Subscribe to our newsletter to receive the latest news and events from TWI:

Subscribe >
Skip to content

Intelligent vibration sensing

A recently completed collaborative project focusing on the use of vibrational analysis for remote condition monitoring (VA-RCM), and part-funded by the Technology Strategy Board and the Rail Safety and Standards Board, has successfully developed a system to detect wear and defects in train door machinery before breakdown occurs.

This early-stage fault detection may make it possible to carry out low-cost repairs at scheduled maintenance intervals, avoiding costly replacement of major door components and improving component lifetimes.

Improving automatic door control

Technical faults cause 19% of all transport delays in the rail network including depot activities, and malfunctions of automatic doors account for 20% of these, ie almost 4% of all delays [1]. In addition, door failure should be regarded as a serious safety issue as there are potential dangers to human life from a train door opening prematurely or closing too hard through failure of the force control system. Delays and cancellations result in lost revenue, franchise fines [2], sometimes compensations paid to passengers and disaffected passengers turning to other forms of transport, as well as loss of productivity through passengers arriving late for work. The VA-RCM project was set up to develop remote condition monitoring to assess the train door control system and hence its operational performance.

Vibration monitoring is a long-established technology, as is the use of accelerometers for vibration detection. However, present commercial practice consists of periodic measurements with handheld vibrometers. These instruments are general purpose devices applicable to all machinery subject to vibrations and it is left to the human operator to interpret the captured spectra. Online condition monitoring involving arrays of sensors is very rare in commercial use. Such technology is still in the research, or at least, pre-commercial stage.

Existing approaches monitor electronic parameters and door opening/closure time profiles which provide only indirect indications of the source faults of door malfunctions, which are usually mechanical. Vibration changes are the most direct indication of all mechanical faults in door mechanisms. The low frequency vibration response of accelerometers permits full door coverage from one sensor location. No other sensor for detecting mechanical changes can achieve this.

The VA-RCM system developed in the project:

  • provides a single on-line vibration condition monitoring module for total coverage of automatic door mechanisms
  • uses a single accelerometer and single channel data acquisition system to provide total condition monitoring covering all critical door components (making the low cost possible)
  • provides automatic early warnings of which components are likely to develop faults to the extent that the components require replacement at the next scheduled maintenance interval.

The system will automatically detect looseness, change in air pressure, wear in door rollers, the linear shaft assembly, ball bearings and misalignments in the shaft and door panels in the very early stages, before breakdown of the door mechanisms occur. The intelligence of the VA-RCM system will arise from the signal processing applied to the analysis of the received accelerometer data.

Test phase and prototype module

The project has generated positive test data in characterising failure modes in train doors using vibration analysis. Low-frequency vibration measurements have been gathered from a pneumatic MK4 train door made available by East Coast for experimental purposes. The problems related to the operational performance of the door are bolt looseness and change in air pressure. The defects were simulated on the test door and statistical data analysis was carried out whereby both the bolt looseness and change in air pressure were detected. All measurements were captured using a single accelerometer.

This figure depicts the results of an algorithm developed to detect faults in train door mechanisms using the Euclidean distance.

The design of the VA-RCM module has been completed and the beta prototype is now in the manufacturing stage. The data analysis process is being coded into a low level language and embedded into the NXP LPC4357 ARM microcontroller. Once completed, both the firmware and hardware will be tested experimentally for failure detection.


  1. Review of Railway Remote Monitoring Activities, Section 4.1 page 31
  2. REF 2 'TAPAS’, ‘Our recipe for a reliable journey'

For more information please visit or email