types of problems would thinning of the shell element be important? Certainly for metal forming simulations or any problem where the physics of the results change as a result of thinning of the shells. 6. What type of situations require solid elements? Any case that requires three-dimensional resolution of stress. 7. What physical entities could be modeled with spring elements? Many applications. Examples are joint definitions used in dummy modeling, seat belt looped through a D-ring, etc.
  5. This forum is for discussing ideas, concepts, and questions that arise during the "Intro to RADIOSS for Impact" training on June 21st through June 28th. Feel free to post!
×
×
  • Create New...