Railway condition monitoring is the use of sensors and cameras to watch the track, the trains and other railway equipment, and to predict when each will need maintenance, before a failure stops a line.
Railways have always checked their equipment. On Indian Railways, staff walk the track to look for cracks and loose fittings, and depot staff inspect trains at fixed intervals. Monitoring adds measurements between those inspections. The sensors are fitted in two places. Some are along the track: cameras that photograph the rails and the sleepers, and sensors that measure how hot a passing wheel bearing is. Others are on the trains themselves: sensors that measure vibration in wheels, motors and brakes. All of them send readings to a central system. Software compares the readings with what a healthy part looks like. When a reading starts to drift, the software predicts how many days or weeks remain before the part fails. This approach is called predictive maintenance. It replaces the older approach, in which a part is changed after a fixed number of kilometres or after it breaks.
For instance, a bearing in a metro train's wheel slowly gets hotter than the others over three weeks. The system raises a warning. The depot replaces the bearing at night, with no service lost. Without the warning, the bearing could have failed during the morning rush hour, and the train would have blocked the line behind it.
The readings also help to plan the work of the maintenance gangs, and they save money in two ways. Parts are not replaced early while they are still good, and a breakdown is less likely to stop service. The system sends the findings about the track and power supply to people who repair them, and the findings about trains to the depot. It also feeds asset performance management (APM) software, which plans maintenance across a whole company's equipment. Signalling equipment can be monitored the same way.
The railway buyers are governments, who buy through tenders. Read more about the setting under railway technology.