First, let me correct one statement:
haruspex said:
If the force is nonzero then it is always possible to represent this pair by a single force that does not necessarily pass through the mass centre.
I was thinking only of 2D systems
, as in this case. In general:
If the force is nonzero and the net torque is perpendicular to the net force then it is always possible to represent this pair by a single force that does not necessarily pass through the mass centre.
E.g. if we consider two equal forces at right angles but not passing through a common point, the resultant linear force and the resultant torque are parallel.
You asked me to elaborate on this:
The net of any system of forces acting on an object can be represented as the combination of a force through the mass centre and a couple.
Suppose the forces
Fi act on an object, with the line of action of force
Fi being displaced by vector
Ri from the object's mass centre.
The net force ##F = \Sigma F_i##, and the net torque about the mass centre is ##T = \Sigma F_i \times R_i##.
If we choose the line of action of
F to be through the mass centre then it has no torque about the mass centre, so the system of forces is completely represented by the combination of
F and
T.
PhysicStud01 said:
does a torque always imply a couple? a couple consists of 2 forces of equal magnitudes, right?
The pure torque component, T, can be represented as a couple: If C is any vector at right angles to T, and D is at right angles to both (e.g. D = C x T) then we can represent T as the combination of two forces kC, -kC with lines of action displaced by vectors +D , -D from the mass centre. ##T = 2kC\times D##.
If instead we choose the line of action of
F to be displaced by
R from the mass centre then it produces torque ##F \times R##. The remaining torque to represent the original system is ##T-F \times R = \Sigma F_i \times R_i - \Sigma F_i \times R = \Sigma F_i\times(R_i-R)##.
If T is orthogonal to F and F is nonzero then ##(T \times F)\times F## is a vector parallel to T. Setting ##R = kT \times F## for some suitable k we get ##T-F \times R = 0##, so we can represent the entire system of forces by a single force, maybe offset from the mass centre.
See
http://en.wikipedia.org/wiki/Resultant_force, though I don't think the explanation there is great.