In order to maintain a constant water deep in canal, how much water must flow trought the pipe ? enter image description here

As shown on picture, canal have a rectangular shape. I don't know if canal length have an influence.

EDIT : to simplify things let's consider there is no turbulence, no viscosity, and that water falling from pipe do not disturb water in canal.

I tried to solve the problem by myself (i'm a physics beginner so it could be totally wrong, please do not downvote the question if you think this is not correct) :

Area of canal section : $A = w \, h$

If I calculate water velocity $v$ in canal, using this and surface $A$, I can calculate how much water $Q$ will flow :

$$Q = A \, v$$

and solve the problem...

So only thing left is to calculate $v$.

Let's say the canal have no inclination $Z = 0$, I think water velocity for a given water height can be calculated like this (I'm not sure about this) :

$$ v = \sqrt{ 2 \, g \, h } $$

$$ \delta Q = A \, v = w \, \delta h \, \sqrt{ 2 \, g \, h } $$

integrating h from 0 to H and gives :

$$ Q = w \, \sqrt{ 2 \, g } \int_0^H h^{1/2} \, dh$$

so discharge for a given height and width :

$$ Q = \frac{2}{3} \, w \, \sqrt{ 2 \, g } \, H ^{3/2} $$

Could anyone tell me if the above is correct (assuming there is not inclination), and try to answer my initial question ?

  • $\begingroup$ Your expression for v is not correct, as Bernouilli's law is only valid along streamlines. $\endgroup$
    – Bernhard
    Commented Jun 13, 2012 at 22:10
  • $\begingroup$ Hi tigrou, and welcome to Physics Stack Exchange! I'm adding the homework tag because your question seems like a homework-like question (not necessarily from an actual homework assignment). Could you clarify whether that's the case, or why you're asking this question? Note that there's nothing wrong with it being a homework question! We're just interested to know what kind of answer will be most useful to you. $\endgroup$
    – David Z
    Commented Jun 14, 2012 at 2:26
  • $\begingroup$ Hi, first of all this is not homework. I got interest for fluid dynamics and engineering recently and read some of information about it. This is a question i'm looking for since a long time without finding an answer. I'd like to know if Bernoulli's principle can be applied here and if there is a simple answer to this problem. Maybe some conditions are missing. People trying to answer the question are free to change/add some conditions (or neglect some details) if it makes solving the problem easier. $\endgroup$
    – tigrou
    Commented Jun 14, 2012 at 11:42
  • $\begingroup$ @tigrou Why did you make this question community wiki? $\endgroup$
    – Bernhard
    Commented Jun 17, 2012 at 10:57

2 Answers 2


From the potential application of the question, I assume that question is to provide design parameters for a water slide. Because the depth of the water will be reduced from acceleration, the real question is how to add resistance to the canal. Resistance can be provided through the disruption through the use of twists and turns. I would generate empirical data of an actual set up for model. A water flow meter could show how various degree turns affect the speed reduction of the fluid. The use of a lower viscosity fluid could help to reduce the size of the pipe for modeling purposes. Because the fluid is not strictly contained within the diameter of a pipe, common equations are not applicable, however someone on this forum may provide something of use in the modeling area.

  • $\begingroup$ You are right. I initially asked this question because I was wondering how river rafting ride builders determine how much water must flow trough the ride to maintain a certain deep. I should probably have mention this is in OP. As you said, the path has many turns and curves which might make it difficult to calculate. Maybe they build a smaller scale model first, or is this based mostly on experience ? $\endgroup$
    – tigrou
    Commented Apr 19, 2015 at 12:11

I think there is a fundamental difference between the inclined slope and a flat surface.

If you consider the sloped case, you have gravity as a driving force to accelerate the flow downwards. As the fluid accelerates, you know from mass conservation that the value of $h$ along the slope will decrease. But at some point, you got friction at the wall preventing further acceleration. Lets focus at this far downstream point where $$ \frac{dh(x)}{dx}=0$$

Therefore, we also know that $\vec{u}(x,y)$ reduces to $u(y)$, only a velocity component in the direction parallel to the bottom, that is only varying in the direction perpendicular to the bottom.

We know from experience that this kind of laminar flow will give a Poisseuille profile, so we assume that $u(y)=ay^2+by+c$ and boundary conditions.

  • At $y=0$, $u(0)=0$
  • At $y=h$, $\frac{du}{dy}=0$, because there is no slip at the free surface.
  • We now that mass is conserved, thus $\int_0^h u(y)dy=Q$ with $Q$ the flow rate. Note that $Q$ has dimensions $m^2/s$ as this case is two-dimensional.

This give $c=0$, $b=-2ha$ and $a=-\frac{2}{3}\frac{Q}{h^3}$ respectively, and thus $$u(y)=\frac{3Q}{h^3}(yh-\frac{1}{2}y^2)$$

Now we will balance the friction along the bottom with the gravity force on the bulk. This is valid, as there is no pressure gradient acting (the pressure at the top is uniform ambient pressure).

The friction force along the bottom over a length $\Delta x$ equals $\mu\frac{du}{dy}(0)\Delta x$ and the gravity force in that direction equals $\rho h g \sin \alpha$, where $\mu$ is the viscosity, $\rho$ the density and $\alpha$ the angle of the slope with the horizontal.

$$friction = gravity\\ \mu \frac{3Q}{h^2} \Delta x = \rho h \Delta x g \sin \alpha \\ h=\sqrt[3]{\frac{3\mu Q}{\rho g \sin \alpha}}$$

From this expression, you see that for increasing flow rate, the height of the film increases. Also when you decrease the slope angle, the film thickness will increase.

Finally, increasing the viscosity, will increase film height. This makes sense when you realize that when there is more friction, the same area can withstand more mass.

  • $\begingroup$ This has been flagged as "not an answer", and in it's current form I think it teeters on the edge: it's rather more like a long comment. That said, I can see where you are going with this and think that the addition of a little quantitative discussion could improve it considerably. $\endgroup$ Commented Jun 14, 2012 at 15:13
  • $\begingroup$ @dmckee I had the homework tag and this discussion in mind meta.physics.stackexchange.com/questions/714/… I've provided a lot of steps that will give you at least some answer. Maybe I can extend my answer later, but not without doing the exact derivation myself. Actually I wrote my answer without pen or paper in the neighborhood, making it tough to add more then I did. $\endgroup$
    – Bernhard
    Commented Jun 14, 2012 at 20:35
  • $\begingroup$ @tigrou Does it answer your question, or is anything unclear? $\endgroup$
    – Bernhard
    Commented Jun 17, 2012 at 18:26
  • $\begingroup$ There is two things i'm not sure to understand : 1) from v = ax^2 + bx + c how did you find that c = 0, b = -2ha, a = -2/3... ? (you integrate v(x), but then ?) 2) you said gravity force = phg * sin(alpha) where does this formula came from? (i understand why sin is here and what it does but not the rest). $\endgroup$
    – tigrou
    Commented Jun 17, 2012 at 21:01
  • $\begingroup$ @tigrou 1) From the three conditions I give above v(0)=c=0. v_x(h)=2ah+b=0, and then the integral. 2) This is the force of gravity. It is mg, and $m=\rho V=\rho h \Delta x$, since the system is 2D. $\endgroup$
    – Bernhard
    Commented Jun 18, 2012 at 6:10

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.