130403-1533 EDT
hisham1986:
Most physical devices do not have a very precise threshold point of failure, and many are time dependent.
Some different kinds of examples:
A steel tensile test sample will have a tensile break point that is quite predictable on a tensile test machine. Take the same type of sample and subject it to a lower stress level than the tensile test machine failure point, cycle a very large number of times, and you can produce a fatigue failure.
Aluminum and eutectic tin-lead solder have much poorer fatigue failure characteristics than does steel.
A Zener diode when reverse biased conducts very little current until a threshold point is reached. Just above the threshold value the Zener becomes a very low impedance and will conduct a large current if the current is not limited. If an excessive amount of current flows the Zerner burns out. Thus, a very small change in voltage can easily produce a failure.
An automotive starter motor can produce a large amount of output power for a short time, but would burn out quickly if the on time was not limited, and there was not sufficient cool down time between uses.
What is your spinning disk kilowatt-hour meter made of? Conductors, magnetic material, and rotating elements. Small overloads will produce an accuracy problem. Sufficiently large overloads, and/or extended overload time will cause some failures. This is not a sharp threshold type of problem like the Zener diode.
.