Applies To Product(s): AutoPIPE, Version(s): 2004, XM, & V8i Environment: N/A Area: Reports Subarea: Original Author: Bentley Technical Support Group Comments, Questions, and Answers with AutoPIPE's Output "Mode Shape" Sub Report: Item #1: There is more information available under the following, File Print Reports MODAL Output command not available in the normal output reports. This report contains 3 sections: Displacement INDIVIDUAL MODAL CONTRIBUTIONS FOR RESPONSE SPECTRUM RESULTS Local Forces and Moments Relative Acceleration: Item #2: For a model with a single segment on one axis (X-axis) with no bends, why are the R1 inertial forces in X directions {FX(Point 10) = 350.87 N and FX(Point 100) = 360.91 N}, are greater than zero, if the total modal captured mass in X direction is 0? Answer: When reviewing the output report the participation values may be 0.00 but the actual value is much smaller than can be reported. See the following procedure: 1. Open the model 2. Analyze the model 3. Select Results Grids 4. Select Frequency Tab in the grid 5. Select any cell in the Particip. Factor X column, note that the value is not 0.00 but much smaller, ex: -5.2860372e-011. AutoPIPE output report is not able to show such small values. Note: The anchor reaction is more for the model with higher cut-off frequency as it has higher level of discretization and thus able to capture higher modes better. Item #3: For a model with a single segment on one axis (X-axis) with no bends, Why for a very small cumulative modal mass of 2.0651929E-19 in X direction (corresponding to the 6th mode) such greater R1 forces values (FX=350.87 N and FX=360.91 N) have been obtained in anchors 10 and 100, while for a large cumulative modal mass of 7.9421875E+01 (corresponding to the 17th mode) lower R1 forces (FX=277.91 N and FX=271.42 N) have been calculated, accordingly? Answer: Look at the special modal report in ITEM #1 above to review the contribution of each mode and the missing mass correction in the total response. This can be reviewed at the pipe force and moments and not for the anchor reactions which may be slightly higher than the pipe FX force due to contribution from the anchor. In addtion, see the following observations: Item #4: In the Mode shape report, what are the units for the columns 2: Translation & Rotation. Answer : There are no actual displacements, it just a shape with any amplitude, positive or negative. Modal Displacements (or mode shapes, they are one and the same) are mass normalized in AutoPIPE. There are no modal stresses. Modal analysis is a requirement for all dynamic analysis in AutoPIPE, such as response spectrum, time history, force spectrum and harmonic loads. These analyses will have true displacements and true stresses. Modal displacements are not true displacements and that is why they are commonly normalized to 1 or mass normalized. Mode shapes or resonance shapes or natural vibration shapes are real and when a system is excited it most likely will vibrate in one or a combination of many mode shapes. It all depends on the loading pattern and loading frequency. Item #5: 2 models with exactly the same input except for ,modal analysis cutt off frequency, why are the first set of mode shapes NOT exactly the same? Answer : Changing the cut-off frequency when you have the automatic mass points option will lead to different level of discretization in the model. That should explain the differences in the frequencies. Rather, I would not expect the first 7 frequencies to be same as the two models are no longer "equivalent" - because the level of discretization is not same. Higher the cutoff frequency, would be a higher level of discretization. Item #6: Is there an animation tool to visualize deformations or modal analysis. Answer: After analyzing a model with modal analysis, select Results Mode Shape select a mode shape number and check the box "Animate mode shape" Item #7: I have run a modal analysis in order to capture a piping systems natural frequency. I have also taken some measurements of the amplitude of vibration from the actual piping which is vibrating at the natural frequency. I can now use the amplitudes of my measurements to find a ratio between the unit-less amplitudes of the modal analysis and the actual amplitude of vibration (at the same points). Using this factor I can find the actual amplitude of all points from the modal analysis. However I also need the angle of rotation at each point, I have tried applying the same factor to the unit-less rotation but I am getting far too high values. My question is what units and factors are applicable in order to convert rotations from a modal analysis into actual rotations. Answer: The mode shapes and frequencies are not unit dependent since they are scaled. However in order to understand the rotation units you would need to make a simple system and test this relationship. I am attaching a simple cantilever model (restrained in x-direction) with 5 nodes at 1.0ft or 304.8mm distance (total of 5ft). MODE SHAPE Point Frequency TRANSLATIONS ROTATIONS name Mode (Hertz) X Y Z X Y Z ------ ---- --------- ------- ------- ------- ------- ------- ------- *** Segment A begin *** A00 1 67.6917 0.000 0.000 0.000 0.000 0.000 0.000 A01 1 67.6917 0.000 0.000 0.270 2.212 0.000 0.000 A02 1 