Strain is the amount of stretch imposed on a material, expressed as a change from its original length. Typically, a pair of marks, called "gage marks", are placed on the test specimen prior to the test. As load is applied, the change in gage length is noted.
Strain is most easily pictured by imagining a rubber band. Imagine a relaxed rubber band, with two ballpoint pen marks on it, spaced 1 cm apart. If the rubber band is stretched so that the marks become 2 cm apart, then we say that the material has achieved 100% strain. With elastomers such as rubber, large strains, on the order of 500% or even more, can be achieved before breakage occurs.
In most, less elastic, engineering materials, strains are much smaller, typically below about 3%. Percentage of original gage length may be used, but often the strains are so low that another system is used: microstrain. If you express the strain as a fraction of the original gage length, so that 100% strain is 1, then multiply the result by 1 million, you get microstrain. This is a unitless ratio. You may sometimes see strain expressed with units, such as centimeters per centimeter, or inches per inch, but obviously the units cancel. "Micro" is usually followed by "strain" to identify that it is a strain measurement.
Typical engineering metals such as steel are "elastic" up to around 6000-8000 microstrain, depending on their heat treatment and work hardening histories. This means the strain will totally recover when an applied load is removed. Beyond the "elastic limit", a permanant set is taken by the metal. 2000 microstrain, or 0.2%, is frequently used to define an "offset yield" that is an important material property. Ultimate strains, at breakage, may go to ten percent or more.
Not all metals have an elastic range. Dead-soft metals are plastic at very low strains. Other materials, such as titanium, may "creep" when held at high load for an extended time. Nitinol has bizarre stress-strain relationships, including a "pseudoelastic range" which give that metal its temperature-sensitive "memory" properties.
Do not confuse strain with stress. Although they are proportional in the elastic range, stress is force per unit crossection.
Again, the rubber band is a useful model. Picture what happens when a rubber band is stretched. It not only gets longer, it narrows in the middle. Typically, if a longditudinal dimension increases by x, the transverse dimensions decrease by about x/3. This is "Poissson's Ratio". (Not all materials follow this rule: fiber composites depart greatly from it.) Strain gages are made of electrically-conductive wire or foil. As it stretches and narrows, the electrical resistance of the conductor increases. This is readily measured by using a Wheatstone Bridge configuration, and a good differential amplifier.
Bonded strain gages are the most commonly used today. These have a thin insulating backing, often a polyimide or epoxy film. A metal foil pattern is embossed on the backing, to which wires can be soldered or welded. The metal foil is usually a temperature-compensated alloy (constantan is frequently used) matched to the material being tested, to minimize temperature drift. The gage is usually attached to the object being tested using high-performance adhesives, although other methods are sometimes used.
Strain gages work in tension or compression. Strain gages are available in special patterns called "rosettes", which can simultaneously measure strain in two or more directions. These can not only measure longditudinal and transverse strain, but also shear strains. By putting shear rosettes on a cylindrical shaft, torsion strain can be measured. One special circular gage type can be placed on a surface, and then a hole is drilled in the center of the gage. The resulting strain reading is a measure of residual strain, that is, the internal strain locked into the part during fabrication.
Strain gages are a marvelous tool for instrumenting test specimens. Relatively cheap, they can be subjected to harsh test conditions that would be destructive to expensive special-purpose devices such as extensometers. With the proper selection of adhesives, they can monitor strains up to 3% or more of original gage length.
Compact and light, strain gages can be attached permanantly to structures, to allow periodic or even continuous monitoring of strain in the material. Aircraft or spacecraft components are an example. When Tuvoc reports to Captain Janeway that the stresses on the hull are exceeding maximum tolerance, undoubtedly he really means strain, for the ship is undoubtedly peppered with strain gages.
With strain gages, you can easily design your own special-purpose transducers, and even turn existing structures into precision instruments. A bolt can become a load cell. A bent piece of flat spring metal can become an extensometer or deflection gage. A socket wrench extension or drive shaft can become a torque transducer. The surface of a pressure tank can become a pressure transducer. Use your imagination!
Strain gages can be bonded to a tremendous variety of materials, even concrete. I have seen instructions for mounting them on such exotic materials as uranium and plutonium. I have seen instructions for using them in "impossible" situations, such as gluing them to Teflon. I have used them down to liquid nitrogen temperatures, and up to about 500 F.
In my data acquisition article, elsewhere on this site, I mention some of the high-speed strain measurement work I have done. We used strain gages for this. Strain gages will faithfully measure strain during violent impacts, if care is taken in their application. We attached them to steel plates and blasted the plates with high explosives. The strain gages easily recorded the bending and ringing which resulted. Similar work with drop weights and hammer blows never challenged the ability of these devices to record strain.
Strain gages are so light, they usually have an insignificant impact on the dynamics of the specimen. And they're so cheap, you can afford to put them in harm's way.
We did learn one interesting thing about strain gages in explosive environments. Strain gages work by having a small current pass thru them. This produces a slight magnetic field around the leads. Sloppy lead routing can result in some false signals. One particularly perplexing signal encountered during the explosion tests was a series of large, very fast spikes that occured before the primary strain readings. What was really bothersome was their direction. They seemed random, sometimes going the same way that strain would go, sometimes the opposite way. And it was quite clear that they were false readings, not strain. I finally figured it out. Slots in the test specimens were allowing the explosion's fireball thru. The supersonic ionized gas was shooting thru the slots, and breaking into turbulent whorls. As the whorls of flame passed the current-carrying leads, magnetohydrodynamic effects induced signals in the leads. The whorls swirled at random, to either side of the slot, producing the random signal polarities.
Magnetized metal could also produce this effect, of course. The fix is the similar: careful attention to lead routing, the use of closely-parallel leads or twisted leads, shielding, and attachment of the leads so they cannot vibrate.
Micro Measurements is my favorite source for strain gages, adhesives, prep materials, and other accessories and supplies. Omega is another readily-available source.
Heh, heh, heh ... you hire me, of course! I've done it for years. I've instrumented test specimens and structures for load tests, built custom transducers, used half a dozen adhesive systems, and learned what bad things happen when you take too many shortcuts.
But, if you insist on trying it, the suppliers above will gladly sell you the stuff. All you need to do is leaf thru Micro-Measurements 3-inch-thick catalog, select the gage pattern, resistance, temperature compensation, backing, adhesive system, application materials, select a bridge configuration, figure the allowable heat dissipation in the gage, calculate allowable excitation voltage, figure the required amplifier gain, calculate overall sensitivity, figure out how to put alignment marks on the specimen and get the gages precisely aligned, maybe design and build some clamps to hold the gages while they cure -- you'll get the hang of it. Just remember, when you discover that you've bonded yourself to a railroad track with reagent-grade Crazy Glue, that I told you so!