Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
facc Aerostructures Extend Your Expectations “From Theory to Reality – Peter Glaser, Ho Stress imposing real world requirements on the world of numerics.“ This document contains information which is proprietary to FACC or other companies. Any reproduction, disclosure or use of this information without FACC‘s prior written consent is expressly prohibited.. Page 1 Abstract The ever increasing performance of computing tools provides unprecedented possibilities to evaluate the last single detail of, for example, a structural product. While this, per se, provides numerous novel and useful possibilities, it calls for an intelligent choice of when to take advantage of such possibilities and when they do not make a difference. When input parameters are of coarse nature, then post processing to a high precision is purely numeric, but non relevant. For example, if you are going to measure the consumption of a car over 1 km, you don’t need to measure this distance with an accuracy of a mm. This presentation will discuss similar aspects in the world of structural analysis. This document contains information which is proprietary to FACC or other companies. Any reproduction, disclosure or use of this information without FACC‘s prior written consent is expressly prohibited.. Page 2 Introduction „Dimensional Analysis“ In the AC stress analysis, the RF threshold of 1.00 separates flightworthy from non- flightworthy. Sometimes complete design concepts get cancelled because the minimum RF on paper remains at a value of e.g. 0.95 and noone can think of a “pencil sharpening” option. On other occasions, simple, apparently non-dramatic modeling updates lead to “critical” Rf changes in excess of 30 percentage points up or down. This curious fact raises a number of questions. Within the scope of this presentation, the focus lies on the following: ”Is enough attention paid to the variation of parameters when computing single number RFs?” The collection of arguments herein are source for discussion and food for thought and may in some instances not withstand full examination. They are not suggested practice at the moment. This document contains information which is proprietary to FACC or other companies. Any reproduction, disclosure or use of this information without FACC‘s prior written consent is expressly prohibited.. Page 3 Parameters and Variations Loads Load application accuracy; Winglet sectional loads; Spoiler load shapes; load splits between top and bottom surfaces, etc. This document contains information which is proprietary to FACC or other companies. Any reproduction, disclosure or use of this information without FACC‘s prior written consent is expressly prohibited.. Page 4 Parameters and Variations Manufacturing / Handling Tolerances Draping angle influence vs. manual layup accuracy. & 5° positioning tolerance on dwg. Measurement accuracy of defect in BA vs. grid density. B-Values “improbable” 95% of time + redistribution… Ultimate Factors related to probability of failure, This document contains information which is proprietary to FACC or other companies. Any reproduction, disclosure or use of this information without FACC‘s prior written consent is expressly prohibited.. Page 5 Operator Behaviour FEM Detail H/C allowable - 3 D Mesh Laminate Allowable coupon size – Shell elm size Differences in attention to detail – e.g.: every fastener modeled but contact phenomena ignored; Multiple elms through laminate radii modeled,… Test correlation within +-10% successful Bolted joint load share accuracy. This document contains information which is proprietary to FACC or other companies. Any reproduction, disclosure or use of this information without FACC‘s prior written consent is expressly prohibited.. Page 6 Sources for Alternatives Damage Tolerance Intentional DCFs Inherent DCFs – eg: MLP cases are not required to be assessed in composite; it is assumed that CFRP is itself an MLP. Maybe some composite failure could well be allowed without threatening product integrity. This document contains information which is proprietary to FACC or other companies. Any reproduction, disclosure or use of this information without FACC‘s prior written consent is expressly prohibited.. Page 7 Sources for Alternatives Statistics B-Values “improbable” 95% of time + redistribution… Ultimate Factors related to probability of failure, How many pcomp layers in the total quantity of composite elements actually exhibit RF of 1.0 ? How is the remaining RF distribution? Even if the entire product is built from material with the properties of the remaining 5% below the “B-value quality”, will that product fail at 1.0xUL? Which features really are critical structural elements? Shouldn’t all others be treated differently? This document contains information which is proprietary to FACC or other companies. Any reproduction, disclosure or use of this information without FACC‘s prior written consent is expressly prohibited.. Page 8 Sources for Alternatives RF bands with probability of occurence. Probability of an RF of 1.0 on paper to be the precise failure point in reality? Dependent on: Probability of Loadcase occurrence; Probabiltiy of raw material performance to exactly the B-Value performance. Probability of the applied load/stress/strain to be acurately predicted Probability of RF calculation algortihm to be exact Probability of correct level of defect in correct vicinity of position under estimation. Probability of manufacturing process to deliver exactl yat the lowest end of the required spectrum. Etc…. This document contains information which is proprietary to FACC or other companies. Any reproduction, disclosure or use of this information without FACC‘s prior written consent is expressly prohibited.. Page 9 Sources for Alternatives RF bands with probability of occurence. In one element? In the element next to the failed element?... Probabiltiy of two neighboring elements to be equally un“fortunate“. RF ? 1.0 - 5% Probability 50% 95% This document contains information which is proprietary to FACC or other companies. Any reproduction, disclosure or use of this information without FACC‘s prior written consent is expressly prohibited.. Page 10 Discussion Options for the future Can probability based RF bands realistically become standard use? OEM position? How exactly will the product variability (including defects) with the variability of other parameters get mixed to produce the RF bands? How does this get sold into the status quo? Are shortcurts available? (eg. Work with nominal geometry, as so far, but determine effect factors or equations….) ? This document contains information which is proprietary to FACC or other companies. Any reproduction, disclosure or use of this information without FACC‘s prior written consent is expressly prohibited.. Page 11 Thank you for your attention This document contains information which is proprietary to FACC or other companies. Any reproduction, disclosure or use of this information without FACC‘s prior written consent is expressly prohibited.. Page 12