Definition of Leading/Trailing Edge Points

[seahuston]

Hi All,

New to the forum and hoping I might be able to get some assistance. I work for a manufacturing company and we have just started large scale production of engine turbine blades. To meet the needs of our customers we are working on developing our own inspection and report generation software. I am tasked with most of this and am looking for help with the definition of leading edge and trailing edge points.

I am currently working with point curves generated from the 2D section profile (X,Y) of the blade. This section is, of course, an airfoil. In order to determine parameters such as chord length I need to determine the leading edge and trailing edge points. The only way I know how these are defined is from the mean camber line (MCL).
Method for MCL Determination:
Construct a MCL for my data using a modified voronoi approach to inscribed circles. This works until approximately 1 radius away from the edge.
Extrapolate a straight line from end point, find normals along the line and make sure the points are half the normal.
Fit new points to a polynomal, and contstruct normals. Make sure points are located halfway along the normal.
Iterate last step until convergence.

This method works reasonably well but my chord length still has error on the order of .005". I really need to be able to obtain values in the .0002" range of accuracy.

I am wondering if anyone here has experience with MCL or LE/TE point determination for an arbitrary airfoil. If not, perhaps you might able to assist with alternate MCL definitions to play with?

Thank YOU!

 

[rb1957]

is this a student post ? (i can't think who else would construct an airfoil section that way)

 

[seahuston]

While I am an engineering intern at my company I would not consider this a student post. I am contracted to begin work as a software development engineer upon graduation. My job description will not change between now and then.
We deal with (X,Y) data because a coordinate measuring machine is used to obtain data points after manufacturing. My background in ME/Math so understanding a bit more of airfoil properities would be helpful.
If this approach to constructing airfoils seems elementary, could you please recommend a better way to do this with the measured data? As we are a manufactuing company we have no control over airfoil shape.

 

[seahuston]

I should clarify that the understanding of airfoil properties I mention is mainly just in the calculation realm.
This process (the determination of a MCL from arbitrary data) is not well documented and seems not common. I have a found a few journals but the implementation of the topics was not successful.
Perhaps there are other ways to define the LE/TE points without the MCL?

 

[rb1957]

so you've got some X,Y data that you're trying to spline into an airfoil ?

 

[seahuston]

I've got X,Y data representing an airfoil and I need to determine the leading edge and trailing edge points for this data. The only method I see for this is to construct a MCL. I've done okay at this but the end points of the line cause difficultly. Given our measurement tools, point density isn't really a concern but I could spline and interpolate if needed.

 

[GregLocock]

Pick up that very heavy phone and talk to your customers. Their spec will either define an approach, or they haven't thought about it enough. Also talk to your inspection people.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[rb1957]

so you're measuring a physical airfoil ?

i would have thought that picking points with a vertical (ie along the Y axis) tangent would be adequate to define the LE and TE.

 

[seahuston]

Yeah, I'm measuing a physical airfoil. I think you are right in that most cases tangent points would work fine, assuming a coordinate system aligned to the chord line. This won't work for an airfoil with lots of camber though. An example of a shape similar to what I am dealing with is shown in this picture where the intersection points of the MCL (dashed line) would be the LE/TE pooints.

 

[rb1957]

if i read your pic right, and your foil is the shaded pieces with a duct running down the inside of it, without any real leading edge, well then aren't you "just" extraploating what's really there ? (like the attached pic) does it Really matter ? particularly to the accuracy you're aiming at ??

if you're doing engineering design development, just be consistent in your definition. if you've got several test results/prototypes, the best definition is the one that matches your data best and makes the best predictions.

 

[seahuston]

Sorry, I just grabbed that picture from google as an example. It is a cooled turbine blade defined by the outer black line. I choose this image since it details the typical "extreme" camber of certain turbine blades. I've attached a picture highlighting this along with the desired points being circle. For my application the only thing I have is the (x,y) data representing the outer airfoil surface. Any internal ducting is not important since this is only an external scan.