Title: WiBeaM: Design and Implementation of Wireless Bearing Monitoring System
Abstract:
A sensor network is composed of a number of tiny sensor nodes. These sensor nodes consist of sensing (internal/external), data processing and communication components. Such nodes combined with MEMS sensors can form a complete replacement for the existing cumbersome predictive maintenance systems used in many industries for maintenance of machinery.
We have designed a system that can be used to monitor the bearing conditions of induction motors in a ship. The system called as WiBeaM (Wireless Bearing Monitoring system) forms a network of sensor nodes that sense the vibrations on a motor and transmit the same to a central base station, where the data is further analyzed through signal processing for defect diagnosis. WiBeaM uses MEMS accelerometers, that are placed on the body of the motors to measure the vibrations generated by the bearings. We have used cheap but reliable off the shelf components to realise a real world application at a very low cost. Considering the cost and the ease of deployment, our application is one of its kind in the domain. The solution has been implemented on Moteiv's Tmote-Sky nodes using ADXL105 accelerometer from Analog devices. The solution consists of an application layer, transport layer for reliable data transfer and a MAC protocol for low power consumption. The challenges in the implementation span from the hardware to software. The aim was to select a cheap yet efficient sensor that could be compatible with the sensor node and build a software that can automate the otherwise cumbersome process of bearing monitoring.
With proper selection of the hardware and efficiently handling the measured signals we have married the MEMS and wireless technologies which were otherwise being used independently. Whilst developing the application we have found that by proper selection of transmit power of the radio, single hop wireless transmission is more power efficient than the multihop. We have taken an application driven approach to design and implement the application and transport protocols.
In sum, we have proved the viability of machinery vibration monitoring with wireless network technology by conducting extensive testbed trials.