A fail-safe is a mechanism or system that is designed to prevent harm or damage in the event of a malfunction or failure. While a fail-safe may not be able to prevent a failure from occurring, it is designed to minimize the consequences of the failure and ensure that the system remains safe. This is achieved through the use of redundant systems, backup systems, or other measures that are put in place to protect against the potential consequences of failure. The following are examples.
Elevators are typically designed with special brakes that are held back by the tension of the elevator’s cable. If the cable snaps the loss of tension causes the brakes to be applied.
Railway trains commonly have air brakes that get applied automatically with the failure of the main brake system.
Elevators and trains in earthquake prone regions such as Japan are often configured to detect earthquakes and automatically stop. Elevators may be designed to stop at the nearest floor and open their doors. In some cases, such systems are hooked up to earthquake early warning systems and can potentially stop seconds before an earthquake arrives.
Flight control computers are typically designed with redundancy so that if one goes down another kicks in. They may also be designed to detect a flight control computer that suffers from “insanity” meaning that it appears to be dysfunctional due to damage or other factors.
Electronic locks that are designed to be unlocked by default in the event of power failure. Some vehicles don’t have this feature and it’s possible to get locked in a car that loses power.
Traffic lights may be designed to blink red in all directions if their controller goes down.
Aircraft are designed with some ability to glide and can be landed without any engine power. This isn’t completely safe but is certainly possible with a competent pilot if there is a suitable landing spot within range.
Some submarines are designed to automatically drop their ballast in the event of power failure causing the submarine to surface.
Many machines are designed to shut off if they detect something is wrong.
Computer services are commonly designed with redundant servers. When one server fails, another replaces it almost instantaneously. This allows many services to maintain an uptime of 99.999% or greater.