The data getting generated from IOT sensors fixed on the machines is analysed using survival methods, regression techniques as well as classification algorithms. This helps to generate the failure conditions of the machine and thus the trigger can be created to alert the failure of the machine well in advance.
Here is a very good example for IoT:
HEAT PUMP (R)evolution – “Google for Heating Systems” https://prezi.com/v/view/bQCuOOex8ZTgb3WjtZaD/