Superelevation and Sight on Curves

Banking the road lets friction do less of the work of turning the car.

Take a curve with a 1,200 ft radius at 60 mph. To follow it, your car needs a sideways push toward the center of about a fifth of its weight. On flat pavement only the tires can supply it, and on a wet road at 60 mph, designers count on them for just 12% of the car's weight. The other 8% comes from the road itself, tilted toward the inside of the curve.

How much tilt, and how sharp a curve it buys, is the first half of this Lesson. The second half is the other thing a curve takes from a driver: the view of the road ahead.

Two ways to push a car around a curve

Anything following a circle of radius RR at speed vv needs a net force toward the center of mv2/Rmv^2/R. Divide by the car's weight, mgmg, and the push the car needs, as a share of its weight, is v2/gRv^2/gR. At 60 mph, or 88 ft/s, on a 1,200 ft radius, that's 882/[32.2(1,200)]=0.2088^2/[32.2(1{,}200)] = 0.20, the fifth from the Hook.

Two things can supply it. The first is the tires, gripping the pavement sideways. That grip, as a share of the car's weight, is the side friction factor ff. The second is the road. Tilt it down toward the inside of the curve, and the pavement's push on the car, always square to the surface, leans toward the center. The tilt is written like a grade, as the drop across the road per 100 ft, and it's called the superelevation ee, in percent. An 8% superelevation drops about 1 ft across a 12 ft lane.

The bank's share is Wtan⁡θW\tan\theta, and tan⁡θ\tan\theta is the slope of the pavement, exactly e/100e/100. So an 8% bank tips 8% of the car's weight toward the center, and the tires make up the rest. Together they have to cover what the car needs:

0.01e+f=v2gR0.01e + f = \frac{v^2}{gR}
toward center of curve 100 8 e = 8% inward push, as a share of the car’s weight W bank: 0.08 W tires: 0.12 W needed: 0.20 W 60 mph on a 1,200 ft radius
The bank's share and the tires' share add up to the push the car needs
The slope of the road is exaggerated so the tilt shows.

The exact balance has a small product of ee and ff in it, worth about 1% here, and the Handbook drops it. Now put the speed in mph, v=1.467Vv = 1.467V. With g=32.2 ft/s2g = 32.2\ \text{ft/s}^2, the constant is 1.4672/32.2=1/14.961.467^2/32.2 = 1/14.96, which rounds to 1/151/15. That gives the equation in the Handbook's Horizontal Curves table, on the row for side friction factor:

0.01e+f=V215R0.01e + f = \frac{V^2}{15R}

The 15 is gg with the speed in mph, just as the 30 in Stopping Sight Distance was 2g2g. RR is in feet, and two habits keep the rest honest:

  • ee goes in as a percent. The Handbook defines ee in percent, so an 8% superelevation is e=8e = 8, and the 0.01 turns it into 0.08 of the weight. Enter 0.08 and the bank comes out a hundred times too weak.
  • VV stays in mph. The 15 already holds the conversion.

Why the bank and the tires have limits

More bank and more grip buy a sharper curve, so why not bank every curve at 20%? Run the thought experiment. An ice storm hits, and a car stops on the curve, or a truck crawls around it. A stopped car needs no push toward the center at all. For it, the bank is just a slope across the road, pulling it downhill toward the inside, and ice may not hold it there.

So design caps the superelevation at a maximum, emaxe_{\text{max}}. AASHTO's design guide puts it around 8% where snow and ice are common, and no higher than 10 to 12% anywhere.

Side friction is capped too, and the cap falls as speed rises. Design ff isn't the most a tire can give before it skids. It's a sideways push drivers find comfortable, kept well short of a skid on wet pavement. Speed shrinks it two ways: a wet road grips less the faster the tire rolls over it, and drivers at high speed put up with less sideways shove. AASHTO's values run from about 0.20 at 30 mph to 0.14 at 50, 0.12 at 60, and 0.10 at 70. The Handbook prints no emaxe_{\text{max}} or ff values, so a problem gives you both.

