24 Jul 2024


Wireless sensor networks will predict motor failures

Wireless sensor networks will predict motor failures

A group of US organisations, led by GE Global Research, is developing a wireless sensor technology for monitoring the health and efficiency of electric motors. They reckon that the technology could be saving the US around 35GWh of electricity a year – equivalent to about 2.12 million tonnes of carbon – by 2020.

The three-year, $6m project is part of the US Department of Energy`s $61m Industries of the future programme, designed to improve the energy efficiency of strategic US industries. The DoE is providing half of the funding.

GE`s partners in the project are Sensicast, which is supplying a self-configuring mesh networking system designed to eliminate the risk of interference and to integrate wireless networks into existing plant networks, and the Rensselaer Polytechnic Institute (RPI), which is developing models for analysing and predicting motor lifetimes. Texaco is providing motors at some of its plants to act as testbeds for the new technology.

The sensors (shown above) will monitor critical parameters on each motor, such as vibration, temperature and power quality. This data will be transmitted to a computer for analysis. If any potential problems are identified, plant personnel will be alerted by phone, pager or email.

The technology is said to have two key advantages over rival systems. First, the low-power sensors are being designed to operate for at least three years, producing one report an hour, from one battery. They may even be able to keep going for more than five years.

And second, unlike many other wireless technologies which can only send data in one direction, the new system will support two-way-traffic, providing the facility for control as well as monitoring. For example, if it senses that a machine is running too hot, it could be commanded to turn on a fan.

As part of the project, GE is building up a database of the behaviour of 31 different types of two-pole AC motors. RPI is developing a software package called Condition Forecaster, which will predict a motor`s remaining life from a combination of the sensor data, historical repair data, and expert knowledge. The aim is for the software (above) to predict the motor component most likely to fail next, the probability of this failure, and when it likely to happen. The software will also be able to provide real-time data and trends, as well as tracking spares stocks.

The project, which started earlier this year, is due to run until the end of 2006.