Institute of Industrial Research
The Institute of Industrial Research has a successful history of analysing industrial sensor data that leads to groundbreaking technologies and developments. Our expertise in this area extends to include:
- Statistical Analysis
- Feature Extraction, Classification and Anomaly Detection
- Decision Making
- Statistical Probabilities Analysis
- Artificial Intelligence Application
- Systems Hybridisation & System Architecture Development
Our work in practice
Working with Stork Food and Bottling Systems, the Institute of Industrial Research established a condition based monitoring project that analysed sensor data and failure records on Stork machines with a view to creating a predictive pattern algorithm that could begin to predict catastrophic failures.
The end product was the Virtual Engineer, which utilises a statistical feature extraction technique and a novel verification model to predict machine failures. The cost of each breakdown is estimated to be in excess £300k per day, with each break down lasting 3-5 days, these faults would have cost the company in excess of £1.5 million each. In an 18-month period, the Virtual Engineer has prevented 14 instances of catastrophic failure.
The Virtual Engineer is completely unique to our company. It has allowed us to add real value to our machines, and even works on slow rotation machines, on which it is traditionally much harder to spot faults.
Luke Axel-Berg, Sales Director, Stork Food and Bottling Services