Checkpoint

A northern state is setting design limits for three roads that all see winter ice: a rural mountain highway, a 70 mph freeway, and a downtown street with a signal every block. Which statements are sound?

A.
On the mountain highway, the danger of a steep bank is a car at the design speed sliding off the outside of the curve.
B.
The freeway's curves should be designed with a smaller side friction factor than the downtown street's.
C.
The downtown street should get a gentler maximum superelevation than the mountain highway.
D.
One side friction factor can serve all three roads, since a tire grips wet pavement equally well at any speed.
E.
On the mountain highway, a truck crawling around an icy curve is at more risk from a steep bank than one moving at the design speed.

The sharpest curve for a design speed

A design uses both limits at once. With the bank at emaxe_{\text{max}} and the tires at their design ff, the left side of the equation can't get any bigger, so RR can't get any smaller:

Rmin=V215(0.01emax+f)R_{\text{min}} = \frac{V^2}{15(0.01e_{\text{max}} + f)}

A sharper curve would ask the tires for more than ff, so round RminR_{\text{min}} up, never down. If the design is stated in degree of curve, R=5729.58/DR = 5729.58/D from Horizontal Curve Geometry turns the minimum radius into a maximum degree of curve, which rounds down.

Worked Example The curve from the Hook
1 of 3

A highway has a 60 mph design speed, emax=8%e_{\text{max}} = 8\%, and a design side friction factor of 0.12. Find the minimum radius and the maximum degree of curve.

Checking a curve that's already built

An existing curve fixes RR, and the question asks about one of the other letters. It's the same equation, solved for whichever one is missing. The superelevation a curve needs is whatever the tires don't cover, and the friction it demands is whatever the bank doesn't:

e=100(V215R−f)f=V215R−0.01ee = 100\left(\frac{V^2}{15R} - f\right) \qquad\qquad f = \frac{V^2}{15R} - 0.01e

Either number is only half an answer. A required ee means something only next to emaxe_{\text{max}}, and a friction demand only next to the allowable ff. If the demand is higher, the curve is too sharp for that speed.

Checkpoint
1 of 2

A curve with a 1,000 ft radius carries a 55 mph design speed, with a design side friction factor of 0.13. What superelevation does it need, in percent?

%

Solve for VV instead, and you get the fastest speed the curve can carry with its bank and the allowable friction both used up. Round it down:

Vmax=15R(0.01e+f)V_{\text{max}} = \sqrt{15R(0.01e + f)}
Checkpoint

An old curve has a 600 ft radius and 6% superelevation. Using a side friction factor of 0.15, what is the maximum safe speed?

mph

What moves the minimum radius

Read Rmin=V2/[15(0.01e+f)]R_{\text{min}} = V^2/[15(0.01e + f)] for its shape. With ee and ff fixed, the radius grows with the square of the speed, the same square that sits in v2/Rv^2/R. Double the design speed, and the minimum radius quadruples. At real design values it grows faster still, because the allowable ff drops at the higher speed.

More superelevation helps, but only by its share of the total. At 60 mph with f=0.12f = 0.12, raising emaxe_{\text{max}} from 6% to 8% grows the total from 0.18 to 0.20, and the minimum radius falls from 1,333 ft to 1,200 ft. That's 10%, not the quarter that 6 to 8 might suggest.

Checkpoint
1 of 2

At a 40 mph design speed, a road's minimum radius is 480 ft. With the same maximum superelevation and side friction factor, what is the minimum radius at 60 mph?

ft

Seeing around the inside of a curve

Now drive the curve again, in the inside lane. A building, a wall, or the face of a rock cut stands on the inside of the curve, and it hides the road ahead. To stop for something in your lane, you need to see a full stopping sight distance SS along it.

Where does the view get cut off? Your eye and the object are both in the inside lane, and design puts both on that lane's centerline. The lane between you curves, but your line of sight runs straight. So the sight line is a chord of the arc you're driving, with SS measured along the arc between its ends. The chord swings farthest from the lane at its middle, and anything closer to the lane than that blocks the view. You've met that gap before.

