Assume we have a very long cable whose ends are not visible or accessible. One end we know is a generator, and the other a load. Can we determine which end of the cable has the generator attached without cutting the cable?
A cable consists of two conductors, with a finite space between them. Let's assume they're insulated.
Use capacitive sensing to determine the polarity of the instaneous voltage between them as a function of time.
Use a Hall sensor to determine the polarity of the magnetic field in the space between them as a function of time.
Once you've get the left-hand/right-hand rule polarities right, the vector product will indicate the direction of power flow.
As you only wanted direction, polarities are sufficient. However, if you can get the indications to be more accurate, like a measurement, then you can estimate the magnitude of the power flow.
As Marcus points out in comments, this is just an application of the Poynting Vector, which is true for electromagnetic waves in all of space, and for DC as well, but here is nicely simplified by there being a waveguide (power guide?) present in the form of your conductors.
Since you can measure
- the electric field (red arrows) between both lines of the cable and
- the magnetic field (green arrows) around each line of the cable
without cutting it, you can determine the Poynting vector (blue arrows), which is the cross product of both vector quantities.
The Poynting vector always points into the direction of energy flow as illustrated in this image (from Wikipedia).
Note: in case of AC the direction of current (magnetic field) switches but also does the direction of voltage (electric field) so the direction of the Poynting vector (energy flow) stays alwyas the same.
If you just want to know the direction of energy flow (not the quantity) it would be sufficient also to measure only the direction of the electric and magnetic fields.
If you can access a relatively longer section of the cable and if you have a contactless meter (aka clamp type meter) :
There'll always be a voltage drop across the cable. Measure the voltages at both ends using a clamp meter. If your meter is "sensitive" enough and if the voltage drop across the cable is "noticeable" enough , you'll observe that the voltage at one end is higher than the other. So this end is connected to generator side.
Yes, if you measure the instantaneous power P=V*I you get a positive or a negative result.
To do this you need to measure the voltage and the current. Both can be done without cutting, with clamp current meters and needle probes to punch through the insulation.
You can do this with the math channel on your scope, then measure the mean value of the math channel.
This is basically how energy meters work.