A bhangmeter is an electro-optical instrument, usually carried by satellites but sometimes by aircraft, which recognizes a characteristic signature of a nuclear explosion. As opposed to conventional explosives, which have one primary flash that follows the detonation wave, there are two flashes as a result of nuclear detonations (NUDET). These need to be discriminated from closely-spaced lightning strikes, or other natural events such as an auroral discharge. The actual detectors are principally in the infrared range.  In other words, a bhangmeter is part of a set of instruments that are evaluated, as a whole, to determine if a nuclear event took place. NUDET sensors are used both as part of the national technical means of verification in nonproliferation, but would also be tactically important should a nuclear war continue over a sustained period.
Why a double flash?
The first flash, which lasts for approximately 1 millisecond, is extremely short and bright. This is the true nuclear fireball and is the brightest part of the explosion. After a delay of hundreds of milliseconds to low seconds, the second flash takes place, lasting approximately the same time as the delay preceding it.
That first is the actual fireball, which is then obscured by the expanding shock wave, caused by atmospheric heating by X-rays. After the shock wave obscuring the fireball passes, light from the superheated atmosphere behind it becomes detectable.
1979 Vela satellite event
One particular event, which triggered U.S. spaceborne bhangmeters, remains controversial. It is agreet that on September 22, 1979, some unusual event took place somewhere between the surface of a remote area of the South Atlantic, and the orbit of VELA HOTEL satellites carrying the bhangmeters. Some intelligence analysts believe it was an at-sea test of a nuclear weapon, while others, pointing to other sensors that did not detect any anomalies, think it was a false triggering of the bhangmeter. 
Part of the challenge of the inconsistent results between the bhangmeters and other sensors is the question: if it was a nuclear test, who did it? The location would strongly suggest South Africa was involved, but South Africa, which voluntarily gave up its previously unconfirmed nuclear arsenal, denies it, and sufficient other information about more controversial matters has emerged to suggest that they have no reason to deny it. The International Atomic Energy Agency confirms that they cannot find evidence, in the actual inspection of South African facilities, to suggest they detonated a device.
Other theories suggest it was a test of an Israeli device, or a test of an Israeli device with South African support. Israeli navy ships were not unaccounted-for during the period of the test, although it certainly could have used commercial ships for the weapon platform and for monitoring.
No convincing explanation of the event either definitely being nuclear, or definitely being something else, have ever surfaced. 
- Mauth, G.H. (1980-05-01), Sandia Laboratories report on the Sep 22, 1979 Vela satellite event - Alert 747. Scientific analysis of the satellite and event., pp. 11-32
- Hones, Jr., E. W. ; Baker, D. N. ; Feldman, W. C., Evaluation of Some Geophysical Events on 22 September 1979, NTIS No: LA-8672/HDM
- Sublette, Carey (September 1, 2001), Report on the 1979 Vela Incident
- Jeffrey Richelson (2006), Spying on the Bomb, W.W. Norton & Company pg. 103, 133-135
- The Vela Incident: Nuclear Test or Meteoroid? Documents Show Significant Disagreement with Presidential Panel Concerning Cause of September 22, 1979 Vela "Double-Flash" Detection, vol. George Washington University National Security Archive Electronic Briefing Book No. 190, May 5, 2006