28.65S/R 28.65S/R R centerline of inside lane centerline of road obstruction S (along centerline of inside lane) sight line HSO driver object O
The sight line is a chord of the inside lane's centerline

Here the arc is just the stretch of lane you need to see, SS long. Its central angle is S/RS/R radians, which is 57.30S/R57.30S/R degrees, so half of it is 28.65S/R28.65S/R degrees. Put that in for I/2I/2:

HSO=R[1−cos⁡(28.65SR)]HSO = R\left[1 - \cos\left(\frac{28.65S}{R}\right)\right]

That's the Handbook's equation, on the sight distance row of the same Horizontal Curves table. It calls HSOHSO the horizontal sight line offset and SS the sight distance, in feet. The 28.65 is 90/π90/\pi, so the cosine's argument is in degrees, though the Handbook doesn't say so. Put your calculator in degree mode.

The Handbook calls RR just the radius. Here it's the radius of the inside lane's centerline, where the eye and the object sit, and HSOHSO is measured from that same line. On a two-lane road, if a question gives the radius of the road's centerline, take off half a lane. The equation also assumes the whole sight distance lies on the curve, which is the usual case.

Worked Example Clearing the inside of the Hook's curve
1 of 3

The 60 mph curve from earlier has a 1,200 ft radius to the road's centerline and two 12 ft lanes, on level ground. Using t=2.5 st = 2.5\ \text{s} and a=11.2 ft/s2a = 11.2\ \text{ft/s}^2, how far from the inside lane's centerline must obstructions stay?

Working back from an obstruction

Sometimes the building is already there, and the question is how far a driver can see past it. Run the equation backward. Solve for the cosine, then take the inverse cosine, still in degrees:

cos⁡(28.65SR)=R−HSOR⇒S=R28.65cos⁡−1(R−HSOR)\cos\left(\frac{28.65S}{R}\right) = \frac{R - HSO}{R} \quad\Rightarrow\quad S = \frac{R}{28.65}\cos^{-1}\left(\frac{R - HSO}{R}\right)

The Handbook prints only the forward form, so rearrange it yourself. That SS is the sight distance the curve offers. If the question wants a speed, set the stopping sight distance equal to it and solve for VV as a quadratic, the way Stopping Sight Distance did. Keep the positive root and round down.

Worked Example A rock cut on a tight curve
1 of 3

On a level road, a rock cut stands 20 ft from the centerline of the inside lane, which has a 500 ft radius. Using t=2.5 st = 2.5\ \text{s} and a=11.2 ft/s2a = 11.2\ \text{ft/s}^2, find the sight distance and the maximum safe speed.

What moves the offset

On real curves the angle is small, and the offset comes out very close to S2/8RS^2/8R. For the Hook's curve that's 565.52/[8(1,194)]=33.5 ft565.5^2/[8(1{,}194)] = 33.5\ \text{ft}, against 33.3 ft from the full equation. Use it to judge how the offset responds, not to compute it.

Checkpoint
1 of 3

A curve needs 20 ft of clear offset for 400 ft of sight distance. About how much would it need for 600 ft of sight distance on the same curve?

A.
13 ft
B.
30 ft
C.
45 ft
D.
80 ft

Recap

  • A curve needs an inward push of v2/gRv^2/gR of the car's weight. The bank supplies 0.01e0.01e and the tires ff: 0.01e+f=V2/15R0.01e + f = V^2/15R, with ee in percent and VV in mph.
  • Rmin=V2/[15(0.01emax+f)]R_{\text{min}} = V^2/[15(0.01e_{\text{max}} + f)], rounded up. Solved for ee, ff, or VV, the same equation checks an existing curve against its limits.
  • The minimum radius grows with V2V^2. More superelevation shrinks it only by its share of 0.01e+f0.01e + f.
  • emaxe_{\text{max}} is capped so slow or stopped vehicles don't slide inward on ice, and design ff falls as speed rises.
  • The sight line is a chord of the inside lane's centerline, so HSO=R[1−cos⁡(28.65S/R)]HSO = R[1 - \cos(28.65S/R)] is its middle ordinate, with the cosine's argument in degrees and RR to that centerline. It grows with S2S^2 and shrinks as RR grows.