Please help! Powelectrics need data sets for anomaly detection research they are sponsoring at Keele University. Ideally, Powelectrics would like time series sensor data, including ‘incidents’.
Examples would be temperature and/or vibration of a motor, fan, pump, gearbox or bearing with the ‘incident’ being a degradation in performance or failure. Frequency of readings is not important.
Anomaly detection involves the identification of novel or unexpected observations or sequences within the data being captured.
PhD student, Andrew Cook has already published his preliminary paper as part of the Keele’s Smart Energy Network Demonstrator. Find out more.
Dave Oakes, Powelectrics MD, told us “We are delighted to be part of Keele’s ambitious SEND Smart Energy programme and particularly proud of Andrew’s work.
“The end result should be a software model that will detect and identify anomalies in time series data sets from IIoT devices, such as sensor data coming from our Metron telemetry hardware.
“I am excited that this combination of academic prowess and our own 30 years of commercial experience developing ‘connected sensor’ solutions will allow us to remain at the forefront of IoT design and incorporate AI into future solutions.”
In order to test the model, Powelectrics would welcome data sets from a range of different situations. Ideal data would be:
- Indexed to indicate when something was wrong
- With sensor scaling information
- With flags telling the model which data to ignore eg) false readings for a known reason ie) when a train is in for service.
Keele University’s ground-breaking SEND programme is a European first … a world-class demonstrator facility for smart energy research and development in a life-size laboratory!
Keele is working collaboratively with partner companies to develop and assess a wide range of innovative technologies, aimed at reducing energy usage and cost … and saving around 4,000 tonnes of carbon dioxide each year!
Over five years, Keele University campus will effectively become an ‘at scale living laboratory’ The campus is an ideal test site. It is the size of a small town and with the same diverse range of activities and facilities. The 600 acre site has 341 buildings, 3100 students in halls of residence, 1000 commercial users on the Science and Innovation Park, 200 ‘standard’ domestic households and academic activities serving 10,000 students. More here.