Grading Contours not true to slope values

[SDcivil]

I am creating some simple grading objects from feature lines.  I am targeting my existing ground surface with a 2:1 fill slope.  My resulting surface shows me 5' contours, whcih I think should all be spaced out at constant 10' offsets, but they are not.  When I measure the offsets I get different vaues ranging on 10.2' to 9.7' (depeding on the section of the feature line I am looking at).  Anyone know a setting to make the slope build properly with a true 2:1 offset between contours?  What good does this grading tool do if you can't get a properlty constructed slope?

[Jeff_M]

Does your surface style have Contour Smoothing enabled? This will give you results as you describe. Smoothing the contours forces them to not go explicitly point to point between the TIN lines.

[KMercier_C3D]

Is your feature line sloped? Grading occurs perpendicular from the feature line and not from the contours.

[mathewkol]

Please attach a DWG that exhibits this behavior.

[Joe-Bouza]

Adding to the others: Is the feature line you are grading off "engineered"? If it has erratic grade breaks in it you are appt to get erratic results.

[SDcivil]

Yes it is.  You are correct that the feature line is building the slope and spacing the controus perpendicular to itself.  However, a proper slope is created by offsetting perpendicular to the contours.  Do you know of a fix?

[BrianHailey]

First of all, when grading out at 2:1, there shouldn't be any area that's less then 2:1 unless it's at an inside corner. Can you provide a drawing with the feature that you are grading from and perhaps the grading as well? I'm very curious as to why you have something flatter than 2:1.

Second, what's the steepest slope of the feature that you are grading from? According to my math (with 9.7' between contours) your feature should have a slope of around 12%. Does this sound right?

[SDcivil]

That does sound right.  I don't have a picture, but if you draw any feature line with the first grade at 15% and then the second grade at 1%, you will see that the contour spacing changes.  Counours should have a constant 2:1 offesets, regardless of the grade of your feature line.  If anyone knows a setting or work around, let me know.

For now I am asking my firm not to uses featrue lines for grading. 

[Joe-Bouza]

---

@SDcivil wrote:

<snip>

For now I am asking my firm not to uses featrue lines for grading. 

---

good luck with that one.

15% to 1% with no vc I wouldn't expect anything less than a train wreck. Are you bucking grade to boot?

[Neilw_05]

Could you explain "Grading occurs perpendicular from the feature line and not from the contours" Kati?

In my mind contours should be nominally spaced in plan view regardless of how they are created.

[brianchapmandesign]

3 guesses:

1) Something wrong with featureline

2) Something wrong with how you're measureing the offsets.

3) Something wrong with the surface

Good luck lol

[Joe-Bouza]

Another thought: is it possible the feature line was created from polyline geometry that would have benefited from a map clean operation? i.e pseudo nodes, segments doubling back on it self?

[Matt.Anderson]

Contours are drawn according to the slope of the triangle face they cross.

A grading object between two wildy skewed feature lines will produce two very differently sloped triangles, so the resulting surface smoothness will be dependent on the number of triangles.

Your best bet is to smooth this surface to guarantee a uniform set of slopes.

[BrianHailey]

---

@Neilw wrote:

Could you explain "Grading occurs perpendicular from the feature line and not from the contours" Kati?

In my mind contours should be nominally spaced in plan view regardless of how they are created.

---

When you use the grading tools in Civil 3D (aka projection grading) the criteria is applied perpendicular to the featureline. If you project out at 2:1, the 2:1 is measured perpendicular to the featureline. If the featureline has no slope, then the resultant surface has a slope of 2:1. If the featureline has a slope, you need to add the two slope vectors to get the resultant slope.

If the featureline is level (slope of 0) the resultant looks like this:

If the featureline has a slope, the 2:1 slope (aka 50%) is still measured perpendicular to the featureline but the resultant surface is a combination of the two.

To get the resultant, you add the two slopes together (remember how to add vectors?). The resultant slope is sqrt(.5^2+.1^2).

So in the case of the original poster that said his surface had 9.7' between 5' contours, this would result in a slope of 1.94:1 or 51.5%. If the slope from the featureline is 2:1 (50%), then in order to get 51.5%, you would have to solve this equation: x^2+.5^2=.515^2. Thus X would be 0.123, or 12.3%.

[brianchapmandesign]

Post an example of what you're having a problem with.

[CADvisers]

SDcivil,

In case you haven't given up...   

Is your area of concern around a curve or does it have any curvature at all, or does the object you're grading from have varying slope.     The slope should be measured perpendicular to the object you're grading from not the individual contours. 

You're probably use to create all your contours by offsetting to what ever slope (distance) you wanted.  You will find, that most of the time, the grading that you created this way, is actually not correct.  By this I mean, It was correct to someone measuring perpendicular to each contour.  But that's not necessarily correct if your slope is measured from something that's not perfectly parallel with the contours.

Do this test...  Create something like a swale with a CL create the first contour for the side slope, then offset that contour 3.00 for a 3:1 side slope.  Now measure the distance between contours going "perpendicular" to the CL of swale.  See image if you don't understand what I mean.

Note:

If you're not going to use the grading objects to create "correct" grading.  Don't bother using Coridors either.  (I'm kidding....  Just understand that what you did in the past may never have been correct.)

[Joe-Bouza]

I've bee biting my tongue for a week now:

"All Contours are lies"; all it take is one look to see : flip a face.

...and from Sinc.."The other problem you are hitting, though, is that contours are the least-reliable form of vertical data."

[SDcivil]

Maybe this will help.  Attached is a plot of a basic straight street at a constant grade (no real cleanup required).  The 10 foot contours are shown at 2:1 slope.  The grading function calculates the controur spacing at 19.6', but a proper 2:1 slope should have this dimension at 20'.  Instead the 2:1 slope ratio is calculated perpendicular to the top/toe, which is incorrect becuase it is skewed to the slope direction.   

[CADvisers]

That actually is CORRECT.  I think that's what everyone is trying to tell you.

If you have a cross-section detail of your roadway, that is an exact representation of the required slope.

[SDcivil]

Thanks, but it is not correct...technically contours should offset from each other.  Slopes ratios should be measured perpendicular to the slope (and thus controus).  I know how a typical street  cross section calls out a 2:1 slope, but the intent is always for the slope to be  constructed properly and according to geotechnical standards.  Along your reasoning every 2:1 slope would be different depending on the grade of the street.  Anyhow, its not really an issue unless you have a steep grade or large slope.  For the most part, although incorroect, it is within construction tollerances.  I appreciated all your responses, this is a great page.