67.6917 0.000 0.000 0.931 3.739 0.000 0.000 A03 1 67.6917 0.000 0.000 1.841 4.636 0.000 0.000 A04 1 67.6917 0.000 0.000 2.876 5.024 0.000 0.000 A05 1 67.6917 0.000 0.000 3.947 5.111 0.000 0.000 I n English units: If length in ft: Rotation at A05 = (3.947-2.876)/1.0ft = 1.071 * 4.77 = 5.111 If in mm Rotation at A05 = (3.947-2.876)/304.8mm = 3.51E-3 *1454 = 5.111 So the factor for ft is 4.77 and for mm is 1454 The development team could not confirm if this would apply to every model, but that it might help. Remember, mode shapes are mass normalized. After increasing the length from 1 ft to 2 ft; the relation still holds. Item #8: When reviewing the maximum displacement of a mode shape, I expect that the normalized maximum displacement will equal 1.0. Instead I find it equals some random number less than 0.5. Why? Answer: AutoPIPE outputs mass-normalized modal shape amplitudes. I.e. the mode shape PHI is divided by PHIT*M*PHI = modal mass where PHI is the mode shape vector and M is the mass matrix. Item #9: is it possible to extract ‘modal stress amplitudes’ from a piping modal analysis? We are performing a fatigue analysis of a pipe system subject to ‘Vortex Induced Vibration’ (VIV), and have extracted the natural frequency of piping spans. We are then applying uniform loading to the spans under consideration, and extracting maximum deflections and bending moments for conversion into ‘Unit Diameter Stress Amplitude’. The verification body, DNV, has reviewed our draft submission and is suggesting that instead we should be using ‘modal stress amplitude’ and converting that into a ‘unit diameter stress amplitude’. The logic here is that the mode shape may have a different deflection profile than that generated by uniform loading. We can extract a mode shape and normalized nodal displacements, but can we also extract ‘modal stress amplitude’ as DNV thinks we ought to? Answer: At this time (AutoPIPE V8i 09.06.xx.xx and lower), AutoPIPE does not have a formal VIV analysis. SSD has wrote a processor that reads AutoPIPE output data in order to perform rigorous VIV analysis. There is a way to output the mode shapes and element modal forces and moments to an external file. However, we do not output the stresses as there is no input forcing function assumed. Please note that the mode shapes in AutoPIPE are mass normalized. You can enable such output by editing the file autopipe.ini in the AutoPIPE folder. You can then use the command Result/Export Eigen values (*.txt) to generate the file Modelname_EIG.txt. The headers should help you see the type of data being printed. Unfortunately we do not have any documentation or guidance on how to use such data. We also do not offer any support on this feature at this time. We hope to incorporate and improve this feature in the future. Item #10: Q1. The mode shapes only give the relative deflection which is normalized to one. A1. Normalization to one is common, but mass normalization of mode shapes is more common as in AutoPIPE. Q2: How can I know the actual displacement of each node? A2: There are no actual displacements, it just a shape with any amplitude, positive or negative. Q3: How can I know the actual stress of the modal? A3: There are no stresses since no loading is defined Q4: From the help file, I understand that the stress is mass normalized. How can I get the stress results in term of displacement normalized? A4: Modal Displacements (or mode shapes, they are one and the same) are mass normalized in AutoPIPE. There are no modal stresses. Modal analysis is a requirement for all dynamic analysis in AutoPIPE, such as response spectrum, time history, force spectrum and harmonic loads. These analyses will have true displacements and true stresses. Modal displacements are not true displacements and that is why they are commonly normalized to 1 or mass normalized. Mode shapes or resonance shapes or natural vibration shapes are real and when a system is excited it most likely will vibrate in one or a combination of many mode shapes. It all depends on the loading pattern and loading frequency. Item #11: Why are the results from our in house program are different from the modal results in AutoPIPE? Answer: "The modal displacements and rotations describe the set of natural "shapes" or "patterns" of the system when vibrating (no external load). The model shapes depend on how the system's mass and stiffness are distributed. The displacements and rotations are not absolute. In fact, some programs report the normalized modal displacements relative to the maximum modal displacement. The frequency and participation factor for each mode determines how that mode will contribute to the overall system's response from a dynamic external load. It depends on how close the frequencies of the external load are to the frequencies of the mode shapes. To duplicate the results of the other program that you are used to, scan the results for the maximum values for each direction and divide all values by the maximum. Excel can help you accomplish this. See Also Bentley AutoPIPE External Links Bentley Technical Support KnowledgeBase Bentley LEARN Server Comments or Corrections? Bentley's Technical Support Group requests that you please submit any comments you have on this Wiki article to the "Comments" area below. THANK YOU!
↧