Hi Ravide,
Low frequencies will not do this.
A low repetition rate of med freq to high freq pulses could.
To use an echo from the distant end in this way you would
need to apply a physical burst of longitudinal frequency of
a fairly high frequency, not too high as it has to 'bounce'
off the other end and if the frequency is too high then a
poorly defined end will give too much scatter.
The speed of travel through the steelwork is not a function
of the frequency applied.
20 k/c travels well through steelwork, and would be a low
enough frequency to ignore small imperfections and to echo
well from a poorly defined end. I would suggest using a
frequency of about 70 k/c because this is within the spec
of most audio equipment and amps to drive the transducer
(and for the receiver)
should therefore be easy to obtain.
I would suggest using two transducers one to send and one
to receive, switching arrangements for using only one are
troublesome. Arranging for the receiving end to be quenched
during the send signal is not difficult.
Circuitry listening for a frequency signal can be made very
much more sensitive than listening for a simple click or
pulse.
I don't think you will get steel rods embedded in concrete to
respond at their natural resonant frequency, which in essence
is what you appear to be describing. But bursts of high
frequency signals will travel through rods like that, and
yes they will echo back from the end, but if the end is
welded or bolted to steelwork it may give spurious effects.
Best of luck with it, John