We've recently had discussions on the practicality of full-scale structural testing of actual homebuilt aircraft prototypes, and also some references to use of models in crash testing. So, that got me wondering if appropriately constructed physical scale models of composite aircraft can provide useful insight into structural characteristics and crashworthiness? If so, is there existing guidance in how to do it, best practices, etc?
There's broad agreement that aerodynamic dynamic models can sometimes be useful in analyzing characteristics of new aircraft designs. The most "approachable" coverage I've seen of this is Stan Hall's article on Dynamic Modeling in Sport Aviation Magazine.
Can we do something similar to learn about the >structural< adequacy of a new composite design (flight loads, crash loads, etc) through the use of physical scale models? Or, has this approach been overcome by technology, with the same thing now done better/more easily with use of computer modeling?
If someone knows of an article or reference that can provide a little handholding, please provide a link.
What I think I understand from reading a few references:
- Yes, a reasonable-size composite fuselage mockup, wing mockup, etc can provide useful strength, rigidity, and crash behavior information. If our length scale is, say, 1/4, then our areas will be 1/16, and our volumes will be 1/64. Our weights would need to be 1/64. We should use the same materials (core materials, reinforcements, resin) in the model as we'd use in the prototype full-scale article.
Here are some basic similitude relationships for a 1:5 model. The source is https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&ved=2ahUKEwiPzsCI2I78AhUKMlkFHQvjCa0QFnoECBoQAQ&url=https://ntrs.nasa.gov/api/citations/19990036755/downloads/19990036755.pdf&usg=AOvVaw23--X_EGXu6p6L5d6CANP4
-- Layup schedules/orientation, etc: There's been quite a bit of study on this, To get an accurate representation of the behavior of a full-size model, the layups in the model have to be as close as practical to matching (at scale) the layups used in the full-size article. It appears to me that, if possible, the optimum approach is to use laminate plys that are in the same number and fiber orientation as the full-size article will be. This might require some darn thin fabrics, but a quick search seemed to indicate that suitably thin fabrics are available in fiberglass, I'm not sure about CF.
--- Obviously, this has limits. For example, if we prepare a surface for subsequent bonding by applying some peel-ply to be torn off later, the remaining rough surface is probably going to be the same thickness on a full-size plane as on our model. Where possible, these situations should be avoided to keep the thickness of the total laminate stack on the model as close as possible to the desired scale thickness.
- Dynamic scaling: I'm still trying to understand all the ramifications of this. The chart above is from a study in which a 1/5th scale composite fuselage section was drop tested to test various types of energy absorbent materials below the seating area. (the study, linked above, is worth reading for anyone interested in this stuff). The full-size airliner impact speed of interest was 31 ft/sec. and they dropped the model at the same velocity. Because the available crush distance of the model was just 1/5th of the full scale airplane, it resulted in an acceleration of the seats/floor 5 times as great (125G for the model, 25G for the real airplane they were simulating). Likewise, the pulse duration was 1/5th as long. I know that some composite reinforcements are rate sensitive (e.g. fiberglass composites have much higher tensile strength at high rates) and some are not (e.g. carbon fiber). So, a little more digging would be useful before moving forward.
Anyway, some grist for the mill. Maybe building a 1/4th scale "cabin" would help validate the calculations indicating that our pusher engine will stay out of the occupied area of the aircraft at the impact speed of interest, it might let us know if the cabin remains intact through various loadings. Maybe we could build a 1/4 scale wing half and test it to destruction without crying a lot over the expense and time of making it (compared to a full-scale wing).
There's broad agreement that aerodynamic dynamic models can sometimes be useful in analyzing characteristics of new aircraft designs. The most "approachable" coverage I've seen of this is Stan Hall's article on Dynamic Modeling in Sport Aviation Magazine.
Can we do something similar to learn about the >structural< adequacy of a new composite design (flight loads, crash loads, etc) through the use of physical scale models? Or, has this approach been overcome by technology, with the same thing now done better/more easily with use of computer modeling?
If someone knows of an article or reference that can provide a little handholding, please provide a link.
What I think I understand from reading a few references:
- Yes, a reasonable-size composite fuselage mockup, wing mockup, etc can provide useful strength, rigidity, and crash behavior information. If our length scale is, say, 1/4, then our areas will be 1/16, and our volumes will be 1/64. Our weights would need to be 1/64. We should use the same materials (core materials, reinforcements, resin) in the model as we'd use in the prototype full-scale article.
Here are some basic similitude relationships for a 1:5 model. The source is https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&ved=2ahUKEwiPzsCI2I78AhUKMlkFHQvjCa0QFnoECBoQAQ&url=https://ntrs.nasa.gov/api/citations/19990036755/downloads/19990036755.pdf&usg=AOvVaw23--X_EGXu6p6L5d6CANP4
-- Layup schedules/orientation, etc: There's been quite a bit of study on this, To get an accurate representation of the behavior of a full-size model, the layups in the model have to be as close as practical to matching (at scale) the layups used in the full-size article. It appears to me that, if possible, the optimum approach is to use laminate plys that are in the same number and fiber orientation as the full-size article will be. This might require some darn thin fabrics, but a quick search seemed to indicate that suitably thin fabrics are available in fiberglass, I'm not sure about CF.
--- Obviously, this has limits. For example, if we prepare a surface for subsequent bonding by applying some peel-ply to be torn off later, the remaining rough surface is probably going to be the same thickness on a full-size plane as on our model. Where possible, these situations should be avoided to keep the thickness of the total laminate stack on the model as close as possible to the desired scale thickness.
- Dynamic scaling: I'm still trying to understand all the ramifications of this. The chart above is from a study in which a 1/5th scale composite fuselage section was drop tested to test various types of energy absorbent materials below the seating area. (the study, linked above, is worth reading for anyone interested in this stuff). The full-size airliner impact speed of interest was 31 ft/sec. and they dropped the model at the same velocity. Because the available crush distance of the model was just 1/5th of the full scale airplane, it resulted in an acceleration of the seats/floor 5 times as great (125G for the model, 25G for the real airplane they were simulating). Likewise, the pulse duration was 1/5th as long. I know that some composite reinforcements are rate sensitive (e.g. fiberglass composites have much higher tensile strength at high rates) and some are not (e.g. carbon fiber). So, a little more digging would be useful before moving forward.
Anyway, some grist for the mill. Maybe building a 1/4th scale "cabin" would help validate the calculations indicating that our pusher engine will stay out of the occupied area of the aircraft at the impact speed of interest, it might let us know if the cabin remains intact through various loadings. Maybe we could build a 1/4 scale wing half and test it to destruction without crying a lot over the expense and time of making it (compared to a full-scale wing).
Last edited:
