October 3, 2026 · Problem sheet · LibreTimes
Vector calculus. Stokes' theorem. Circulation
Statement
The circulation of a field around a closed contour equals the flux of its curl through any surface spanning that contour:
where
The orientations of the contour and the surface must be matched: seen from the tip of the normal, is traversed anticlockwise (the right-hand rule). A mismatch gives the answer with the opposite sign.
Green's theorem is a special case: a plane domain, a field with no -component, the normal along .
What it is for
Any surface will do — and that is what gets used. A contour of three segments gives three line integrals of the second kind; the flat triangle spanning it gives one surface integral, often of a constant. The reverse also pays off: if the curl is complicated and the contour is a circle, it is easier to compute the circulation directly.
The steps
- Find from the determinant.
- Choose a surface spanning the contour — usually a flat one.
- Write out the normal consistently with the direction of traversal.
- Compute the flux of the curl.
If a problem asks for it "two ways", both answers must agree; that is the built-in check.
Common mistakes
- The sign of the middle component of the curl. In the determinant it is taken with a minus: , not the other way round.
- A normal inconsistent with the traversal — the answer is right in absolute value and wrong in sign.
- Forgetting that the normal to an inclined plane is not a unit vector. If is taken, the integration is over the projection with ; if is taken, it is over the area with . The two must not be mixed.
Problems
Problem 1 (Circulation around a circle, two ways). Given the vector field and the closed contour : , , , .
Find the circulation of the field along in two ways: directly and by Stokes' theorem.
Solution.
Method 1: a line integral of the second kind
The circulation is
where , , .
From the parametrisation: , , .
On the contour , so the term vanishes; the term disappears with . Only remains:
The integral of an odd power of the cosine over a full period is zero. Formally: , and both terms integrate to zero on .
Method 2: Stokes' theorem
The curl.
Compute: , ; , ; , . So .
The middle component is non-zero here, and this is the case where it is easy to lose: in the determinant it is taken with a minus sign, that is, it equals , not the other way round.
The surface. Span with the disc of radius in the plane . The traversal with from to is anticlockwise, so the matching normal is .
The flux of the curl. The dot product picks out the third component only: .
because . The same conclusion without computation: the disc is symmetric about the axis, and the function is odd in .
Both methods gave the same result, which is the check.
Answer: the circulation is .
Problem 2 (Circulation around a triangular contour). The vector field is given.
The contour is the triangle cut out of the first octant by the plane , with vertices , , , traversed in the order .
Find the circulation of the field along .
Solution.
Direct computation would take three integrals along the three sides. Stokes' theorem reduces the problem to one flux through the triangle spanning the contour — and here the curl turns out to be constant, so the integral can be done in one's head.
The curl. For , , :
Orientation. Seen from the side of positive coordinates, the traversal is anticlockwise, so the matching normal points "outwards", away from the origin:
This can be checked with the right-hand rule: going from to and then to , the thumb of the right hand points towards .
The flux of the curl.
The quantity is constant, so the flux is its product with the area:
The triangle is equilateral with side , so
A check by direct computation. On the side : , , the field is , the dot product is , and the integral is . By the symmetry of the field under the cyclic substitution , the other two sides give the same, in total. It agrees.
Answer: .
Problem 3 (Circulation around a triangular contour (variant)). The vector field is given.
The contour is the triangle cut out of the first octant by the plane , with vertices , , , traversed in the order .
Find the circulation of the field along .
Solution.
The field and the direction of the normal are the same as in the previous problem; only the size of the triangle has changed. The curl does not depend on it:
The area changes. The side is now , so
The answer grew fourfold, not twofold: the contour was stretched by a factor of two along each axis, and area is a quadratic quantity. Here the curl is constant, so the scaling is predictable; with a variable curl such an estimate cannot be trusted, and the integral must be computed again.
A check by direct computation. On the side : , , the field is , the dot product is , and the integral is . By symmetry the three sides give .
Answer: .
No comments yet
Be the first to share your thoughts.